diff --git a/EMRI_DET/nn/model_train_test.py b/EMRI_DET/nn/model_train_test.py index 16f0a9d46c553e57e0fb84e3f17bf0c34fee3861..1b279522e1a2076649baf5a95e722f4296653782 100644 --- a/EMRI_DET/nn/model_train_test.py +++ b/EMRI_DET/nn/model_train_test.py @@ -54,7 +54,7 @@ def model_train_test(data, model, device, n_epochs, n_batches, loss_function, le for param in model.parameters(): param.grad = None outputs = model(inputs[i * ytrainsize // n_batches:(i+1)*ytrainsize // n_batches]) - loss = torch.sqrt(loss_function(outputs, targets[i * ytrainsize // n_batches: (i+1)*ytrainsize // n_batches])) + loss = loss_function(outputs, targets[i * ytrainsize // n_batches: (i+1)*ytrainsize // n_batches]) loss.backward() optimizer.step() current_loss += loss.item() @@ -74,7 +74,7 @@ def model_train_test(data, model, device, n_epochs, n_batches, loss_function, le for i in range(n_batches): outputs = model(inputs[i * ytestsize // n_batches: (i+1)*ytestsize // n_batches]) - loss = torch.sqrt(loss_function(outputs, targets[i * ytestsize // n_batches: (i+1)*ytestsize // n_batches])) + loss = loss_function(outputs, targets[i * ytestsize // n_batches: (i+1)*ytestsize // n_batches]) current_loss += loss.item() test_losses.append(current_loss / n_batches) diff --git a/emri_data/schwarz_data/grid_dataframe.csv b/emri_data/schwarz_data/grid_dataframe.csv new file mode 100644 index 0000000000000000000000000000000000000000..06abd449e4153a5e6e245aae2bcb60db06128f58 Binary files /dev/null and b/emri_data/schwarz_data/grid_dataframe.csv differ diff --git a/emri_data/schwarz_data/samp_dataframe.csv b/emri_data/schwarz_data/samp_dataframe.csv new file mode 100644 index 0000000000000000000000000000000000000000..7541c07824187e217ed1513ce1c0ea02727ff9b7 Binary files /dev/null and b/emri_data/schwarz_data/samp_dataframe.csv differ diff --git a/emri_data/scripts/data_preprocess_dataframe.py b/emri_data/scripts/data_preprocess_dataframe.py new file mode 100644 index 0000000000000000000000000000000000000000..72379841c2140dccc1cb4890dbf27de08de5a056 --- /dev/null +++ b/emri_data/scripts/data_preprocess_dataframe.py @@ -0,0 +1,62 @@ +import numpy as np +from pathlib import Path +import pandas as pd + +data_directory = '../schwarz_data/{}' + +# Grid +snr_grid = np.load(data_directory.format('grid_snrs.npy')) +intrinsics = np.load(data_directory.format('grid_intrinsics.npy')) +angulars = np.load(data_directory.format('grid_extrinsics.npy')) +plunges = np.load(data_directory.format('grid_plunges.npy')) + +intrinsics[:,:2] = np.log10(intrinsics[:,:2]) + +grid_out = np.zeros(shape=(snr_grid.size,10)) +flat_snr_grid = snr_grid.flatten() + +num = 0 +for intrinsic_set in intrinsics: + for angular_set in angulars: + for time in plunges: + out = np.zeros(9) + out[:5] = intrinsic_set + out[5:8] = angular_set + out[8] = time + + grid_out[num,:9] = out + + snr_here = flat_snr_grid[num] + if snr_here == 0: + snr_here += 1e-6 + grid_out[num,9] = snr_here + num += 1 + +save_dir = '../schwarz_data/{}' +Path('../schwarz_data/').mkdir(parents=True, exist_ok=True) + +cols = ['logM','logq','a','e','Y0','thetaS','phiS','thetaK','t','SNR'] +df_out = pd.DataFrame(grid_out, columns=cols) +df_out.to_csv(data_directory.format('grid_dataframe.csv'), index=False) + +# Samples +snr_list = np.load(data_directory.format('samp_snrs.npy')) +inds_to_keep = ~np.isnan(snr_list) + +intrinsics = np.load(data_directory.format('samp_intrinsics.npy')) +angulars = np.load(data_directory.format('samp_extrinsics.npy')) +intrinsics[:,:2] = np.log10(intrinsics[:,:2]) + +samp_out = np.zeros(shape=(snr_list.size,10)) + +for i in range(snr_list.size): + here = np.zeros(9) + here[:5] = intrinsics[i,:5] + here[5:8] = angulars[i,:] + here[8] = intrinsics[i,5] + samp_out[i,:9] = here + samp_out[i,9] = snr_list[i] + +samp_out = samp_out[inds_to_keep,:] +samp_df_out = pd.DataFrame(samp_out, columns=cols) +samp_df_out.to_csv(data_directory.format('samp_dataframe.csv'), index=False) diff --git a/mock_data/2d_function/models/model1/function.pickle b/mock_data/2d_function/models/model1/function.pickle index 60ce0949f4cc72bd21c1d2e540e7cb6ec93acbe4..e1117cd46fd5300e779ff8a256b3b0626fc4926f 100644 Binary files a/mock_data/2d_function/models/model1/function.pickle and b/mock_data/2d_function/models/model1/function.pickle differ diff --git a/mock_data/2d_function/models/model1/losses.png b/mock_data/2d_function/models/model1/losses.png index d430db4f2150e7104a70d3e7f0f60f87ab9db0e3..2bca0cc6e65da937fdffbf30654414167b88c662 100644 Binary files a/mock_data/2d_function/models/model1/losses.png and b/mock_data/2d_function/models/model1/losses.png differ diff --git a/mock_data/2d_function/models/model1/model.pth b/mock_data/2d_function/models/model1/model.pth index 7223bc32cbd1eae9ceead30bce4c84954a5797ff..48a8cca02d9d118244103ad2b34abcbe309b0ef1 100644 Binary files a/mock_data/2d_function/models/model1/model.pth and b/mock_data/2d_function/models/model1/model.pth differ diff --git a/mock_data/2d_function/models/model1/xdata_mean_std.npy b/mock_data/2d_function/models/model1/xdata_mean_std.npy index 5bc507df2f022689dd26495a54970333c6080786..aee69c76f75246f125bb6d648edb5c01bc0bdde8 100644 Binary files a/mock_data/2d_function/models/model1/xdata_mean_std.npy and b/mock_data/2d_function/models/model1/xdata_mean_std.npy differ diff --git a/mock_data/2d_function/models/model1/ydata_mean_std.npy b/mock_data/2d_function/models/model1/ydata_mean_std.npy index e2d9a31bc98fa19b41e7985ea865d522d8361061..7290c4c422be333149f86b202a8247001f8caa5f 100644 Binary files a/mock_data/2d_function/models/model1/ydata_mean_std.npy and b/mock_data/2d_function/models/model1/ydata_mean_std.npy differ diff --git a/mock_data/2d_function/notebooks/spline_example.ipynb b/mock_data/2d_function/notebooks/spline_example.ipynb index 2ededbb19187786d6b407a151eef1089626e2487..0a2b9efea755bac462623241f1041d8d10daf01a 100644 --- a/mock_data/2d_function/notebooks/spline_example.ipynb +++ b/mock_data/2d_function/notebooks/spline_example.ipynb @@ -57,7 +57,7 @@ "outputs": [ { "data": { - "image/png": "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\n", + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQsAAAELCAYAAADOVaNSAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAACkyElEQVR4nO39f7Btx1Efjn56Zq19zrmS/IQUW1xhKH/xo4Riy/mm7JSBZ/MF+ZcQsgXEwCsZKuBnOZgE1yMVgpwH8s8iEVVUChMUsF8QdkxeEr18Y5eFnzHG+dqYHwYSHjJRbF6MEpv42gIZla1779l7rZl+f/T0TM/stffZ59xz7z3Cu0/t2vusvfZas6Z7PtPd09NNzMzY0pa2tKUDyF3uBmxpS1t6YtAWLLa0pS1tRFuw2NKWtrQRbcFiS1va0ka0BYstbWlLG9EWLLa0pS1tRFuw2NKR6eabb8Zv//ZvT373B3/wB3jJS15yiVu0pYtJ3eVuwJYuP733ve/Ffffdh4cffhhXXHEFvv7rvx4/9EM/hOc85zlHvuZznvMc/Nqv/doxtnJLl5u2YPFlTvfddx/e9ra34Y1vfCOe97znoe97/OZv/iZ+4zd+44LAYkt/9WhrhnwZ05e+9CW89a1vxd13340Xv/jFOHXqFPq+x80334wf//Efx1133YV/9s/+WT7/Yx/7GL75m7+5usbHP/5x3Hrrrfhbf+tv4XWvex3m8/nkuWfOnMHf//t/H9/wDd+A5z73uXjTm950aR5yS8dGW7D4MqY//MM/xHw+x4te9KIjX+O9730v/uW//Jf49V//dTz88MO49957l84JIeDv/t2/i+uvvx4f+tCH8JGPfAS33nrrhTR9S5eBtmDxZUyPPfYYvuIrvgJdd3Rr9BWveAVOnz6Nq6++Gq95zWvwq7/6q0vnPPjgg3jkkUfwj/7RP8KpU6ews7OzNXGegLQFiy9juvrqq/GXf/mXGMfxyNc4ffp0/nz99dfjkUceWTrnzJkzuP766y8IlLZ0+WkLFl/G9Df/5t/Ezs4OPvjBD05+v7e3h/39/fz/X/zFXyydc+bMmfz5s5/9LJ7ylKcsnXP69GmcOXPmgkBpS5eftmDxZUxXXXUVXvva1+JNb3oTPvjBD+L8+fMYhgEf/vCH8dM//dO48cYb8eEPfxiPPfYY/vzP/xzveMc7lq7xr//1v8bnPvc5PPbYY/jFX/zFSV/Es571LDz5yU/Gz/zMz+DcuXOYz+f4T//pP12KR9zSMdJWL/wypx/8wR/Etddei3vvvRf/8B/+Q1xxxRV4xjOegR/6oR/CM57xDPz2b/82br75ZnzVV30V/vbf/tv4pV/6per3t912G175ylfikUcewQte8AK85jWvWbqH9x6/8Au/gLe85S341m/9VgDAS1/6Ujz72c++JM+4peMh2ia/2dKWtrQJbc2QLW1pSxvRgWBxzz334Oabb8YNN9yAP/mTP5k8J4SAN77xjXjhC1+IF73oRbj//vuPvaFb2tKWLi8dCBYveMEL8Cu/8iv4qq/6qpXnvPe978WnP/1pfOADH8C//bf/Fj/3cz+HP/uzPzvWhm5pS1u6vHQgWDznOc+p1tKn6H3vex+++7u/G845XHPNNXjhC1+I97///cfWyC1taUuXn47FZ6FBN0qnT5/G5z73ueO49Ja2tKUTQlsH55a2tKWN6FjiLE6fPo3PfvazeNazngVgWdPYlP7yL88ixifGSu61116JRx99/KJc+7du+X9U/1N+L33jmu/y/2kl3B5/wR/8HP7jc36kuo4zn+U8rq4Tzf3lGMORnO/BIJLzIxMCHCITIoCY7uLyPTi3xl4T5lw23zMI35LaG/MxgInyZ3vuKmmJZD6b3+l12rYo2evpNVw6+OL3vWXp/IspB8dNzhG+4iuuOPLvjwUsbrnlFtx///148YtfjMceewwf/OAH8Su/8iuHvk6MfCiw+P1n3rl8janrZqE0x5LQ2OOR6nPlf5j/y29e+T/fhf/9WT+Uz1klfJuqbvY8FU5Kx10+zuUY6/dsPttzGB4FBMJnPo8uDa886InhqAw5ZwaYEpEAgyPAuQjnOL9iJPNyCJHATNXvmOVYZIBAiEwokT0EhnzHqMFm8ZlHEMyxkPo+kP6/zBsly4+atwk4Gp6t4lHL03f+rz9U/e8BvPYz78J/eNYPoUt9L08l13zxH/3Ciis/MelAsHjLW96CD3zgA/iLv/gL/OAP/iCuvvpq/Oqv/iruvPNOvPa1r8VNN92E22+/HX/0R3+EF7/4xQCAv/f3/h6++qu/+qI2/D8/81VpMBSmRhA8kvCk/4uwUCNc08BgQSGa74Ba2ADgcWd/xysBQ3+bP1NpB5CEi4UZBMpgoM9HLO8dKAtiBwUFKqDACi4CFB0YPt35FI3wTgCCiOEd58FPJIPb+/oJiABy6XzPcD7CdQznpeUcCTEQOAIxJO0i1IjDFaAUYImM/B7YgRlJQ5Fre9M7Fd8SUATDp2D4FywomP62PIrpmJ08JnDSyBGbz0I9gD796ksOmDHQEWXgBoD3/q+vgQPgmZcA/lsf/MWJO55sOhAsfuInfgI/8RM/sXT87W9/e/7svccb3/jG423ZAeTBedb06VjUmQo66Bhjo6rqzKQgYIXOgsPULMQARkISHeCLjhGMAE4Bhv2fqT5XiUBwBHgmdAC8/q9AkIDEg+D1MxO6BCgdGD1TZUp0YPSImFEAAOx0AZ2P6LoI75OG4CO8Z5ATAFAQIMcgB5B+1wFuxqAOoI5E4hngCPDIQEyfI8CjgAcgQMGBEEcBkRgIITiEwSHG9DkUrWSMjJCQtANnXua+JGSgGIkwEDDCHgMCOPEIGJtBbsFCeaFktYvl8xmBC8c8EWZw2EmzxuOOscNAzwRP5XqegY6lrR0EJAhAx4TfeNYPrdUKn/vHb8dJoyfk3pCPP/OVaRauTRbRMjjPzkXjQDKdkwpMXFRcMzuNzewUzIw0ooBLSPd9HBEDKWDUAhXTpzIrssymWYB1YCe1HaI5eH0RoYcrIAFCD0ZHhJ6BPr3vsPzaJ82CIECqQDFL2sLuzoiuC+hmEV0XsobgOobrWcAg2TxuJoBAHYE6B5p5UOeAzsk7AI4MRAbGWP8fGTzGBCAMHiJ4ZPAIhH1CHAhh4QQ4BodxlM/j6DGODmOQ63cUAQYiHByAASRaBYBFAgp9BTAWBAzgzJ+BhGfCu8KbmACAuQCB5YPyTs8NYETDVwLQw2EHDrsk09QXEbFHhJ6APml3CvQxySWzgEz2/wAyExHDT1jev//MO4uZCfUz1X4lS8/441+aOHq89IQDi//yzFeKVkHLYMGQmTZwAQ2hZoYCZTXUAkWoZqplgBhRBBEAzlLAwDEJaQENVaUZQECsNInInIUQECHN5gQ5AwwOPVz6TOjJYQeEngkzEjsfSfissIkJwugoonMRfZc0i90B/SzA9wy/k8yJGcPtElyvoCDAAAUH70GzDvAO6DzIe8AlUY0RzAxE07ORgRCAMWTw4PkAjBG8P8KfC4gDI+6PiIsCHOPCwQ8RfvDwo1y/dxEcCZ4ZQ+JjoKJRDATME0DMiTFAPi8oYgRj5AQSYAyIedArnzj9H40MKWBUYMExaxg+8WgHHgt4DOm35yggJE2jJ0qAAuxAtSSVu8QrIiCZJjFpwAoqSO9TQGGd0upfoqSBfeKmH6x8PzddBPB4woGFTw45B7G1LclskVRBFl25VWWBAhIjUQaKYUKVDUAGghHIWsQi6dmP84gRjDkHEUKOGTCQ7q7HAsd8XD8LUBAcUf7ck0NHHj07+QyHHfKYMSOQwy458U0QoUvPq89THKLim+h8RN8LWMx2A7qdiG4vFpDY8aBdL5rDzIP2ZsDODNR3QOcB7+X/rgP6HlWHJ5W5BosIhAAOARQjMAZgvgDPF+DzC9CpBdz+CN4PiPOIuIjw+ww/j/BzBz9n0Fxm685FhEh5AFtgDwko5sRYgHEejECMOSIWHDGHAkbEiIhBeYBYeMQJMMxn5YX9TsGCElD05LGgDrvUISTJOqv8J4ceBM9pGksTErNMTqp1KDBEwqRWodQChXVEW01DvS9E6QymiatdOD2hwOKhZ/4gPC178ZWYCMQCGiAHMKVBWVTBQLXwWaBQ1VVV2xEJLIgxgLMWsUhCcp4DBkQsOGDB4yQoiHBGjDGkGasIJwB4clkQRbMQsNhxHTp2SSgZkbyYGQQ4cuhZ/DHq3AMXh50D4ClmHwWADBT+FMPtObhTHehUD+o96NQOaHcH2NsF7cwEGKrXDPBd1iqIHFgdE9G8cwSGRQIRBoYFsL8P7O+Dzu+Dz54Hn0/AsT/C7QfEWYTbj+IfMTLe+YgxOrhQZlz1JylfFCj2KWbe7CNgn0MGiSF/DhU/OGkV+T3xxaVGZMBgHaAkIE4egwsIFBGcDJ9zPCKSR2BGTw6zBOips6Q/uKw2EYv5S6pdGNBQv0YxOYx2YbQJnSydWcJ2ybTyFwcrnlhgYYHCk6JpAQxm6cgAB88RIJdVMwcBhJbUmTWqECbVtgKJBBQBMnvts1zpHA8YE1jM44iRw5IgBkQMUQR15FB9J8/k4BQoyKF3HjPqMHLAzHUZVFya9TomDKldvTpmDVAQimDpCgYA+B0BCn+Fh7uyF6DY2wGd2hWQOLUH7OwAu3sCEDu7oNkO0O8IUOgLAMiBuAEJQMBBlkaAMALDAD7/OOj8eWB2TrSUs+fBs33wbA6ejcC5AeQj4ERv0GVV7yNoWJ4Q1JcUqJgdC44YEDNv9nlMAB4ELNJLgUJBQ4EAEGAHasczGzPFJ41vdF60DVcAf44gYE1A0N+Qz87LDgI2I2ffcDJrChXTw/5/MFBQbh+DSZxsETKx/vU/vm9C4o9OTxiw+MRNP4jOAAVNgAWgdmdEgAMxJycTJyUf+dUSg7OvYgDnARkSUKhqO09CCABzHjFwxIJHLHjEGEU4rfYwxCKoYwwIMVazGIHgnUPnZNYa2SO4iBkJa7xz8BTgkw9jhFn+My+Y53JU4hy6PoHFrmgUdKoDXTEDXbELumIPOLUHOnUK2NsDdvdAs10BjNkeMNsF+h1QPwO6HdHWnHGvKVDoexikL8MIDHNgsQ/qe/BsF+hnoP4s2HtQ58UPQvui8bkRHBl+ZHCQ9jqXlnejzMSAzMbFn1R4NEdMGl7EPo+YJ7CY84ghjhkwsmbBETaNS6VhsPU5JRlLPAqUQMbVMrfPo5gcDIA8JHAtoiMHz+JjcSxmp66uVX6mxhQR7cKaGcrfZaBQXkssC8MnO4dbG/0Y6AkDFh6xAgoNFLJkPdsMBrN0mjBKzRFKzFjuzAjRLtIcJ5pEmrXGBBr7PGJMw1MFcuSARRyxSNqFAoUK5hhDBorAMQOGCqGPDsFFsJEgcgTPDgMH9OzElEFEIGeWaxkt9KmgadxEBtUZiQNztwPtzkB7qlGcAk5dIUCxuwfsXgHsngJme6AEGNQVDYMo+fOtZhGGGjSGOTCcB3c94BzId2DnAEcg1wAOAESG243gEOEWaaZMbZ/SLPISN2QmV6BYIBRNLwHFgsclwI7NNQFkniiIWFNR/ndg52W1iAkLJlDyDSw4yGREZTVrZAEzD5mw+uRUb2+dAT7JqNUu5HuufBTW9FCgAOTdgRCSqUq8/IwXSk8IsPjkTT8AnzqsBQqrWThQFgRicfhQQnT1WyjCO6qjIy0xcQYHBY3RCOOQBooCxRiLqjsmkyOCEWLM/1uwYONQY4jXi5gQko3t2OVrVGv9MA454rWOLBEkwHWpPzxkxaNzZXWj74HkvKR+Jr4J1SYsUMz2QL4TwCAHcmWhkQ1YcBiBMIqqzVFMkW4G9CMojOBxlNWSEIBhAI0BGEbwEECLAHIRCYvSoKifyS6Fl5cM6YCIkYU3YzL5Bg4ZKJQPam7IPZL/qhlYrb9C70oJIKyGAgADB3QQQA9wCMR5idbG5/Rcx93kdkzxD83qRwMUCqIq/zlydmICOS468WDx0DN/EH0yP1R4LFBMmSERouLL7C5DkowpchjKYJGEMTJnIVEQUOeZfqef44RmYWcrIC2lqtrbqL/WZtZnY5RZlfO5q59KtVFyZBDEpemHygnOAc7ryalxOboKzBEUBsD3EHRz+VRGD4RBHJ8xCIDYVRI1X1z6HZGstLgEWo7AMn1mnVwjR12jjlvAsECq2mD+Y7PsyTp4OQOE9icgDkYH9f2Y/iYdgJRNRl21aiWp0kJMu5gYkXVJdjXZlaxybFl71uPAsgl+senEg4XKd9tBm3aUgwRgKa36VcvIKkgneZnL4NRr1Z709jvVIhQMKAkcswit9Vm0jk5HotZqOzQeQwO2bLuZkPZGrO4T1gCqoOYCywzPDIQRHEYBg+Rv4KQ9MCDOTN+BfQ/QQr7zXQIBD8QAHhfyu8U+MM6BmK6lr3EsZoreO0aJxdCALl3uqPgiwViFL8vU8iofJ8pd4kAFBJvfJKMVuqgpPgDO/HJUVqs652VVJPmYAMDD8Ky59jpScGifiZpjU2EClvgiLZW2dKLB4r8885XoDCjUKx+0EjBs52kIuLV+bTi3goCS7MUg8Y+Y+UO91So80M/NLENJQPP8QwIGTmcXszxHjRD2SRDV866fva6EmKAtz2WfiHYDp+et+yK9B4DHCB4jaBiBxQDuO9B8LrP8/rmio6TYbe4FQHhcyEqIagiUtALfiaYRAzAuwIvzAjzjPDs4sdgH5ufBi/T/fC6vYQQPI3gxAosAXqRIzwQWbDalWd5UfErmZZmVpT+Ffw6epW/hks9BJ5xq8ihmHTFV2pxeT0G8dzVv1Am96zrM4NHBZR516UVc9vpUmwJR05S/Qtq6flLUPtLNegyJUmYQ/sszX3mskZ0nGiyW4iiyOifqWStM2mGARh4TGBY4kMK7izpvSZnZAQiQAKmeGSOJs7TXiSr9sCcPdrJTtne+qCdOTAgb4NPaxe2s1RmgmLkOM9ehT+v6PfkMEhoS3pvP6ovJ/QBC0E1bYxKmBYN7Bu+P4NkAdHOQc1mD0JmexxG0M8hg73fAnYmxcN4AhpoWvQDL4rz4LlSTGBaiYcyTtjGfA/vngf198Pn9FKg1B59fpIjOiLgA4pB4HOUZGMuA0Zl35dEAh544+Qy89Lnj7GNoYyayTBjT0ZOrNBNvNL2ePPrEE/2/S0N7j3rZK0Jeom3h0FHhU+GRALvj5Dszn+tnLA7MKcreLjspJoCIF1HLOLFg8UfP/L+h00HNIgSyU0A6I6T+XY7iLEIWEniMIIwoewtE27VqqKj3kTnvFfFpmt0hhyABHMJoknMBYM/NZPaCw4JHeHII7LPAWSFtfRUWKAgks1ZSZzvyGSx2qMuRnGqCdNDNZPLS2RX6bAyEtFFrHEVVDvsE6iKoC+BuyP4KAsBjAA2DDOidffBsR2Iu+hmoS05QNTkAgJI4q3ahAKHxFSGAh4UAxjAAiwQa8zmwPxewOLsPPjtHfHyBeC4gnmeEfcI4lwEYgjOzZbFOiJH2zSCFvrsqaE2sDVlJ0tWk1vdjSf1K1pwEkE3EnjwcOcyok0A5JOBIPAGAq6jPoOB1koGE5sveHrMpMP3fLuNP+StquRaWxaT1cpJX4blqrWR27673kRyFTixYACmEmQGQBFkFiCqparcjGRQWgRVhM2AYEySQLF81ZrHMUmlxQU2THrLdGEkYHSQ8fIRDl9hwBfUiGOQx4w4Dl7AvDchqg7CA4oV3aqagzGAeEubtyWEGnyICPXbhsYMSTkwoM6zNfwGIVhRYNmmFQQR6PO9ALgU/uUHEnBkYRqDvgJ0ZMOuBvpNVkp0doO/BfV+ckcbmz8uf5MqyaRgTUklAlq568DDksG/MF+Bzc/DZBeK5AfF8RDiXgOK8x5jCvcfRIXDZRQzIDNyR7LztgbxHhm0IPGQQ96kfh0arsMSQVa7se2rOUQ3CJx7M4DFLQNEnXgDAVdRlX4U3INGDMANhxrISortQ88pcNosKKUhw8hjp8ztwBRhofslcTO6yR+R4tYwTCRb/+ZmvSqsaQtIRLtnnZRlJZ+uhGf2KroGlw4pmUdv0OuB0RoogzJQjkCVWn+L6O1Ba02cRTgBXuZkEZVFIQge0Ds/stW8db7A2d9lEVvwSujeEBCTI5W3RuyDssPouSt4LAPkZAxPG4DAMafDNnSyjuuSnH1PI9e4gG8f6fdk01ndgXVY1e0RyfMSUp01nuHGUz2NaHl0M4JCWRxejvO/LK+4HAYp9IJx3GOcO49xjWKT2hrRtPfHLJbYoUIQMFADY5RiHngkdEXbIS7i3BlI1/q0pXmn/IfWngrcHCWCTAraYGX3q9SvYy1I2RNspWgYSUBhtECm/hZogy72ZKSZNttrfxLUGspRtjMv5rQl3oXQiwSKiDFYNws0BJ5CZ0wKHBQCLrGPSLNT0qByZaZbKXj39kgQwHEw2JnYYQAh5O7rQk9ClKEKftBKzDZrL+r88U03FQZcclaS2rVFnUbaq90yy6xQkuRPS+4xr7UL7IDBhjAUsFnNhdTcS4iLCzfQVQJ5S0FazFV3jMRzloKq89FoxLK1wRAaPIYGFOFORnKrymREXETww4kK3rAtQDAuPcWHALbqkSSqwyqysuSt2WHiFZJZ4ZvQgDOQwQ5Rdp4lfyrciX/XOYAUSCxhiLri0rZywiwTW5DBjl3KJSNt2kq9AtQltq4BX8q1kzYLTtZNDk+ul4QIMSf4n/Cx2dU99dDY1oc3rcpx04sDiY8+8s5olFTAs4gPis5CBVttt+ruYgUK0i5hV/+QYIgArAEO84BrnL8I4QPwhtg74VegqAMlgwZzBw8YDKFlPvgULVWVFmESrsSqtTyDRJ7V2lmZZz+XpRZOS3w/RgUZ5qPl+hxgIfhHR7Uf4Pko+i15yWlAX4boAko0MEsSVItfIEdCRvAMT0VKJB5GBkWVVI3DOY2GT4sjLgQMhDCmnxeAwDpLPYpHAYoiqGZbZk9Kz9obXDiQZwUh4PbAslUuqAY2RKXyBspm0vwpAtGBBhhe75GULegJt9RUBAhYWIBS8RYso/qVZ2n7gG62ibFNfniSRrhcKntRdb64hv6ClY8dFJwos/vNLXpc/lwctHQbUqBpQotumkr9qgpusNCDNwBr9SKKdIKVDS2H1KWmKrvEnzToFeVngelJ2JJVVegsOYSIzli7bdUBZVkufgWLPiolkMmMlkLDCJ++c7OAyEESjcRi1UQD25z3G0cM5yY7VdQHO1ynzxIfJJVNWzhHHObqSWjUGyE6gDAaBEFNSmzguL4FyTLk4IxCD+FY08Y0mvwlcnNKlXxgdp6hXCyCpbzStwJi2akfGUiYzlQ37zsZEsd/p4OgTSGStjuuQrN2Utaws4Qqpj8JDNIoMHrxsIChAWcAIaTJcpSG0EyRgQePLwMFZOks6vwaNqXNLOjlgubPagJ6IAhg6WPV+LtnBHkmrzrNPsgUbe/2qqOn5qAKJnFyHFbQKw0tcQElcs7zmLqCiApjBwfgnbF5HD67GrwCe3EXV1P3g4YKDd+Lu9a7LuTiJuMrF6ZyG1ctvSaMqSUAkhxw71SjK4GemnHdTVjTSNfRare+AKQOFpNZLYIE6NgbpWSOl1YRkKvZIyY4IGBkYiCreAZY3yzOtBX9ujmfQnjApYH63x20CIiHlrQUKG09hl0yt3OskZ6+13O4GgCef66+wZjHFLKWw5sHDBMLKZwMUaVm0BYyYBqWcz7XgcL3ECq6vf0VE3hwkx5NgEvLSldrYliwQ2JnIPrMm6VVA0ezRKrzaqjaPY3lu9aTL0X32AqopPwSF6Q1KU5m+q81KCVDaADl9bxPxyrXqHZL2Wnp+SEveIbVXHdOtZthx0TQ1w1SPov3NkoaxPLfUNrz2bRU+biYHe15OmGzAGkA2SXebKTyvdnABBgWKqclBm6p8s9dZJ/erNIepbPbHQScKLJRsqrFNz5+iVZ2lQqIMLGv4VOUYkIs0M6H5vCuhctXxmLQTFXJJ10cVEKjA2EE+5QBdJXAwv7Xn6/0lCUrt61mkM6rfqPMMyLsU223R9trtDkhL1gy0voY2ecsUOAVOYGGC6AI0A9gybzvdGdrcWwLopmfmNvCJmu/t9VbJTdEIkPeTAMAO11E09jz5vAzoQJmInPGlBBQTpzzD6mE/1T/HDRJKJw4s1Bo9jAp1UOdEWr6Wnc1VhfQbdrOetTOxdm+DiNRvMnI9O6pqWmsDy7SJwE2RAgZQZqZxsj/VH1Bm61WUBxcvJ5CV9peBXjzypu0oq1f5M5W4gsDFGa3Xs1qF9oc91vKNYXKwqi9WzaB8jZpnrfsFwFIczhR5LqshnYLHxL3aGjBTNAUaShcyFqbk/kLoxIEFcDzIuGlHHTTw1tGUnZrzbya1WAeHFQhN7w+UwWRrnOhxoAYJex81q1bROgFsKZhBvo5cNm1Wq8g2JyjMefob+X0CJ9Y8pJyBQgGn1Sr0eRWIHcOAUSFSbdG2uxnESis1zzWdVWtzcqJHnaHb/n5Kg1lFLc9sG4932B+NThRYHEd2n01AQplxWORtfzclAFZTUPsaZNX8ZW0BADQsVVX3cs/lax+l7etoagBsQtMDgNMzLJP19AMKIiXLdb3sV9rWtkgdyNp/1sSrfFLmGsutTNei9TZ+q5FMmY6rwOUwQGHJaob2OrYdwLTGoUcYy6bXhdKJAgtguaMsrRoelap6DB3UOrjycapn4HUzsc6A0qZatbUq64FtMbPolF9jHU0J3CqfxJRArvKltKnpc1tR8koGuGoQr1v+W/WdalwKGBGoKomFBBrA9HJkuUfTTjIaEBXTxlatsxqcXJeyudHxZv60YlIfDjBavi2ZTtUdlsn22XHSiQMLYD2yKmkHascoHbWDWuYXgWzbVj4Pzew+de9WGPP+AwLARdjtb0MWz3INO7OtavOUM0+v0WEaOFr/g3VmAsirIrYErWymS+s/2v7m4SXTed0/h1nOyxoC6iVP1Sq0T3WWD0SVxlZH9Zo26PNAHdBk+IMq6E6X0/W6gWTJWpMyZhmk+rksr7IGY645+bwrHbAtX6b6qKUyNo6TTiRY2AKzwDpktaCB9D7dQcuzixXAWi2111klaACw0P8bjcNed+n+OVJTBl211t7YrFaNVQApcRq180zJmRnWzv7efFZwsMWaVqVtswlhyyPUqQ11D4N+pzt/NWYi5qXVsiOy1QN05cZWaLNaiforGIAtOZmfG1ZODG9XaIMWHHQDodaNsWSXruVdontFk5GT7aTh2PiAVLM0vGoH95QmbIHITRzLz4aSz6Ntc3nK46MTBRbE3FSiXo2qLU11ULtpbNn2Lb+YUk3Ld0XgrICed8s29tRa/aplUdnyTmhVdY0HaZdXdUBI0Fhtt7dmhAUVAOgRK3CwJRWci1X8hMtBWCUuwhmVxgZs5X13QE6Fx0wIKSrTFkIGUMVgqObBoBzLUBLPLM+OumM4koZzNx3baHpVwBVNA4bWrrXV5zQiV/pPwUFKRra1ZgFgbnhod5LqsZZXq8ySKZBvx0G7K3UVHbdWAZwwsFDKM2fzP7BajVMVV6uoA6qO6vfG9jWzkqqiFiQ0LsKqp63gAcA5glGJWw0nva/gWZdmY1s1Xc+3KqvGZORoQEjUYtYUUr3MVssg1FXUZxTRu7hURd17zoWSc4FkE6GZg6hSx5MBEqUqPBxAHGXPRxhS4eMUyak5Kmwx5JxKIKZ8I6lze8Ts99D+1OVUre+SA+Cavm35FgxvrNkn1xTOaWlKuWYpJ0i6N4dEthQ4PBWwOE8WzKkGd+bk50gDmKYBY9k0bP+v/URa3mKdp+av/K5TCxTLoFHPOkoKEkszdAUey+G+Fihs8JRW+bIqqg2yUmA467iaicp5RjjTvVzzGaSbjtJSamUW6XPSknD2hBSSrtvppZHVCgoKUPRpsO96raIepCaHj/BdzPtCfJ/AIoWHFgBARrICGOVG5bPxEYwRPIZcz1TBIwYJBVcAGbP2UYBDoz596tFohg6bl9altaCg/T7m7wsQ2M18Cgj6WUEilzVk4UHOe0qmSDXrTuC0CTHJgYf4Njw1IA9Kg0zkmLkGDBj/3JTcT5UvVNmQybDIde2rOXxy6oPoxIGF7TB1yk153606VlRWPb/Y9yWSUM4ORvDUs65AUYSshGmLairCpBuV9D5nEavdjXamaitvM2oQICBvS887FZu+0FktZ4aCZv0CdtKVFTCKgMp8QmD0JAWSAWDWj+i6iH4W0HVl16mfxVJFvYNUTncAJanXnacKBtXOUzexEzVVUecxSl7NgVPKvAAe08ayZsepmCxaRV36yDd+Eu3LkShXj1tXwForyi0BAbSZ9Y5gm15A3zNYsIKFAgXlJEQA8DhxrXVAdwonYGLRZrvEK0qAobEZUxOkb0CiLQQuMlZAQ6nVpo+TThxYAAUoWlSt9itksDDH0j8FJNRhVhDXdl8EZaAYUFdSz/VOGyHU/AgA8HiqsRm43gptiyMjtcHmuAA0B0Odv6J23KZq3CZF2w6V3ByAmlKEnRzeLd90YPSI6Chi5lNh5FlA1wcBi50oldRnDL8LUE9wMy/b0DWnhaP8nnNYKDhoPgv9zlJKfsNjTElvQnmNXMBjnxAWEd28rqJOKfmNVlF3rJokVeagli8cSYtXo6ok1/IlGBDQauq5yTDA0TgcFdA1GW9ngEM1i3MUE6CIJthDJxXKgDFjQKunT2mDKpsWKKYc0KWbORdCjvVlsswDwG8/89X4pj9+G46DThxYqBPPAoVWTgeMA0ltz7xPQBiluQkLSJTOa9f0VaXNMxMBA+oCybmEoR6DCCEAfCkVQ9ZjUmSGc9YsoNYqbIEbAJiRh2aQVsAA1OmpmbAiduBSrknCLjtoJi8kgOlAeb+E3JNTXVgpjgwA/U7AbHdEtxvhdyP8LuB2HGjXwe12KVuWl2Q3s05S6XUelN4tYFTJb1qwCJL8hkIADyNoMQhoKHjsj4j7I3g/ws8lpZ6fR7jzvsrapzy3RXPUFMwaBQELlHKTc3Cuoj4gVkWibIV7zY2q17S0lIeTJaemh5Qi1NybkihYGnwOEX3Kp+FZAL2nktO1V34nwJguiLwaKNQZrbxlUNJQLGgYMzbJvzN9dxx04sBCqQWKdmkvU+oL6TgAKIirpypgVCsUVDzrChgjSoHkObEAR56tJPvSAKnWDQBneUSAVvBOBYUQlxL0jqZytyWb49En0AAAzy5pFR49JFt1ZIZkpIkoaeTScl/yZYCN0BGj8xF9nzSLBBTdqQh/ilIl9R602wE7PWinlwrqNg+n5t/sOgMWreMikQKhptYbBtAwSKq9YQSlHJw0X4DmA/jcADo3SBLhczH7SfQyzkX46PJeFOFs8U+oqTGAcZ4Y+xywgACFFkrWCuqBFThKcl4LGLU46XCkzCOvWkXih+T49Dmt3jkOov2lCuo5WW7WJMqAnbFOBmKG2EC2CjBIYlnshruiWRvgIMqgIbJOOav8cdOJA4sIZCTWmWXdFmpVwbSKtC0975RZsHEKddEh9ZDbytyDahgJKBaImHPEwsxYAHCWhwwWuVxeKoysiF8V49XZzMxcXUr1r7UpiqYhs9ceSW8QAU7rYUBsZM2gpc5VJQHZmBya8k23k4DiCoI75eGunIH29LUrFc73doHZLCfrRS9lAEhLAShgAOVdK4+l1HocUjGhVA4gZw1fLFLNkAVwfg70c7jOAW4BkCjtYCCOAm6djxhDhG88OWqCxMSjeQKKfUTsI6Q6p/I+JsCYKoxsM69b53L2d1EqAGXKAXSmNMMMEUMGixE75EsiJEqeI9UC0z0cBOQUJMQBrwPf8K8BirYQuPpyVGNW0LCyf5wahdJGYPHwww/jrrvuwmOPPYarr74a99xzD572tKdV5zz66KN43etehzNnzmAYBnzDN3wDfuInfgJddzQ8Mj61SXTVzvOEUk+EpWMDXAUYxbttN02V5dFsB6NUUg8GKFRzGBBlBktZvM/zUAmjraKu7VEhXZXlW4XQCqVPqefZSUZHm58zp8Ln5G9h5JwaqnRqP+lKBwBxYu4S3G7SKPZSJfW9HamkvrcH7KYK6ju7oFkphlwXGFqhUaRqY6RgMc6BcZTiQjtSN4RmM6DfL+ZM4jPiPBdF9imzjPesPlRYZbLwCsgV7yF8WnBI/BmXilZPgYWtpt6WNfTklvjTk0dwHQJJAuA+9cUCMTkxWaSMAZCDVhZzJMmexbQs5lS5HxvTu5b5yULgxIhcgENzdTpKmgULeCA5Vo+LNhrJr3/963HHHXfg9ttvx3ve8x7cfffdeOc731md8wu/8At4+tOfjre97W0YhgF33HEHPvCBD+DWW289dKPqWXJCVWySrygV4IgVYIjjaHl3ZfaMw667q4OTsxYxR8zCqOAAAPtxyEChAjnEolkEjhhiQEy1T+We9YqNli60hYZ2fQ8HwpAqag0c0SFmT7z6UVidftwIHyU13kd0vTyt32W4PlVR30tgoUBx6hSwtwfsXSGV1Hf2gJ1Toln0O6lUoV8GCgApjx4QpR5qrqqumkV/HuhnYH+u0kYoRknsOwbQboBbDJJAeK5gEeFcBE0kn9TU/Rngk59oSIWlFSjmcch8WcSxFKhOIBEMX0rfleJPXapANlJA57yABKR4keglMnwWHLIPQX6v2rCkPNScqOJTq0Pja6e29dMkzboBinaStCtsFjAmhs0F04Fg8eijj+Khhx7CfffdBwC47bbb8OY3vxlf+MIXcM011+TziAhnz55FjBGLxQLDMOC66667oMYVk6POtJSjCtUUSedpSUMHzZd5cKcxUiwFoRI+cYoVb/qgFbohLwAygxmQGOKYi9q0ldMnhZIILkp5wy4V8PHe5dT1eQYkuWeAy0468bGYJDfJ6ZlnqKavqEPJ4D3rxD+xMwPtpIJCu3sCFLtXALunpHL6bE+qoPu+mCKWbOHkmAonp+rpPM4B8xuKcg7GEdT34H4EZiNoGEEzebkZ4BK42QFS9mDQZOH4aHmWgCNXteeAeUigbvhh69C22p4Wf4rMGcyjOoSqYBlkORCellWsjglMjFFXRMoCxVqSa4hDU3nYZhgDalm3gLEqse9x0IFgcebMGVx33XXwXoXZ4ylPeQrOnDlTgcUP//AP40d+5EfwvOc9D+fPn8crXvEKPPvZzz5UY178Bz93yOZfXvpPZz56uZuwMV373o9c7iYcim745PsvdxM2pg/+2QcudxMuCR2bg/P9738/brjhBrzjHe/A2bNnceedd+L9738/brnllo2v8YHn/Aj2P/MX6MHYQcCMYl42amfLVi0rfgv5zClNm+Z1HJgwQOp/DOQwJ8KCJK5/7iRkdz8tv+1TxHlE7Cen5j4H7POIBQL244gFj/idz/4feOZ1z8UijlmrWMRR/BYxYAxF3eWs+RT7WIsmd87DO6lvutP12PMz7Lg+19bcdT12qcMOdThFPfbI4xQ8TpHHlexwiglXMuHKCFwVIp7EAVe6AadmA05dscDuFQO++vc/hC985/8G/3/q4Z60A3rSFaArT4GuOAVccQVw6grQ7ing1JVifuzsgXauAGa7SbtIpojvkOudAkjpuUWzyI7NoRRJHufg/bPA/jlgfhZ87iyQXnzuHHDuPPjxc+Cz+4iP7SN8acT4ReApv/Fh/P/++ktw7uwMZ+c9znKPL5LHl7zD4w44R4yzxDhLEecQ8TiPOMsjzvKA83HAPg8Y4oj9OGAeB5wPCwxhxCKIKRKiKSvZrFxZvuTq6c5JnVPn8/vMdZhRh//8ud/C//ZVL0xFiLpcjGgXHrvkcIo99iCVyXaYsMfATpQsa1KAiLHDET0kNkYD6TzFnEzZu1qrsKSyrnKvUbCL6HAOHfbJ4Xkf/0UAgHOEa6+98hCjuqZVWy0ynT59Gp///OcRgtjpIQQ88sgjOH36dHXeu971LrzsZS+Dcw5XXXUVbr75ZnzsYx87csOUVq0A5eIqawrBEkogy1SE5CZkg6tWLUepDX1YcknldSjvUrU7mSgTHm2NDbGh5ACOnseD2TgpQ/nMWrd0BMIg/ogmTiTXPyV3oJmCGMv/+fsEomMER64ur5OC9c855srZSWxrrdT9CBj/Q/qfVjj7yPzp79pl7lzi8AA+W55Vn/lo8gesl3FgGkTKfY/PLjmw/ddeey1uvPFGPPDAAwCABx54ADfeeGNlggDAU5/6VHzkI6LqLhYL/M7v/A6+7uu+7sgNq3I6as1LozVYav+f6rx2cLVkN+8Qq/A1odk6kFX4zJ8WONbPnfdZa3AkL+88vJPj+lLHpqO63ql646cAQ9u7ilQ+2Oxi4wik0vISOBWCxECkQcwZGEZgmEth43EOXpyXKujpnffPAklrwDjPzsysWXAE22PDAlosWf0Z0JKGIUhZw1StDA1gOFeCkSwpACh/tGC0BraJg7L0o9f+15cz/zvpZ31557J2of8rb13FcSo1ayvZMIWGGrnalA47vm129YtJG5khb3jDG3DXXXfh3nvvxZOe9CTcc889AIA777wTr33ta3HTTTfhH//jf4zXv/71eOlLX4oQAp773Ofie77ne47UqLxKweL1XUU2j0L1e7P1eZOOr9LuEwyj05JlCs3WsN+Q1PCePNil2SZF1ziNrYjimFRnZ3U/1SKcWQXRF5X33iyp2jwP+shTgKHr95wqakdN0KB7NhYBSHVHaRgl9sH7skTqzpt+TKsWxmFJvpdar5MlDBNQLPbFDFmcBxb7wPw8eL6f4ixSNfVULDmHhJvqZcrDGPWZy5KiW/HSoDaJT/ESBOe87F7lrtL8PFw2QXQ5u4raJKqAJYM6+WSKCG86Sn48FKDqNFZGQQsowMYiawfxEGhkt9knU51nTBD7u+PecQpsCBZPf/rTcf/99y8df/vb354/f83XfE1eMTkO0kUkXTeWyLQSygqIBtHGxR+WyLx3kOVVD5TZKu0FiOC88UxV0d51QJTNVQSCT0udgSOCK7axFUZgWSBVGGeuk4Af6vK6fgc7Q1pVO10LJQ5BzJOihYXgEFOFr7gA3ILBiwCaD8BsARCBHUnRYyCbGmSCqth3KThLVkTYOVlGBWrfhS6fJrMFi/0CFPvngf3z4q/Y3wf25+Dzc/D+IIWSFwFxYMQR4JSkwmpFLa+0SDKlQdmTg+eIHg5DCmbTZU0mRnRltWBwrqyCGN5UUZtJu8hgrnEWBsR7J34kAJhRl3klKQREdrSSnMqTLuEX3q0XXt0oloFsAgCsti2KY8mQftwLIycugtOSxD9Mrzm1mZvsMdtxwPJGm/w7TMxOzOgo7TSkknE6kgbblN/vUi8qLY/oyS9FCkbHEmPRODjVxFCTpTcz147rDVjoPgSPGaR6d67QzWX2muy7WHJHAJDt4vMI6gK4GwG3n/uAI4OSaYJhAM92JDCr6yXU2/eA7wQ4VmkVgAAMR2CYC9CMQ3JuzoH9ffD58wIUZ88LWJxfpL0iEbxgxAUhLBS4lI/18+WdmYzMJ91CPhJjp6n84hxVy6G9id5UzUJ5k80KNWOal2oTJZJT2jpL/NlJ4eCy+c8lHhV/RdfIG4AqnB1ACrZKyYDS5IjoQMRSGNyYZZVzE6kqnH6elIoLoxMLFhozH1M0nOeYtQuppl40jFZFs2nbYvM+RZmBSQB7UMqBQDI4ONnJVGxlANhzPXp26FmAote9Bxofkf6XNhXnWI7KNHa1CuGOE4+6h8uzZpcEcZYFkfKMNeW4La4JyppFmDtQFwAXAQygyHCRU8XzAhQ0n0vcxf55cN8X86TrQDmKk5bDvZlr38ewSKaGhHrzfA6cn4PP7wtInF+Azw2I5wLinFPuC8l5AYhWFGKtUiuPHJX6r7LDVmbzAIdIBSw8OdH2WPp5pFDtC1GetJXKS9Rs7e/oE196s08EQAaKWQJz3SXcg2STHyTEmzi1nzm9T8ujmBK1Ng2Wz1bWKzlXszNplReDTiRYZKBIA1M2UTlQSkxCKQxYQQNYttEsylrgsIlCrC3poYFZlCtbgSW8Cxoow8U8AYBT1CGQz5uUhgYk2g1LS6HeMEKYNpOJSuvqbdEg7CQhncFJNW+k3AlcyhmqDiZVwSSpzDiKQI9zlzJZiVfDxQEYI2iMoOTDwM4M6Lu0LyRtJuvSDlTvxQRZFfKtqx1BzBEMA3iQHadIm8jkXUyPeG4A70eE82m7+lxyXAxpi3qMadZs+EpIBX5I9lX0aXDskMs2vk99OyCig8OYTLuWN1azUHIoK1GUfBB2o5/yRPmjcqAbyXoQZlxAQ4sRqWbhUXjVTl9W3onE/A6smhEvBVypjOtvS80VaoL1jodOJFgApuPSoNYNPxHInQjUy2GtM9Oi7CqHj2OGT9fNG3xIqmXnvC/s4InRUwq1TluTr6QeIzhHdQbitMNUZiwRSlSzlyVxXEJsYlCauSiDRa/aRTq2k2YsXbOXGayYU+W5UxXy4OASWOgglBM4J6Zxiwg6FUBDAPVzYNYJYHQ+b1OH96i2p7cCmLSKVLg0b1PnQVY+eDEC80EcmfsjeBER5xFxnxH2CWHuMM5dTsUHAOPoc8q9HDRpeOIZmAGygQpAZAeQDMy5MU0GOAGIhjdW02uDMlWzUJ9Im0agM58B4JTyzWgUypsMGKpVYFkTzN0IyvIeWJ6NwLLvZ0JbsCUjE1tzf102B+flItnXofMlmbwhxl+R91yUzcX59wfYb7ZmpVd/RPZAp4pZaSu45CmQ7FO6wHAldQjMGCiWxDdWKKlOqtImxLECaUGiU6AgWlJrrRDusDCwQ1FpJWTdSbuiA40CEsPcQ0MlujFt2low3CzAzSPczgiaOaBzkvRG370mwymbv6bBIpkhZnmWFwHIWbNM8pv94p+IA2GcS6asMDgMg7R3DA5jKphcp9kXTUo2/IkWqO1RUPdEGBK/BnI5DFxC6NMKSLMkG9GABaaymVHmDwFFw8z+iaTxJR7ZyuuqVWQTxMbv5HtqkqYyQSKZIpPV35qJMObrfZn5LDi/UyoWy7CgUYhWfBbSpcSlnX4GKJhTPkUDGLrxTHNG9IlhgTnXlriCfUrd5sq+Ek5CiSZdG00Vzy1BRSqENl6gA9BnIZS2SNSfApjOWAUipfq4COrITvwSABYLL5m2g8M4Si5O36e0evsCGtQFSavn0wpPR5I9SzNnTaxTc07bzSlOIr2PoqZxkM8cZUWGR/FNhIX4JzSpr02tBwBDdAhwGFGSyYljU1R0zymVQQJ1MqDeg0uaRJR9PmVFa32RIHGi1nE2uhJlYyg0rd4pVk3DpNVLpq0tI7Au5kINRM27YsFrZWBi9dmaILXJfVx0IsGiVFEvU70zoHEY1LSd15JDWr+mpKEQZVVRBhwSWFBKlSZtUs3iFNSuRgaLCGSNohwreTmr+1MRPmqEUMFiSgh7iPDprGX9/wqpsus2Yki1OxZDhxijAMUYMXoH5xldF3PCXtdJ0l4QcvJeN5MOyW4KCxgagZkeNsdI5CmO0nKoA0fJvxmDvGRZ13yOJGChDllTVd3yW5e4tSuFV5SPt6Cuaf4tT9rwbqw4loOqUtp/yyMFeQDYZcqraRYkVAvKjmjmScBo5V2vWzSFaaq0aHMt+e746cSBhU7uiqyx6TjXvNvvpsh2Xit0+u6yRlHyD3jWPIo66I2GkjhxBZe2qa9DQEHvzfkaoQG5HKSDNvdmEcSDZip5n1ZrZQARkPwriyAqvScnSWV030FTCoBIsno7klwYrpPyAFO704EUYjERE5Edb6NoM5EpbSWZLgsgCzMu+4NGdik5cXHWOTA6ThNIWl6UzOfSFyPI9HfhSUAB6k0HkT6ROsGznKCAiNIOaLlkAxdgU83P2Wss+deKvOuzaiuWVruqz3XfsznnuAHjxIFFS+0Dh4T7q9F2mSzqtqaIRjQR22I/MrA7FOZpzghb1+MUN6n6zDl6P2kz55IDSlkYV3C0gwCJnZnsykcGCtQzVdQ1etQZwQZ2CDFlYIpuaXOegoctMqQlA1xKFNmu8QMrgqfMQ6nmkCNKo0aXpliYWCqU6aY/QAd+MT5tv3Wsq1nyvJ5lZUzLJARaBva6VkwJwFpHJRai3Hvq+91YL7+vAokCQKZ/qJQCsKZHXbJxmlrRuRh+CksnGiyO4+FX1bwE6tnCVjHX2UESqhZGMmrn0c6KBlpA0WtqaYH2/rZtdty5BiTqWasWwqoae75nfYcBkgwowKe4kTq9vM0ebYHEVitrSQd8lXW9AZwYZVXGAkR55lJcCEimRwZY+TyaZ9BnjVTAfQrYwfXMqrkkCk+mgaLVXltyK86ZcRPgNwESLY+mTBHtFXu/VTQlehejCpmlEwsWasfp56PSEvrmKD31QgsVFZPzoCUTAKN+Aeby/Y4CSfbGJysyaxjl9yNPA4JNb78EZlxWOqzdq7Mr0ACeeebi75EbDnC5P7PX34aga6Pt96gTx1b34JpHmgbO/k5JUwUsq8yUVjvMuenuAcUpreUc7RKq1QhzIanMj3I9BQ/bN5am4hCc4aud7fP1GuoxbWooSGibV2km2i7RmFU+1/vm1o2Ji+GvAE4wWADrtYIjXa8RDCuAFtctcEy3S2jGRUgj1bMIUMwSQIoCjbQiFgRJ6+BlMFEBbLWJKZU2t43KfhptAQNm1lZarYzn6zLg4OGbzXA6qAksNV7Mg3niCugVFLR9OkjU1Gj7Q88LoIpnFuB96tul56eSXn+KlgbZmp2GdmLQNvmJ83peBoQpTWKVpqCmCJtzDzNBrnqC4wzIAk4oWJT5sPyvtDTguRWWw3WQBQxgWkEtdqSQCmNnZjIroCo03gxUoESlKjEkNZ46VSOAVttfBxSHoVWz1CqhbAXbmWFiVXGCxDPIZ2l8aPc7oIALUMDKLou2NBVXMNWutk22DsfU96smgPZuqgXq9UpVu/T/BKjb6+uxVpZXkfVdXIyJ8TjoRIIFsJm6uO74YWidbTh1ntVGrCYydb5VaSVCtMwgI4kDU2zw4rBrV02m7F9pB1X3nhLMKTBYpbHZvrQYbAW5vo60RZc3c1wASl7I8n+rYlNeYVpFh+HtErg1JqKS1QzW+SgiGr5SXSt3WWvg6jsy3626x9I9jTl7IXQxgAI4YWBxcd0zF07Fmy4U1jS4LXwr1bSL7a0qdB7wpALKyxGSk21J6jqtBqsLJSt0lTmwQphLvAAmYgVKbbHQgIel4pQ8ujRoX9v7A6t9D61fQqkYcfr76TKYU79T2lSzsLQKnDcl+/zHSScKLFqynXaUzrvQDrMCVYS4CHkwA7alohFw3kNgr6XX2eSplrSs1s5P/24KGquE3jd93P7fqvDFEVp/v2oWXXbwLVe234QOWiXTeJep+1rAt0DbrlSVa+nvlnlYt4myn6L9jrEakFo6jJyvMtX1Ol8WPgvgmMyLDfwZ6wTPAoScS9WKxkCYnMGAspzoWT63voYafMqxajYHsmlCaWNMawvbFYkpqitsTx9XWuU0PUjI10FeSdhSzpFKXLWDso1AXKUdKNXmU/2dNXfkWnI8YBmkW+1wyndRXXtCLDMvQctOJ3P/A5dmDwCKw2gzm1zvsHRiwaKlo0DHFNoC0wN8SvjamT/XFkm/nxNVg30pQhMyWAjIuyW1HS1wtWrxEhEh76TL16kBo/g29HNxNJZ6mtPg0VasBwrgrQKDVTsb6w1OnFcRdBB7lA1/xf9TD3DCcj+t4pH9zmpZefWH6u8ENJbbbuXjMAN3pOJT0ntWV6eyMiL8Wu1jWXWP6tyVq3T1vY/bOn1CgMWqGW8dxWZZcN0vrbZgAUKBwNrRI4rwzR3S/gP7+yI4QIo/SMc01+c6n0QrRCVfgfyOzaYxq2lMOdxsdGeHuhJ9G0uhYOFdOW9dCvrSPqrepxKySPvbSt/p98Z/UW1Hz/4crswKawYWc24i0td0sZ0c9NyxYYECe+ajTV3Y9K2eX+4lfAm0/L34rorGUQC+njza65d2bS7zdch/2a99XHQiwWIS9cFLA2kd2aAWCxyMesZq/QdW+OzKhBWyIV3sPJU9H5qfE0gMNzOc3YNl94Is7XfhWtiqlRAIcHjShD3p+ahdti0agjd91qcntMCggVRaT9OGfHtfIjG1pmapilVATBPrtuHc+h2ApazsbRo4G4uRl0ATd0pekxLGbfkUGh4pT4HaZMhAQQXgLSnP7G8zP5SXyhOuwWJInVLzist1Ul/ItXgSMMxtVmp/hyO5xqol6KPQiQQLS+0seWCSU9M5paN56Ts92gKFqqgj6pnLCpqCxT5xzmMxAs210/4Da35AB6rOdOWzSw+rWoIMdsllkMO9YbY7E+U4D1TXRwYKjxIc1VHMmkMLDrqRTPeBZLBwnPeFAMh7PiqwiGWDmG4Wa/eN6MYxzdZt94eEWArjyA5f6a0ue2UkO1q9MiN8GhNQhIZH+rlw3pogsgu1jZPQc4r/QcG9nJPBIgO2fHfeFRDRiNt8DhJwUK1ltD4sCxSt47il5UC8ZbluI5SPg040WOSVENSzpJK1pUuexlqj0N/H/B1l77SeoxqDvFN2hMnW5jSDpWMDSvKb80i5EohRayA1M6kBsKXnYKq+U+1Bt8iXfJPFs94xZDbL8RlFw1DA6CDAAAC9S9XdnOw49T7K9nQnu0u14nreqq5goRKiQGEaz0ml0qQ6nEYb52MptWEsW9EVTOzW9BAcxsAIkbJ/0Od5t8RwBPP8yifV9PSz5qrI/Ehdr7+zO4EV0B2V/1vQbwHC5j1VP9DcmJoBqAE+ATsawJia9FrT8aDVJfu7di+S/u6glaPD0IkGC6DuMFWfp6loD+JpL2v+VpUFjA1LRXuwQDGSzFaqXYykO0dFGIZ0r32KuXAyA4hckt7Y+y1HQxbwyIJBmpMh5X6EJnKhvAW7T8/S629Z1HAHNpGLyfcAAYeOpBW9D+gSKHRdmMxl4foINxOAoBmBHECpkIqWC6geJsIkv5HPHADJmKXHAB4JYSGdnHNajJSzY8UY4TX5TVrD7IkBjohwaRAWnoXEp4UCRQLxgVB4gWXgFi2k5LmQ/pfr5mxadr9MNi1KzpEOAu4dSqasc8Q1qLPITIca2BUwPNfgXvjWatJcfV+eY1lfUJkXKmPhMD6Pg+hEg0WtUsfKKbeKij9A7V1RZVvAsLRuplKQGNKMJKnaElggJenNoCTAUa67rp2UnxEQYFjKmEUlpZ4kFFaBMxWvGBkQVUjKzCc1MwFgZxYkO1YX0fURvheTw/UMP5MkNzQjuN6VauspvV7OktWm1cvey5IhS4sZ5fdRMmV1KWOWZssKC0I3xJxSry1V6CFAEbCET4lfwqfBAMXcaHoKCJbfVcYsc10L+JZvjmuwyMWEKPEocW8fMR0voO6R/DFm/DqrCZqBrKe0QLHKT9fmLmk/twFxx0UnGiyUirOP1y7n5aU6hgGJ5OVeOlcYpjOVCqBuJR+SwA0ZKLgcS1c7xyHPSCNL0t5iK9ez1FQWaUBnLxVGV9LqpeJGNlkv2OUZqqTkK0wsiruYH94xOp80iz6gnwV0O5JOT7UIvwtQT3AzD5o5AYldSdorYKEJewlIpf6k4Q5aBiCv0MQopQWYJWHvYgSCycG5CHAmUXDcJ9Fo5gxHvgYLxwgRaYOaCr9wPaAGijkxFok/GTBYAaD0+6jfZVCVsPRBSwMY3omMFY2iStRrUh0CwDmKKTGvaH2RKC0PA5rCHyiO0anNbiLjnEF+agl7ikqBKatJl2t+WTg41atPkE7TAWC3RFuScam1FshULiuzSNtxVl0NWZ0VoFjojAWduRgDoiTCVTMEIf0v6eUHNPUo8n1Waxial0HrafapqpkWzRlAshWePEARjl2qXyKRoZpvs7fqs+mvDBYJKLrdCL8bC0jsFoCgmReQ2OlBOzMBir6Xd0qVyyaye5Msg4BjzECBEEBjAA8DaBgFOIZRsnvvj6D9AOoY1EVx7DbKSuciQiyJc5UiUdEqDFDMiTFHxIJj4luszArhNefs3oCYjZZnADKQAKj4kkszwGXgGFSz4IiRBNRj0vQ0P6g4ccoy5mi0i6oPYU3t5SXtKYpJm2knSQWNAFq6z4XQiQULoKCtOursct8SERA5ITpLAtfARatoTZBlx1dRb0V7EKBYkAqfzEBzxGxq7HOQKmQoNUO02nYuLpTao5/r5ys1KrSeaUc+VUPzmEEK12RJI5+FyYMwJMdaVHU3+y2QnZldJ+1QoOhORbhdgt9zoN0CErTTAzszUN9J/ZCdmdQNsbVDtGaIlgTInSlmB3EExrEqWoRhABYJMOYL0M4CPFuAZiPo3AByZS6MgeBTgISs1KCKcWCU1aiQTA8FdwWKOSIWiFgkrU+zqotGUoo+Acj1XgBUtUQAVHzJ5SPZlGtIwAEAc0RZCSKAKcWTsAOQNEASn8II0QTUTxahjtBloG/ziEwWiFatIqUj8OqDYwAJKFaFsR+FTixYFI8wloCiTa4CIBVR1h2QKQKPkACjaBU23Vwk5IAfNT8qIUwgMU+z1QKiPSxYWLDPIxapZOGQyhZGlJKFVXk8LDvP5DkNWJjKVznRL/kEFGJDOwI6FnOlS0ISCFXuDADwFNH5Aha+Z/hdAxSnOrhTPbDTF01CQWJvT0BitiPvqyqS2apkqV4IcyyV04cBlKqS8XwO6jtw50GdB/w8DbUBHCN8jOCUwBeQQDBPYtnnkPA0GLJzU/kDxoIj9hEx55DBYkiAoX0fELPWpwWHlEe2GFTLF6116slhhBQuCvDihIVomBGqVRBADrpEPhBShjLlXTJTJiM2NV1B0qaNvAO1OWLlHZRAiHXJV/wlxwkUwAkGCzVBfLJU2wCiljylJTnmtBSWEoqQBv4U21cpaxRktAoqnnExPTgDhQrikMBiziPmccRoa5xCSxjWxWymigxlfwWZIsnkESgimsgfneXUqTYQw7PUx+hZvf7pXBQVVpdDAcDviOlhgYL2ZsDejgDF3i5oNpPShbt7wM4OaLabqpPtCGBksNCVEaNdhBGIERRGYJwD4yi1Tuf7QN+n6mb7BWwkUkvKKI4DeBQ/hvNyTe8j3OhlgHDhVeFRcTarxjfngDmC8IkDFjxWGoOCg/Ip6mcDFFrMSmSq8KQnD0cOgWIGc40JWXAwAV0uTWycYzNGKqUbYnr01imtpCa3Ay/lSW3lHUAu2WJlvo2v+J1nvhrf+MdvW5K/w9KJBIsSX8HF8ZM6jgxYOFKVq5AnIMTSeZTDo6nWKozH2C6zqVkSko9iSL6IATGZG6FoFnHIxZAXccyCp/VOrXbBDVjkEnmkldGTVuGiFMBxSZMAMCSQmJM40QYqpftGFT4qgqfCpcFVAGS1oydxYu52BSj2doHdHdEmMlicAu3uAbM9qaDezYBONIysXbRkK6gPC2BcgIY5qoLKSBZV0kIoxlQ+McANsiLj+5j6pwyQ7NgkzWXKObhKeCPOzAzoHLAfB+ELjxkoWnDIn60GmExYq+31LmaQYKerKQxOkRaDmjYyyaNjknoyBIzJ6SnmqfqYTLdBl/tb+UAFFETT2rTKvCjSCcBYm7NZYuJN6USCBWADlGwi2fpdVFUTVpwHS6nmVK5XZ7sG6mi9ytmZBDGwCqXMXgPHNGMJWAwcMMRxGjBiAozsYGtnkFSRzGgVkSSVjDpYHAjOUVZ/ByYM5DBAKqNJ+8RObkEzh2prMFUHOAUK9U/szApQ7O7WQLFzBTDbBfodUD8DOtUuRNTJAIYM/kG0ijiAfQ8MqaCykwVG1sitEIBZLz6MYQR1ycG6n8oOpOSi3sel2bTyLUEAXR3M82R2KFAseKx4o0ARK0CX6YG5LmeYi2CnCvcMKRzFjhGjVDNjo/ktEACW1SydZMT5LAWPshPayN1ql3eZHFugaPvDJfsj856lJiqzyMTamxyBTiRYUH4v9lqrVdjPSlLyLQ0UUBIE6bi2qKwSI0VpUgEKq2WMKOUIy0tmEvVVKFCMMWCIoQhkjJNahbS/aBaeHLxz6JwHoggrOYJnh4EdOvbwiOi0FB9L7IcdPEBxcuY+s/2l6Ns50KzLjkzq+2WNYucK+TzbE5DoZyAFC+fRFhGhMAA8A8IIHuYgSkWUAfFl9AMQ5Hs1SbgfUhHmDtSNUv3MhC/mdht/BfR5SfhTNAx9j9kkHOKIBY9LGp8FcatVtDxSnqjDmokLkOtzp0kqcEQgl0EiQiYaphIxjNxu4VFMm88OImpAY4q0NqoWC79YdCLBArABKgVlAUx2nP2sa+eycUdUyk0RthJIIHvR5ZVsXpSVDhXAIYbK/Mj/N9qFtL84NqU6NhVBBOAcwVnh1hdp8JeqwGUDW6tVaD8BqCuJ2eAqjZ/o+2xqUNeL6THbLUAx2xPNwiczxJgUmUInJgjpYrU1N0bA9yA/gMmpfZEdpOQI3LkUJYpq/8mq+IIlLVD5wDGD92ABnEPmhQXxEAUspvwVkZoK607AYYgB5JLpqHKQZUNqzAYqgV8WyLOvYtlVkWV9ilYBhW7ws5+tzB9n9CZwQsGihCy3dlyjSdhiNrHsioTukOQV0jZB0bxzxewiSPpuVzj0zw5sZi7CGWsve2QDFHl1PYKYyiyXwSB56tWLT7aGqpo2VLW/7i/z2aEARRqs1MkqBzovgNDP0spHX4ND+1n9D84JIKCs7squsgQSzsn1fFevoKQiy0RUNJAMZsvPMVUXBYk3xYQsfWZ5MXDAGAUsrHbBRrOw/JHdtQQgwjEhRNmzTCCMEKAIHFPVOrvsmlZKyFUgw8SIh5BDG3Q4BRCtzFuakvnD7tZeRxuBxcMPP4y77roLjz32GK6++mrcc889eNrTnrZ03vve9z78i3/xL7KT6L777sNf+2t/7UgNy0tIE/28qhO18w5Sx1aFwR4FhyMvq7BtPMWSumk8/AQGjErKE9dbbufmLbXVwYp6ZgKriMoqh5oYU6sdS9c1gMHJZcdRtAvnwM6ByMxtOUbDXDNfY7ro8sp7b3xmoQzoRttj47OoLs4Akt+LqIC3vY5nJytWKEBlqUw0lH63mjbVANwUYq4gMvc+LtoILF7/+tfjjjvuwO233473vOc9uPvuu/HOd76zOufjH/84/vk//+d4xzvegSc/+cn40pe+hNlsdugG6Vxr/RUXQhuYhStJzHwFIMrx/c6gtzopVzWV0vLoQQCgYdSk16vacbxM1zDtlcQR4CDnyf5z0Q6qU2KlWZRqyGvuqVpH0w7WvSUbkDPvec9G4kkds1L+L7+ltbEHyit9d0TH1veHvQqD0jagWpM4DGAcNx2ooTz66KN46KGHcNtttwEAbrvtNjz00EP4whe+UJ33y7/8y3jlK1+JJz/5yQCAq666Cjs7O0dqlPorgOnBnlc/IlUvWyKPuWRm2kQOjW8thWCrb0Ff9R+aYxoWrJ+9K0FWrnmpQHpy+V2977oyooKuV7XC3xZTXkdVLVLd7CUJJtIA5jKIY0jxEgUost8hjGVpNAzyXRgLmGhgFkdwGMz56ZVCwssrzegNcB2EOQBgC0p7AxAedXBb+yKizBcFeenbab7oy/6+rJQUGSk8Kf8rmNGECeIYpXrdJjy0iYOMvE+dV91nw+tvSgdqFmfOnMF1110H70XV9N7jKU95Cs6cOYNrrrkmn/epT30KT33qU/GKV7wC586dw4te9CK85jWvWZolN6WpB2Um2IQxQEFe21Ftp218z+RRLrs2NQaCUgnBsqEIQCWIDEbvkjqeBN4l+1b9EXZpTp2bdiXEu3K93pWoQUcED5cBTTc3EUrdkOn+Mp9HzjM4SclycAgSYTkG8LAAeS8+iWEBdj6bGlmhjSE7MpfCvVmWT3lYpAjOBCoJYApwhAJUMYpGYXancnRLbbf5GfRd+WNfPXlZYqYggVSuhN8XHlNygrvsu7DU8mUl+ECOqxwogHjUO1Vb3hznAJ6S+c0h6PB0bA7OEAI++clP4r777sNiscCrXvUqXH/99fiO7/iOja/xoj/4ueNqziWh/+/nfvtyN2Fjuva9H7ncTTgU3fDJ91/uJmxM//HPfv1yN+GS0IFgcfr0aXz+859HCAHee4QQ8Mgjj+D06dPVeddffz1uueUWzGYzzGYzvOAFL8CDDz54KLD49ef8COaf+XPsgNFDkrfk3aZN6OuqNeecro0lx2NgwsgOIwgDXNre7DAnidsv78CCGOdJ8iIsKOY9B+dR9htohOCv/9mv4Ruv/5a8li9r+2Ft4I8lq9J6Es1CZkOPGXXo0vuu67FDHXapwynqsAuHPfLYg8MpdriSCaeYcGUErgwRV3DElTTiVDdgb3fA3qkB/8uDv45HXvTN6K4i+Cf1cFfOQFfsgk7tAlfsgU6dkqCsvbRsursnId6zXYne7HdkJcQ5CcpqlyyMj4PHBTAOwLgAFvvA/Dx4/zyQX/vgs+fA+3Pg7HnEs3Pw4wuEL40YH2cMX/L4qt/9j3jo/3wrzu/3ODf2eJw7nCWHc47wJUc45yThzDlE7FPEOQ44h4B9DjjPI/Z5xH4cMOeh4o2G5OtKyDreWC3CGW2vI4+Z69CRxw51+Mj//A3c8tXfhl3qMIPDDnnswmFGDntwmLGkF9hhYIcJM5YauTvMmLHUSe2ZsYOYonJF7n3aG6OV7IHppWSb3zQyENjlvKYjO8zhEED4W3/8djhHuPbaKzcejy0dCBbXXnstbrzxRjzwwAO4/fbb8cADD+DGG2+sTBBAfBkf/vCHcfvtt2McR/zu7/4uXvKSlxy6QercjIDs/yddUkzpz9LWc5tAVklVMQUK5tVqmUvKtVUVZSsyo0vX8IS0/8Ll5Tmbp7FP0YxwgE9byxUwHEd02Czcu92D0LtOdp1Shw4u5bVwxQxCyalAkFR0NkO0CpCkrEvq6UjggcH7AdyNgF8A3oE6CaDKvZS2mtNOEDOi2wHGhcRIKGAAsnICVP4NGP8GxrlsIBsWAhKLed5QhmEE5gvZtr4I4EUUE2QsG8nUB9X6m4oJAmN+OPQpviHAS3+nrEDkCC6WIDcb6m13A1v+ZL8FqDILO/KFR+QxS/zfIS88Isk/YnnUmTaXYMNl0pBvdetz2tcUmYHoQG3cR6LKT6c+OtBFMUc2MkPe8IY34K677sK9996LJz3pSbjnnnsAAHfeeSde+9rX4qabbsK3f/u344//+I9x6623wjmH5z3veXj5y19+pEbp/nzJAM0AOXiOGTACp7j5pkNyhjeTNbq+Zn2stYEdxG/h08afDhIi3qfOjySorZeYUR6uCCTLaR35OuYC05vIAFQbyCjZ3I4cZtRlYdSXCmJPztjpsouxK00SQYNoVCE4hFGeMg6EuGBQH4H9Ea5zgFtIvEVkAQj1J4QgGkI3A/VzCdxyTvwZQK1ZqEcyO0ejbCAbR9Eu0q5TzOfg/X0BifkCvD+A5yN4ERAXEXGR2jgaR16zoqn+Gs1v2VFKPUiEHbhqp24VNu8IHbzEW2B5k5/IjPFrkHFUm5cFig4uTxYzKH/cEmhQaq/PTs8aONZR5HqCbGlVtnT5vMENDkkbgcXTn/503H///UvH3/72t+fPzjm87nWvw+te97oLapB2ie7IcwYwiFmzylWxFJSABaiBQjttnYNdAUIFsSM5v0OanQmAgkRqoCo0O9TBJ5CQKMsCFKu2qtf3NisiKNvTPTnMkkDuJKCYWe0izVidtru5dMz9IIABAGHhJAlvxyAXwW4wP4igGMFjAMYRNJ9L+Hffg3WLeqVVmFBucw3o6sgwlLwWw1BrE+f3wfNBXvsj4n5AnDPighAWDjG1N0bN7FCDuzqhNd9lD0k/0BFhxg4gZEAnELwjDOwxcEDnfQbwdvOY9BtnvrTaheSxECDojCYBALMMFNKO3vCoB+WM356FZwoYLWlBaeKkeTYatc26VRdymp4gN1hYOhSdyAjOCA3bBixgOEqJbJgRmEzAlvUK10AhRXiX0+oBSfCguwxL1awuqYBAYkoCDAUJDQnepS7lN4hpU1PMUZYKGG1SldxiKsugnVV5k4bRk8tgMUuCOCOxf/v00ozfbd9FpHwGZoktDATXOcT9KAlnXBLYmJZRB8lkRcMI7jtgsSjJb3LiG5oO1NIl2LSbVIGCQwAWg2TIGgZgPoAXI/j8QoDi3Ih4XtLrhblDGAhj0oS0RIDdfq+cLishYjL2nLQ+SgOffB5wAwieInp2FV+sViH9ZjQL1HEWypdOE980YLGTgFyAwqFnw5+8ilY0inZXNVBC2HVJWP9XwABkk1gbbDhVyMmaJCoTx0EnDiza+bfVMBwRAiQGHlxHv5VKVaXDrP1mr+0YuYKUVQ8l54DcK5LMXHlJlVxaO5fu3yWPkNTfkABDw8EteJTQ7JpyrEBaiiMAHZXUbZJ7UwRxpmZI2vKssxQZraj0IUm7YsyaxTh6uIFTyHcSxQjZIp62idNCZn/0HTDr5d172Udioj6pAQy2YGFzcCpYhCggMR+ARUDcH8H7EXEeEc4Rwpxy4l41m2SZvIkbSHwQtR6YQTYJiiaFFH0KEEuGKU2AnPmDtHvUgMXkJGL4oxnMlB+qXagpCEDAvOKPJlpOQNFoFaoNroqvsv66KAyFQz1BtkvLNqbIVnk7TjpxYAHUACHOzsYkQZndpxKS2g6T6013nB7JvgoAkYSpLLcEktYB1viIsmV+Fz4LoIKG7gRRwFDBXGWGZPtYZyIjiOqjKMJHmFWaxbJKq32lfRCyZuEwUsmkwCPDjxE0ENxikDR3s1G2i6dcnJhI2CsdZzU5ndZiCvqKKUlvyFqLJutldWYOjLjPCPuEOAhQhCG91GxiLWc4HWORtQtm9HkEIQOGYymrMJDLO1KVL5wd1ev9SQrmyh8Fic7wCwB2DFBkviVfkvqU1K+0zmdhM9A7qB8l+enMBGlpHUgct5PzxIFF7igDEHpcw4McipOzJTbnL9XXnLifgwi8Tyju1exQFE/aDCEVmGHkoKxTmv6OywYv3XFYBBOyg3FCKF02RcosWJUCQLF/BSDEBp6ZWSrbwNVMU4DCp7YOgwCFVA0L8GNEHKUMAM8iaA5QH+FmAeioKgVQbUAD6u3nueNZojI1SnSUgCtehBxwFQcGLxhxBMK+QxwIYXAYFlI/ZBxdNkMi0xLoqwqvan0EBCjUkQXVPiTtYIAsS45gMDmUFAScpSmuWIK3EbyreKMANmOXfRN90iR6iAmiWqBjrSbH2fy1FIH8TA6SJEeBpch6ekYqPgtgtcwft3Zx4sBCSQabTu/WhwEUeKVmVrWfqULaViRECDg7MdUEybe0mgUXv4aDZD8CgF12YOKmElYCCC6AoTtZW9KZKwMGWaFEXh7NzrwkcFpHxHMZQLpaoM8+QpymQ+qUMav34oP0g4PvI+IQJRV/L85P1wVQB1AXQJ5KnokDNnuVYkMCDIjIRYdkWRSpXogDB0JIQDEOsmIzjl7AwmgWAcu+puyITuxRHimou9xvaas4odb4DF+kr1bTOt5oVnYA2MtgguyjsEDRsU40dcUxu4Kl75UWDQWPMklK35Q2Fu3ZmODm2ApL50h0osCi7iz7v36rGYspd8MUs5eRtiYtjOwYGSjy7NQAhtqOnrXgUMl4tAdZrlKfRN46noRU27bONtZ9Jirkql6rw7VLZ3RptlJzqeflZdO2DwKXfSzD6KHFi0Nw8D7CB4IfHJxPWap8KV3oOgAulk2o+l49QD3FcZHSXImMIxBHAYhcjSyVLpRqZOLUlBKGrphNrI5pa4rIIMq+JWAS1LU6WGRKQEFLfOE18mMfswB64Q8Z/gDADMWRqW3TmrQKFF0C9E33gamGYduo71qG09KUNnGcQAGcMLBQaqtrlWMwx1fPclPI2gqFgE7RLpA86c4AhjpBtepXl2YPFbTdpJUoqJWluDIrAMuC2QqjHtNZxwJEBg+dnbA8W3WNWstAvn9IWtAQVb2XNPtaY1SKITvQoi6GTCkVPzmuktIsd3tzgLG21qmChS7rhkAYg0OMTrQJBQvU+45z3zHgIWDfFQVTeGX4MyIVkEp8Ea2Clniyjsg8W/E5UOaTdskO1yBBBizEVOSsEalGu7zcXSbJVtZbDbr10y1rF9PHL5ROHFjoxK7l3dSHUTqgIOi64ivAwZ2mDI9J2JB8FDFnGlL/iWoMSHUfkmMLcl4RwHoWKLVJVjMrO1m5gEgO4DEAASBXtLKzlW+up20gMyMDCTSikxk5ytKkd4xxdAISmrOTDEgQl9QWOT3fwYOMYwoQigKcChohuAQSpYJ6rqTOLjs1pf/ElLJLp1YT7JImqH0SUVZKtIaG+DyKZmFrxWwWFrXMH6tZFLBIx7lUUc/OZwMUm2gVKu/aB/XxzeRdz2+vcaF04sACqAGjHKlBA1BzZPm3Sq0Po6XieU7BLwYwuuTwNHlZKycUIEJS34sq4ACjStXX3jt/biZtnaVcc24VPJaAIh/j+ppAPQONLDtjJYyd4ZgQowDCGOqEsAoc8hn5GICN8ymUlAH2c0kfEKLGBFBZ+eDS5tCYj60mKKtWxpxEDex90grtQKrlYT2t4o/lSQ7O45o3q0BilVZR2lQmyLYNwMHybp+rjl4+HjqRYAEUwACm1LJCbYe2QrAKWdVvIZ8BUA0YuYCtBmfJKUnNFdpJN2tTC1Q2ZjrZ7pqfcsDb5+ga8LDfqyBWQNHoLkXoDFikwaiJjCmt4Eg0rM61CRAagLCfN01oZTc45f9RQpLbAKLWkx/MsdwGFO2iS++WTzoIszaSzskmIePATIsrFkdWAjcgwFSOcw0qbCcBXpoYlNoJ0oLGYWjZl3F8dKLAQqMm8//pfRk0hFZ16CqAaDvOJVVWIuWwBBgAUmFeo12YNqqj05oipQ0H3LxtS/PZbgyzpIJoy95N56fUmVZnVsr2eykcnQZb010ZNEyjN0lLoqCjOSfbyEL5bnqpr50R1y37KRCq9iWFhks+ztwHVMzErPE1fXXQYy35dPXdXKg3MrKsJS7zyGqQ2k7HNqPqKk24bvw6E+M4QULpRIEFUDrOkv2vnUFX0WE6K5sjCTAAWRpjWIbKu58QEjVPDmrDptZjK6DVd0abWDdLtdWuxjxj1QVzpwJ9soA3F1UQWeWDsYaDBQSlNgZAj2mb1x3X9iifHDjzRLNRLWuZhT/ttUqbN6Ml0DAX61f13xpgWsXjVbIOHOx/WCXzcROk34BOHFgAMObB8uMfN2LmZVQYwAAqfVRnsZb0mDVNDqIptvGK45PtXQESU7NU5eDM96L8P61oUV2ot6YpP9Cq86fOXWVj23OmzrNknY56/4ksc5k/q6572CG0ysG4CTjU5y/TQZPkUem4gAI4oWChZB90lVq+yW/196tAqAUMOWa+X8G2TjWLQ/Bj0mSggz3leo/aBl51bgGMfGxipq/aZc4OlRaiv1++ViFa6iPbl1XbNjQRVwm5dTbbM6bAPH+30ULpwbSKzevuDazXFC2tmySPQrYPP3rT38U3/5cLq3d6osHC0oUipP39FFMsYAAHe8uBIjx+WZM/8DeWDhK29pxNhK/tr4NuYVdOit1c/3YdD1r5PmoF7035vOkAnLwHcCg+H3S/C2nLFB3UB5uAyXFqFEpPGLC4GNSqfpX2scHvV81ym9BhBWzVctrGv99AeLQvpsDhoLDoiyGcF4taZ/ITjVq5PW6NZBV9WYMFcDxCvkrlPk66FEK9ri8qE+0StOWwZLWFC7lGSyfxWVu62CCh9IQBiykkvRy0aWDPQZ7341B/L4SOImCXQqs5DG8PXK4+JE05P+01V/th6nMuJm3Kt4sBICceLKYe+kKdnSvP2+AcXvH54DYc7MSc/N0B368SziknrqW2R9RBedDy3GGL7cpKkl1SXU85i9QBk8O65VDrcF7lUD4M6TVW+WHyAhqOX/u6VFrDJnQiweK4O+igWWvVgFz20Nf/j4fUcDYV0qkBOSX0q4TT9h8176uWObOz9oDhvClYKOjo9eqVmfW/Wbqn8SVZYtR9OvX7ts8PK1k0cY1197BRpC1tCiAnCSAsnUiwWEWbjLV13TwldJap68BhKk4gHNEaOkhtbSuqWYGtBjvX6nJ7LTLnd1gGkDb4alX2g3VxFy2tC7c/KCALE22c1BxQgGIpqIuWzz2INgnWWvfcupKkfGrjLlaZNC2ti9q9EDou6DlxYHGQunwQrQp6ste3+z3ac8rGr2VwsDkqABODcFgNw7anaRtQQKiNp9Bt2PlutBygpddQoChA0P6vQLF8DEDZH1IBxQZLdhMcmEr9pufWmbDK3TSkm7CslVigKJ/1Xum8NZPCcpuXqV0xaU0QK6cjlfMqHlENGsBqcL9Qub8UdOLAQmmqs45iL+u11v3SqrPVLsV8zJ5XhHtB04ByWCpRo7QMEEi7LZmNdoG005IqwFCywKDX65AquuXjSZAJ+biOr1Uby6SJ63kwVWd29SYyht1EtrziUm+Ga4EikD1WJoBl4EB1jcOQBWpLZGJygspB4pEeZ6ZJbWNJa5kwG+XeF64T/JXdok5TzsyJyMBNyO7/n5qhlFqgmJqpLEDYzUpDCyaN5rFuW7qdWfIszvV5CgoyS5WMTZLuvgaMOmirAIX2gUfMIOGpgMOqspDtlnUAK7enaxpMAEvVvZcq29sye0iDnhUoOD/9OhG3QBE0uzeWAd7yBeb/w65EOdMgyyP9fkhfa1o/TjtPFTgyn9IFpgDD3OJQZt8qss940GS5KZ0osFBaZVPb79ZREbkCGi1gTNnWFigsSNQzWvntQDV4tKn0yrX1eaaelcp3VMBBhbFoBkizMtDlJLVlxrMeeX134Oxg9AYkvKsBwuekN5oAB0vJcAAcmPiGl4ACuYyiTekXoqyQaE4LF6V0gQzEmPpDC0dx9gkoDxgFKEYs84btZ9QAcpjAb5s9Iu8opZo/gJghDoU/Chou7YTVy6wCd3PKpMzb79dRZW6jlv3joBMJFsCU860+vo6sKbGu06Zs34AaJMIKIQSkmLLNvVlnYmq1DH2eug36PAQRJkcKElQBRpJTYRgjpf8TwJDksfVspECRNQtTbFfBIb97fY+SUi9lyHLEOWMWgOUcnInIcQUUmotTs2RpZqwYCM4xfCp+5KMAh9ZrDaDceW1Mg/VzRKqBYqSaT8U82Yw3U6RmUPEFTfAmXWygwieHlGsDy5qG7b7sm7ERxCtk3n43Rdbcttc/7mDBEwsWQD2zrkLcaSrmxyad1gKFFUIFibE6T64/gI2Q6k5PpHtPaxdYEoiiWbg8a0l26lxUB4SAZncrS1s1+YulMvMVzaJ3Up3bOQEL76MMXB/FHPER3nNO3ruUf5MOoVkwqjycPqTEvcEhhIgYJOV/jAwKUr7RRcCl1H+AmE1RjKXSpwlLQu7fAhQjCkiUsgzLvJnS/Kao0vjW8AZIYMFijnXpHq0mqD4n1S5sykhrLk7JvG3Hip6v/ltndl8InSiwePav/RN89K/fuXR8lcd+FZXZmzNgrKIq2QrRkhCK8CGn+x9RZqk5cRZcFVBN+V/U4NV3J1AtJKk+pg52T0U4Owaq7OPyeMXcojohDiWg8FQ0i85HdF3IIOG8JOetPs+imDed1kVNbU1qzirtQrQJGR2cXoiEOCbgCCR1SgIhDAJKYXRi6oxO+jQ4wKkZMg3ykQqflEeDfoaWZQBGTPOm9Vu0/CkgUfPFankdFBjk3DlYijRDTC2f2qClAUYIYFDSBKXq3fIGwnWT4zopttrqcZselk4UWCi1DrrKe78CbadUVpfwVT3t1v4FjC1r7WAsz1bBgERA0Sz2kWpToAhkNOKnNSqAYoaUdgkRkIsYaW0KB86FdQmSkYuT2aHZnxRQZOBMOTjFJ6FgoUDRdRG+E3DoulCVAMi1Q2YAdVInhDzljiZHq6c4NT1SJWoODI6ASwWGeBTgiIOT+w0kWssgjtNxTA+Q6oZ40sLSZQjlxLvKl8QjBYqBCkgoz2r+lNohIg3TlAes4UsxPwpvdDlVfVcdTxSrgjFL0vkHTXerlrhXUwEHrTfSAsZxQMeJA4v2oRQopp2dpQNFRa81ClvFadX1s3qqforG9i1CJ2bHiAIWcy2CjLoqmU39b8Fj6lkdKDk1S80QT1KNO4Dy9ncG8rKDCu8I5LygthSB7TdPMooVKLo+oOsjfJ+0iT4BxYxBHeBmBOqluBB1DnDlXS68QuxSJTKtSMZjhBYc4lGKDbkFIy4iwiIBUfKLaAEkvYzy1uUBUPhozYkRMlCHBBQBXHiUeDOCERJojyj1TW1VspYnMHwBNNt6qkrX8AaQ+4vmIsvBXeaVXNAlPq0yG2t5WJ4cV8m8tH9Z5m1Brqkl6KPSiQMLS2Kx8jTSTjx7roM6oVGsS2SqApjf1X8BVEChgjik2XqBiMBJIBU0UGsTbQWs0j673KblEVPVbi5l8npKqej05woYkBqYnQWTfP3Sf3mMO4bvIro+optF+FmsNIkMEjNXly7sXKl1CpSqZE0JQwUJyrVOE1gsgnweGdFHMW06RlxQ9oeoLyRGQufLaoh9lswnEgkI2X/BGdQFNJRPjDHxhlF4pLywWl9LVrOwvOlAIK2QnkADSGaIaV8EMEv5XV0yQ4qjHSnRMFWh9S0g1D4Mc14j963MI99nuZrbhdKJBAtru1nKy30r2Cy2Yqq2ngHDqmatuUJ1wA4VgFDV1moUChRDusqcYwaJAMaY/tfZyxbeXVkYGWJ6aPFdLYwcEmhIG5NNr7MulfKFI2lpw9b2ltUO51SziGKC9AIUfidpEzOC6wk0I9BuVwojew+adYB3Agydz/dGU0UdbRX1EEFRAINmYwKLCFoExP2QfCEMcjGDjzhESzZw7xgUpvmc+UTyUiCfJ6AYEh+GhjcCLnEtb7I2Q5R9FVO86VN5RSSZEPNGkziLkxwooK5aYKSysU7BxecBzwboCx+nomlzezUYjXkJIBxqs/tC6USChZJ69LMpQjXqTpEuWVnACMnWXNVx7VJpNj+SDTwCZtaKGLloFkPSLgJiJZS5WjfzSkOEMlhISztyGDnCw+WK3VmrICfPn+qchvQ8WsckEnJ0JExf+TRr+z6i2wnodhJQ7DLcDsHNnGgTCSjQdwISfQfqO1NFPZkizmEp1TcnjWIM0ogxgEMAjQEYRjk+jEDn4BxJsWSKUD2fI4NjRAgEGjUoKz0DFxPEgnnIfWBBXIBCwCMm7cIAOkfErGGs5o3ypbwLbwJzBo2YNA2VAwblNW6XZNAlMCcuy98qo1MJnvNzozip7eQ4LfdJg6blSbKOFLlwOrFgUaWhb4DCyqojzqnnAeRAGNt5MnxXAEV6t7ZwBo70aoFCNYsBEQsOlVYRrHbBqmWkGUxnkJyNS2YvTy75IFwCCvltr0sQiCm2wqU+QKrknZyfXAADDGjhP10iBcRn4XuugWK3gATtdKDdWQGJWQ/0nWgYXScA4VwBi7aSOjMwjgAzeBxBwwiEACyGBBQe8GMCHIITlzEQGW6McCPB+9LeepDU/GKSZ82OTdRAoSA+NDwJiPlzy5ta9tSJLLwRbYIxIqInj8CMSEUPGDiWZSICBqMBuhTSnmN0+OAEz6smx1VbkFqZV1n/Kx9nsQQS1ee6KI4e96TRfmkN23ReYLmmrtdbDcMifaTkq6BlOzg7yxJQDOmXCw5JQCNGDgIwHLNQRrDsF4CqnYnpXMDCk5MlU3Ky1ImICJ81E2mqy8/viRFYK3ZJTc8+aRWaIt/lviqVxbTgsfgoDFDsdqCdHrTbAzszAYqdGajvgb4XrWIJLIzDE5ARwFGAIwTQmIBiGMB9BxpG8HwB8vXvXBzBIcKNDB8iwkBwPmkWVJ4j3wZ21YozkAcSn4QFCuVNAGNIgK68iUkTtLxZlkMSc49d4k0BjT5JlP52QHp+Uq1I/RzJ/KBSyFmfAVhePlVNetXkaOXeUivzMicVOT8u0NgILB5++GHcddddeOyxx3D11VfjnnvuwdOe9rTJc//0T/8U3/md34k77rgDP/7jP36kRrUAWipkIe9lsMcBO2vXnVdcidPUetezLYzaqz5wzAKZNQsDFAPHJJQCFJyEcxIsVL0lgmN5H9mhI4+evFzd6q1Jw+2YMBCbma5UdM99ld69U59F6hfP8LO0LNqTaBMKFHsz0M4MSK8MFH0vQOE94DsDFK72W8QEFByBoGbHAuh70HwO9oP4OlTqkzMUowAFd4yoy7ejzqJpkDTjo1qtQlmdKkAel0B8MACugKF8mQIMy5tADgEug0aO1zAb+wb1iHEEkUvlIUUr9UTpHiWGZ2q5u5XQqclxymfBINGuk8zLjlerXRyf32IjsHj961+PO+64A7fffjve85734O6778Y73/nOpfNCCHj961+PF77whcfSOEXZ1o4DLIBYb3By9CTA0L51WN5iHKkE+LTedUaxidVfEVCEceBihowcknAGBJb/7Qy2yjamZGeqZuHJZW2iV4uW5LkCl9nTp5k0pnqf4OVZQ9XYJc2i43rFo/eg3b4Gip0dYGfHgMUM6DqQb7QLQIBDY7uTVsFhFHOk74FhAJwTM8YlDrKUTkQIyekZxcm6EEDTgeHSILGxCWw+a3RtoALgCuoDRywgIDFywKh8sYCRAP0g3nh2iOQQyCW9jwF0FZiLGQI4uMwrn5yc6lMKVKI719kgGsnbbvTTPmnJTpKBC1CINlGDxoXSgWDx6KOP4qGHHsJ9990HALjtttvw5je/GV/4whdwzTXXVOe+7W1vw7d8y7fg3LlzOHfu3AU1bCmE2SCs3RFZvi+goXabA4OTKTIVcwEUE8S+j+AsiKK+luU40TBkgFigGDhkoND3jcEChN55MDPYhFQ7GHUYhAERXQKJoKbVxOybhcssS4oJAomfSCsetFtAAjsz0N4eMJslsJgB/QzUqXbRF62CJjSLZIZQGESrCAHsF8qsfB6xODNpp0+rJUEAo0uahfos3LQbMJuMMGHdXPNGAX2c4M2hwSKZIF2qPtLlNnVZnEZEOFAG8zGZixqMZYPJgClwn9YcWqBoZZ5ZVlpybAqJ/0JlvjzP8dCBYHHmzBlcd9118F46yXuPpzzlKThz5kwFFp/4xCfw0Y9+FO985ztx7733Hqkx1157JW77/L850m8vB33wzz5wuZuwMX3lhz98uZtwKHrWf3/gcjdhY/rfP/3ey92ES0LH4uAchgE/+ZM/iX/yT/5JBpWj0KOPPo6P/fVXpd2SQA/Z/OQR87Zq3SkJTAWoYClnQmDJlzDAIYAwgDCQw5wIA8G8AwtizEmCbOZg7FPEgiPmiFggYp9HzJM28et/9mv4puu/FQsek6obMMSJGSxOx3Dq7EVE6MnDkUPvPDrymLkOM+rKO3nsUodT1GEXDnvkcYo9ToFwBROuiIQrGLgyRFzBEVfSiFPdgL3dATt7I57+8Q/gkRd9M7qrCP7KDnTlDO6KHeCKPdCpPdDeLnDqFLC7C8x2QTu7wGwH6HfEDPEd0PWA70BZu0h85mTgxSgmiGoW4wIY5sBiDt4/B8znwP55YH8ffH4ffO48cPY84tk54hfniGcDxi8Ciy95PPX3PoT/8vRvx/n9HudDh8e5w1lyeNw7PO6Ac8R4nBjnKOIcIs5xwD4CzvOIfR6x4ID9OIhWEceVmoXlDTOLXwUQbc9qFpT8SYk/6luaUYcP/88P4tavvhW75DEjjxkcduCwSx47IMzYYQeEHQZ2Wd57BnaYscOMnoGeI/qklWSZTzuEp/KKWLKV6gNLgiGV+QDZ+j9CNiJ+40P/T1x77ZVHHp8HgsXp06fx+c9/HiEEeO8RQsAjjzyC06dP53P+/M//HJ/+9Kfx6le/GgDwxS9+EcyMxx9/HG9+85uP1DAyqlnxBpfvW59FVsl0YE5kbFpF1ste4i3E8aWRfyNH44WXe4hfQgTPAsUYw6RAxhSqLc8lYOHIAS4lto0SITlyKMuqCeQCRMUeSXwYTCygqI+6wiTNzjGXoi81KrPzIO9lSVNXPXwH6mcCFLO9ZIrsyHHfAa5fa4ZQGIHQgZ0v33MExV1wTM7PtEpC3oM7iRSlzoF8BHWy41X5uiorl1XnY3JQBk7L1+poNmbIEpBH4duSCZKWS5kYjpxsCswroskZnZzTAYSQQuk1dkP9FSonMa1YacwDozZB2uLVU3xTmpogI9fyX/wVYorIkjonn92FGyMHgsW1116LG2+8EQ888ABuv/12PPDAA7jxxhsrE+T666/Hxz72sfz/z/3cz+HcuXNHXg1RsjZcBg/iJaDQzwoYgaVTGSwJV7A+KAuwXnYbgZki/bIw6LcJLJJw2llLgULfI9ced7Ulsy+C0vWcHBs5iBceEY4jAgm7SzBRDWyr+i0nr3EFLKCA4VOglQKGrnb0swQcfQYK6mZAJ4AB34MULAABBA33DqMiUhkCukoSIxBmAhbDAOp7cD/kdsjeE+Twb7l0yquxYjDZZUjOfEO1RGq1iArIY1xaqbI80Shg71yO0W7BgogQknM3cMQIWf7OMpNWSypgWwPqcn+u36mA5pRz0/oryjFGOMREeRhataeloje84Q1417vehZe85CV417vehTe+8Y0AgDvvvBMf//jHL0rDSgOLh7wFB323r/Y8YNqLDBRBs32rAzEzHA1oJOEDIFpFI5SVZhGL1hFixBiSoMZldXjpmNFaQgYKaUOESeqCNaBhZ2cFCn1pZGYFHF0CBXlloOhnoH5HtI4uaRuzvfTdTtZIqE+OUQWXflaupdfNsRvpnnkPCoF0m+0ED5XsM9v+yJ+55tMSX1b1fZw4J6aYDDa8MEuvGSwyiBdttPyvy6VJvmh5L89BZCfHVtanAIVoepn1Qmkjn8XTn/503H///UvH3/72t0+e/yM/8iMX1iosL/VMJY09KHnsqhWQVdTmb9QZy5ogqn0ARbOIzLUwGaBQzcLaxvJOaZ9Fer5kZxJTuS4VoddcDBbMqrY3z5IFx6xywkEGptN7m6VQjaNwPmsRqmGQT0uo6TtqklqwBmSpVhEjiHfE9OiTRuEX5R4pfJy8B+cwcnnZBDtLPqkpnjV9YuNblv5nrvjDjRmSeQIJuee0TV41PUqahSyn1vEzGcBTzgr9f1W7V1G7D6SV9crsSNr0Kjqs/B9EJy6CEyjqTissUyaIMzsWM3ikjpR3bATjU4xV1VZnCSt8+n2eecBZIHV2UxBRwKifRfwRKrCBRRg9u/Q7zoCRd0kaAS2AsUZY7Fc6IKXTDEgkcCDz2QZgkQEP51K8hQELjqAIsAOA3kRyjuUavpPl1yWgUICgvKEsg9sGMl54VpaoS/9w7kc2/cnMFVDEii+6t0KiFCj5AQJHMU/UV6S8NpNVkQ4bGZzOSQN6FWg487LPvio8QD8rYEgsTjreyMQlD8q6lLQUybZGpbLZpp3jnFla/RZHJTsolVTwrEBaAGGuhdMCRzuDpR/nSEAizqv3sb5qFugcNYhaSFs7uNXIrM8iD0p95SAro2Ek5yWlQU72mAKFsyteHqAAYlHAQQ7kHNh3oNglzaGJ/LQvKiC2KgvXJI8MD+wxdVpW/YfS/xnUwZMgLlG/SYszgCBp/o0MWB6oNpHe9Tqbkj13StZXOXutZrFOyzhEt66lEwcWq+ggk+M4qL2DNTnKObWQ2G3OVjitp701QXSJLnJx4FnBZXNdTft/bBuCDEiQXdHQz2qK2ChN5wU8FCgazQLOy0oOOVkJsd8bYCFaE0e4BjDa/SFTtVkiT1+57VP9fwrEdY8mg0GQkGzS31R+rTJBVPfP5sfxyqoFhPa4frfOHAGOBzBOLFhoyHJLNmX9paIlwODp/9vjq75T4cvC2QgjEZUlVqOPWxPkSAI56VJvwrdzI/z08ZZsyLc918ZiLGkjzf3XNXmD51wFFEotD5a0PD0PnMPnCQUkKo1xKtbhmMHBrVj9WAUI7YrIxaLj0lC2hHpgX0o6stZhBmrrtDwy2b0iGrDFEYgrapDH9a1vM5wdtZUWUI43y0NNbamHo1DccOlzCjz4oj7bCadV23LX0RTSbvKgU92s2ayq82j9/w61ZkCg6hw91v5+6brpLxetqRxXlxCY+HiMILbXOQAkpmgVD53pN7cCsFcdr/gAWnvupiQyczSydWGr4w0w6P9t1PLFpBNrhhxEMVJlirRl86bosAycOr+qUqXmAqeAHhhg4BTIQ7LRCTCgkM7RcG8FF5uhST8fx3NkaqUpFi2AuZnDOaBsEAvyfRinGxADwDEtocrnvCoSzXX0ntFqH1y1bQqXbLkGALkymCVNgwcUkAUSME8At6agkwJA08Ctny14TGklqzQVjcI9bjrIP3Gx6ESChW5LB0Ql83aJaoWzR787jo4kpqXZRZO3KjBIO2uAUPBgyL4COCDEmIXUOjgVHLwr+xAyeBifhWo2JTx8+fns2F1KTNyCqExF9YAtJ8v/4yBxFWEE+x7EMX3u5HM7oFn3hYzyOQFCjr/IwBHl/lHPqYHiMEQo+THrvqgHdpU7hFIZxfTuJhzHGbQboHAWdOykYH63Uqtp3i+UDoqvuFh0IsFiinRN2ZJqF1NaRduZ6yIdgWmGav5FvVLe/GUyXelW5pjW4h25tBxaBMrnzEpF+BQocrYsogwaKoz5+ITmoQNlnZ/X+v1ksmdwZFBO2y+JdSURzVgGdhr85HvZGGZ8GwyUzWQa6m21CN1MZuMtovnMSZtQwEptklfRLA7wWTa8MwOcC8DqANZ+DYgVX8ApOU2zIqK8yJxoNEBnvgMKwCDJCmG9mbjqm4iUci9pTlMyfxQ6ruCsJwxYrCIFCgsOeTkJdDihS31qA2SyicD1DA8kVZXLhjCfYiC8cyL8LEuUIcYsmPq7nC/BuSp3goLEtImigjjR9rZfND4tmgP5pTN+GrQ6SoMO9jTgg8RVsKtFn8mVbEJ6gzCIxhAG83v7OZkjCkpjEE0raxvpZe60tO/hgOe2M74mFbKaRcsXXU5ttZMM4ok31TUUzFXTBMr3qLWLSo40aCoHTzX8Qp29Q4tCM1P50YakMn/czs4nBFjYjos5yq6YInat+TAarc4CZGSUmu+B4rDSgar31wHOkCAdRgmeUjOEmAFXL52qoHnnkoC5pZnL114Qo1HU2oUVyKV+swCaN1XILA5myb49psxWYRRtQkFC/3cOCKbMT4zLy51RtZEEDnEAjwsDPKMxU7gAR4y53giyUlODfyvw9plLvxS+OJIo2NEAcDBArjtJiSUPp6a9U74ov5U3CjItmKvWh9wW04askRZtA8BaLXAdRZ5e9V5HF2NV5ESDhWgGutYtdUGq79fYbRY0NgEQW+g2CyJRqhZWBNMnwUM6pmARyOwncGxsnggkM0TJahZWGJ3Jn6Dg0aXM36raFjV3DUgYQcl9lMyQkv8yJE0ivSxw+DSwh3kFEqxAQT4HdLEukWaTYwAPC/F7DIuiVej29HEEpyzgcs9QAEw3ekA0xnX8LYCRBrcEPUuhpgQY1kzMQG7yWWo2bEbJZQGg5g0t8yaDRuKAZGcvblUH4+sw7V1+hvWC2cr8QX4K2S178XwZJxYsVJuojuUY+1q7mFpGkv9pyeF3EJWZimuQMOaIN5qFRvL5FKcQKVUul+0FICIxQya8945qE2Ry9kqfO3J5YOhSqrR3Rf8xyYDTmTppFjxGkAEIDgE0DLIbNIzgcSgh3sk3wQCoTxfRYxqXYZ2YYRANYphL8ptWq9Cs3woUyQyS6mW1ZrGeR+VdgEF41pGTcgpsBjpcpfXBISWHiWklpA6IKw5N4U3n/BJ/nOEJUCaWUuawgIbKT9v2VSQ5KkjKUpr9H1PaxaVcNgVOMFgoRUjHSUeZpdJ0bBXatmrYOtCo1chyzEEKyQwoQqkVqQBkxyUT51nG7hlwToSS3LITzTreJl8oQKHiZ7UM4mJprzJpmctMwxHgkBybYwTZQkApIc3KDN66srEqU5Z+H0bRKMZFypY1l+sOi1qTSffGGEqZw9A6OKc1CzUZcx9QAY28K5QEMAK5vENUTcV8nWR+lCLJxXdxEG9c4o9OEN5OJpDq6jbWwmpB+v8mZM3vwzg6L0YFdeCEg0WrhkW7bJnQ1lKLsuscPWqFr3KaieOSM/NDUjs7KpmyOvVXoCRyrTYYoexWnEo3P5W+TYVQZy4VRJ9E2RtBrDzyXAMpG5AAAB5JBuMYQWPMgJFre4SUvr/z4GFR9o3oNnONt2j3fuhNxrQCMiSNIgEFDwk4hkX6fxAAUaBIbeG06aPVLCyPHTNgnL0uaZhk+khfHUkKf3YeMaY9IEmrIKa8MWxVAhzR4JJ2QX6SN9r3tq5IZ7UKPlxwVkSpf6orIlbmI7hKeNMCauTi0L8YmsaJBAspCFR2YjKnjlR1zAz+1sk5dS2lg7ULghY2sn6LshJSZhEAOX1/3kSWnJmOKeenCBzTMljNPV3CW6VVlFmMKtAoQFG/WtJwCgsauaK5miLDKFXHhhHoBuRdqEC9ApK0iryVfYWDMzs0h7lk957vJy1jKC/1UyQtI5sgIyOORdg3WcXSZ++WQEI0Pc8C5n3SgqzDOCBmE6R1PtulauVDTx6dK/k3tVIZkkwUn1K9epU1wRUTU0slwognY4zyjucKJKYnxGPbfJjoRIKFJeu74DSYFTDa4BTVKjZ19LR2v4M4OrNqi1I9OyRBZAa6xMAOBSx0OrERe8V6xdIGpGznGr+EJu71ZqbS97aid7ly21+Fap8FJROEZTZfBKAXwKBukHyYChbGeUkxisOzLQeA6kbZZ6F+Cc5JewtQcH4fwYvRaBa6GlJrFqvC9i2YF/8AmVqxJLwhnzU/W4xJHNdUTBBrntjBbvwenfOJF4U3CkKicRbNQt6NecvTslZ1odEq1G/hk3GkZThXyXt7nU0nyMPSiQYLBYpSYUyi7ixgrKMIY5JkVW55Ka59z4OSJXGv1Shi8l0Aoln0yDeoHGWElPg1O1sbM8QKJIot3JuZqyNfAYYHwZPRLLgWRNtvspJkZupAiKOcr6aI1iDlzoP8ACaqzQ/rqwgh1SpdARa2yFAYs9mBYQDm8wQUI3gY5b7DCCxEs+DAiKP0YYkKX6MFGkDvkOq+sgJpMRUlW7YM6AFSF9YlrVX50/LGrdAs5LpeQCEfK0uj1lRUE6QsdRfZWkerNGoxPaad+vl3htd67LjpRIKFbBGWJLviyEx1SicAo6WpjjtIHXNFMZA8MJCOCdDlUbleAKUZoDi2mFy5fna+URZKn7SPSc0CVHntVbvwxuzoyclgSLOW7jdw+Tr1+1J/VGYIIS4Y1HOe1bEYpQZpymDF7bUiA52EeqsDdGVavbzqEXJy3qxRzBdSJHkxZK2Ck1YhL0IcCTG43O5Sj85GRprZmpCD4rIpYhzRyh+lkPqbzcBTrbBycKL2WxSw0BKTDp1xcKpW4YDMK5UhC+wqa+u0ChGjkgk8+zBUw6Bao5g0P1pf3jEBx4kDi0qNhmoP1m4raAtoKKv+tla/DttJalvqL5XhDEozlYBFb4SEm9GlQTkhgQWvcKIBqmq6/K6zVodkI084zlQQtX0AJoN9ojp70wwdgwxGNzJ4YPAi5GS5GEZwN+THYCCFgUegC0DwgA9AJ5mwGKhWSgCkGAuu4imK6THWWkUIRasYGTyq/9Ss3qzK+pT62/aBzuitI5rzwyh/lbsp1b/6mVo5MFGbrR+pN8BhfRZTPiWClanNAjGjmSQ9WXkvJolS2aFK+byLuUX9xIGFkvVVVNpFPlY6UKkIWn2dVXabVh0no9rn40TwAMak8qp2YZPTd3DIzcx+jrLW3qZ0s2T3HlhTRG3hMnsZL39qZyuU+ZraD7k/TGzKSKLmj1Q5OrEIssLg6ihNZlOT1HsBjaH2aZSON+HjGksRggDFQs2P9L4YwQoUiygrr6OYSeKzKG1f5XcqfgvhkVSXNxoFETp2pUO4BgCbDbzlTRXFaWIqPIoGmHljwKJTrQYlFseh7N1pl09bKpqEPdZq1LwEBq0WnY/j+JdQTyRYZHRNI9G337EUHpoSJouwVRQnaq3FsRSrbaldZejS7xgEJg30EurJFdNDBTJtKPPkciJf0LTPogaLWhizGQJnnGYyIAQwiofdahm2HyqfBSfNok8z+SjaBXJOzjHFgySBZtloJnU9OgGAnDNzysHJxW+hgDGGAhTzZH4MITtYWxNE9pgVzUJ4WfNG39VMdQA8EzpKTuis1VGZKEiMQQUYCxYtb1q+5GVZdWA2JiJQazYdGvNjgjeOeW3o93qNukyW9W+KzF8s7eJEgkVLxV5LqyEJMDb63QHnZF+FvtQWRlEhPcMUISrXVLMEyZRwHJOwxbQ/JOVMaISyqkqm92iEsWtAw9rCuZ1YrdpG1GZICA4+RMSRQCODBpYizC7kJL5ifujO1E4Gv/cSQOVTiPdUGjwNBVfAGE3wlZogi2SCWKAIAI9IQCH+ihiWB4HloX4rIFF8TNZf0S9pfJzB3EP2hGS+2IvmexS+ZK3PahgJKFSz6NP/hBIYZn0syqtVy9yW1mnUChj5XC59tMSSdOw4gePEgkWrkinl5VOousjmN6WDrFaxSYflPSDJQdkxS96DJFIqkJHKbO1BADlpA0cJ4kmCyYQyg/GyUNaAQSuFMccPoJggGTSMilv3ndqytYMzjiTqvpoiHYOMhmHGVyo3mCqWBSlExDmpb9OfupPVgIVGaCLEDBS8COlVOzbBxV+R63bG2ny0mEg6+Ej7Q2q6FK0i9UHS/EhNzaTpEZWI2qn8nXU6gOLApIYvvTq6swMa5qX7ioq/wk1jU8U31ajbfSN6rG1uK/ObTJBHpRMHFhYkyoOL9zp7gxNgyPl117edVrIsLbPIzlIBxvRIg10dimmzJqA2Y/qhCEtiDTmEtNdAhTGXDSCgFcmkjwDQjUhIfguqhFFnLQ+ZNa2KO+XkVEFRgMw+i+AQAyOODPIEWqQfkajncFrgOA34zoNCBPvk4ExJetACRk5gE8uW87TvgxejAYxQXupkHcVfIasg8gq6GoKyGmKpffYcD8OESGoaxnQ0ZjB3SXYigM7wZsotoryxZogz2p01D4FifqjPRLWK1hyBeSfDsylSU6TVqKeonRztcXu9C6UTBxbA8sAqx2vA2JRi825JgcIKX166ZRGmrFVwHX7epV+J7SxlFmMSTLtcWgaw/F+W6MwSKpA3i/UwAT4qoI0g2gjBqTiL0l/pWHIYhoFS/RAGjQA7Rkz2PCKDZl7MEE47QR2BQhQz5QCwSDcBp52kOUJztGDBiAoUC0IcHMLgEJIJUsKV12+O0r6j5NPSmJpIMNpfzRtidSSyqfDWODgneOPJaH9G09P8JLVTs9YqPNcAp9c+DKk5Iu2bWg0p/+sEedxAAZxQsACK2qWfkZxZQOmcqT3+rVaxmQki6B2AvGlNnZui6cj9GSgqB1IyGyNrhKShkDjRyialaSJYhxoqgewgTrtlpxmh42mQaPuhUuuDgwuMSKJhuECIC2mdqL6xXE+3sXcR1DkZ/KkYEU11OmC2vwto2H0fsDEVQcyPuEhxFY1WkTWLCe3QGd5YLTBQcXbqUjdq/6D0d1r2LXVd0uOaE0tCm8Ib68CkBsQBVBpfOdeAhF0ROcDXNrUikoMQJybJdfEWx00nFiyAoooVk8Myc3VQ1lHJZYbI50jqu6hBQ3XXHiQ+DramjHrbKW8qa4WyEkgDFNYn4W2gke6mBNJSofFX8LQAapawYJLJxECIXgZoIAdQFLPeMZAAgyKDZiwBWxqHYcoM8qosLDknhWoXsbyrfyIw4oIRF8n8GApIcNSI02mz0f7XBtGJZpF+k7QL/VH2b6TvQ5IQSadX82Udb7zlj+ENkKJIQZk3HatmaoDN+Ct0yb7qPqwyv2vAmKISnVxPkF92e0OUivMHuQOBFWoZzFJS/rxM4jASW1fNDwek2Hy9j3jdQZX7DwCysIDSYGctkKsxFtoevXtrfxuvO4pwtkBRnGfLTjOauHIxu4rghEAgKtoFuQgayq8ocnLUIvksSDpijBks8muSQQ1YRBaQiGxWPhJQJPNDNYoYZNUmhOXl7iWecUlUk80xZnTGGe2N9qfsciCMafC1fCmytJ433mxBV94ABShkabs2OxSsrH8p9/l0T9bdioPlXc+Lzf/HTSceLKx20QIGMKGWYdqZqbReCSzkkJamkkmBJcAoQkk6Y1Hxq2icf9FIABtSrPewM5m1jdVp1yegmFJvrbOspQhtS9EsVLsIpH4L05YoLaTIoE7Ag0YS0Eg34oPMECCBRYrjiAUkEJM2oUCRTBD1VYTg0lJvcXDKcxhAa/hjndHazzojw/DGpb0ilE2bZb7I9Zd5o+bIkoaRTI+lczGtVagJss505OpzWTLVgW9BQ8lODPZ3F4NOLFiIbVn8FnbJ1Hbg6t+vx+1W8Dj5LdiYItoQn0BCPOlFs+gA2YqulayRnKNU2jAVnqv3dmaGtDOZ9axbFVfVWzt7TS2b5ngQY5eHSKCkXRABoTgo6n6JDBcZ7CCgEUR9EZDgVC0dFWgUoEifOS1IxLLigUg5UjOO4mgdR5+dmyGQLJdWmLPMw2x+ADn+pSx1qyovbVXeIPVtJGMiVv6kCc2CCnRYE1F8EIUvaPgzFY4/9QyraMk/B+u7W/ebVtM4fh/GiQOLtrPkWEHVVTEW5bfL5kj7vaUWNKwpokttTs9kxkgKGGW2h9rQZE2gxGBqTRG9lz5DvbKhm8QUKPLg4DKTqtOsDvaZ1plyn5i9IjEBB4xIQp8jAjwClGIwKN2ATJJSCV+o72eziGsKvwokIlVAoQFY0fgqdPOYNGVa2KvZmmrtwpqLYnYU3jABo33iDfiCNbyRdyEFCgtinfIIhYeHGb6tRi1tWubxlMxfLDpxYNHSlK9iVYwFMN1pjIM7scwiyNpFznuYLQ8BjBzuncwPl94jq9oqP1CnlVV5Ye5j31UQK+0iq7yoTJF2KW6KSoxCmkFZZm4iUflLmrak9ic/g+tYllUlRCGBRQINsMY51dOmYk1UjYLyZwWJvFEs7SwdR2dWQKRtNg1geQbbZ4wAyn6L4ri0zmck01Gu0PKmmCzr+aKf7Q7fSrtAmTQUKLwxP7Jj1ZggACrH9CpqNWoAFWi01Mp8AZDDAdRBdGLBQsGBsJlzE1jdaavIJTW2vDMilQEp1zQfFDCyg1P+pySUaULNexqKQK64f/Ous5h41ZeBYkkQgWolpG2z+i2A5LTNPoGYtplbazjm+j9EEofhOjVBUvtc+TxFFiDEFCkgIeHcxUdhgSJG8VcEmxZuVZ+x7OlRUFeNTv1LChhE2qNFy3AJNI7Kl6IlFMDQ85Q/ebWqOr+9biu35ZntdVvf3Cp513PlvQCFvh8XYGwEFg8//DDuuusuPPbYY7j66qtxzz334GlPe1p1zs///M/jfe97H7z36LoOP/qjP4rnP//5R2pUjazTgCHfLXdDK2SHcWhqFCegwoe6tyvAkBkFhAwUmpJP8hMUgZzcddqou1OC6FSdNYKY/RWs15mepTJYsTY9bcxiTqChYhzzoPZetmwTMVwng53SnnBSU6Q8fk16n6xVpF26CSQ4rXZkJ2tw2QQJUf0VJcS73ROi/MnPnJ6xs3wxgBEneFNWx6b50jo47fsUb/S7FiharWJVl1lqQ7xb39yUc7OcW36j3WFp0zFwEG0EFq9//etxxx134Pbbb8d73vMe3H333XjnO99ZnfOsZz0Lr3zlK7G3t4dPfOIT+L7v+z589KMfxe7u7pEbZ4FB/l/uQHsumnNXfWdJVVtAhcBoF0ngqjTsXAtSx8WpqYIpQMFJIPXK09Q60awgZpOEywDRe7YOtGnAKCp0yOvDlWcGgEsFpl0GCgsalP0UZsly4malOJABDE4aBRfAkCVSyQMSgsvmh4LZOsf01E5hGxsjDUAGjCne0AZ8sTwBGsBoeKNtsEDRpWPF31RMkE1m+XaCBJbHgh4rn2ugmBojF0oHXuvRRx/FQw89hNtuuw0AcNttt+Ghhx7CF77wheq85z//+djb2wMA3HDDDWBmPPbYY4du0BQK2mMtCOjLfr8KYVdROztXKmTjrLJqYom2TJ5wnf2h+wXqEOH21adz+qSJeJbrdPbarPbxssNMl+LWCaCddTSEOs/maXbXdzEPxJ8wDi6vVsRAsmoxUHq56jXOl78bR5+u4fI1xtGsfCSgiNFJYuO0cawOXV5eBpx6djXXsgnHq3nTVX2+mied8iRdexVvgGmNwoJ5C+zrSJ+3DPppeZ8CinXXO47VkQM1izNnzuC6666D9yk5qfd4ylOegjNnzuCaa66Z/M273/1ufM3XfA2+8iu/8sgNW/ZTWFtu2Tu8vKFMj09/P0V5ruVs9Kb/S4yS2shAEdDi5yhRdnpfD5SIwvZ+6uxC/Z5t5AmgsL9Z9UTRXCM7OCHp5AIcECO8A0KUK0lOR9EwnCvbuRGA6FTDMO2e2Bdv65MARavQvBqqTcQEEBIZLkCh5sfUxrHyBDVVfqUJDZA24A01t2oVpnV80Tbo93q/VgvcJJWCUmuKWA0DWPZ1ALVctzJ/3HTsDs7f+73fw8/+7M/il37plw7922uvvRK3f/7/ddxNumj0f//Muy53Ezam53/u/325m3AoetHn/+3lbsLG9PefQHJwIXQgWJw+fRqf//znEUKA9x4hBDzyyCM4ffr00rl/+Id/iB/7sR/Dvffei6/92q89dGMeffRxfPSv31nNLYqmrRq3xseWaZXjJxJV6p4ktiHzefncWJ0LvObP3oV//tXfl++zpD6aBq707NvP1mFpvq9XPuzeEDYqN5fvWLJai0ou7y/6/L/F/3Hdd8OTzN0ODO/SZ4L4KZIGAcj/znGeefW4c1z9r1TVsDB7UVSTsMWf7OfAyXcB0TQ0w9ULP//v8P+57v+al7wDJI9ISP0q/xeejDnJUOFPRFkmrfm5nifr+KLfZU2Dgb/3Z+/CLz71+2ofBYpGUcwmZA3H+i6Wzy19a9uxSos8SOaRrhlB+L889HZce+2VK5/7IDoQLK699lrceOONeOCBB3D77bfjgQcewI033rhkgjz44IP40R/9Ubz1rW/FM57xjCM3CJhe7mkdPAcpd0dRxSqnZnOPYqKUxvnkQNPgoMKoFCCUrrHOXrVCY98VJPRYCxTU/N7SVBKVqNdUoIyAd8gp5kNMjkDSAc0GJNKzrMiJqZSXPSvAWAYJDUO3QLEqGGtlFGfig13BavmjZonyxvLlIB9CPWGV92oHqR5vgKJqp/l+E1rl1NzEmLmYAVnAhmbIG97wBtx1112499578aQnPQn33HMPAODOO+/Ea1/7Wtx000144xvfiP39fdx99935dz/90z+NG2644YIaeJBHePo39e83oZLzswYM67/IbWgFjs1yHNXCu9m9hai5busss0BhbWdgvf9C3lPIcwprV8DQRD0KJJy1C+Rw9Lpi17TYtsVvANSZuiaAQh2a7VZ0AFmrsDT1jIRpHxMUJAxvcgqOFX3VUjXwV/Amt8F+t4JHB1VNV5rKlHXwb+rfXwzaCCye/vSn4/777186/va3vz1//vf//t8fX6vWULv2PnV8itqut5qABmUBZbnenqPLqS1Q6TKa3bWqwAE0eQkm+EeVIDXvKwRxUyqLow3QgSrA0MzRORKBjXZhgCO3s/UK6rXNs1iA0O8sSAC1RqFAsckmKAECzcOKSVBXDYKBAhpABerANE+Ami/Aet7o8SmgyNdr27+CVmnUbTumvivHlh/quMDjxEZw2o5bhbQHCdZkxzXCXoODCKGqudXW9ZxHswh0FhYz6+hvltq3Rkg2sZEPsoEPojpHQgGMAAIxo2hWZV9Lnk1Nn7XFkvLxNtt0NsPszlcDCoyNwGFTajUMoNYCtUPtJLHJ5N0ChLlUo2XU/NHflHNX3yzScgLq6Tych6Pm0S+YTixYtHRY1eywaFrAoTZHKq0BtUBOgUIbr7GmCt/S+av8F1NAsSllxx6SVpGOOhBkQ6kxS1iePzDlcGlHjJABjFdPx2jCj5saLusSy7a5Rw5DVruQ/5cBAzCaBg7PF2CZN/bzKqAo53F17Cj+i03PV2p/dTijZppONFi0alnr4Z2i6c1lhyfVOFrAyDMY6kFd24yFrKPTXnuK1oGEPa7H1s5W6QlaU80CBvT/FKjgKJX4A1fAASCDxyY0BRBAo22Ydk5pGAcBh4K68qgFjCmzseXLqt5r7zoFEHofe/4qM2X1MyzTccl8e83joMOawhedvumP37bReRqp2b5a2qSjls0JXnFcXuU4V8Cx7uXNa+r7ui4mrxXEg2zgqWcuey5sdCTl/3XG11mfQeJXSOcFs9Fr6qXfhxRgFeBWXleBwGby0jZuMiOucuw6c0z7cFOeENbzp+WNXZFqV0Zsu1qt4kJoU5m/WHSiNQtg2ulzMak1R6yGAdQ+CQsaFa1wAq6/b32Ng2YrO8usu1sVTwIbFWtzJRQtA0A2T/TqNjlW2IAbq2q2LMe90NKxg2idjymfgFrLKD9ep6msplW8qc9Z/n4KKDZ5ykst85vSiQcL4OiddxCYW/PBCmH5vgijXq/1SQC1egtMgMeGNDVLAstA0R63ZJ1l006yGjDkmFI515m71eCxnlalfdN723NWBc3ZY6toHaivcjIfB1+AZWCZ4s+UyXAYGb4QwDgus6OlJwRYAIfz7F5oZ9mlVOvwtKBhz2v9Akdl8iohtPfWNh1EKmzW/HDmWL17twUNPX/Z07JqBt5kGW8qMdGq3ZKb0GG0wAvhC7CeN8Bq/qwyPzax/zfxc637zXHTEwYslI67M1ZpFy1gAFgCDT1PqQWOI7dpA5CYWmI9iGxQm92c1IKGPX/qGrY96zSATbQMOW/6fvm4WX3SQX8YLfBi8AUofT9lGrbnX6hZcTFBYFN6woHFxaYWMIDNQKM1UVqaWqpbd/6y6nvwbLUJKWC0mkebBFnOWS/iG/kvlv5fBgl73lGXvIHVWuBh+HJQ3076K9aAxKrfLLdjs2Lfl5NOJFhcrI5rA7JWUTtzWYEENrNHl2a+Iwjh1L0OM1tZu3cqbN6adssbj9bTYaJnV6UPaH+7Lk6gpaNogVPXPipf7L2qYxsAxaVegtxU7g+iEwkWl5raOAlgGjCA9aBhhf3C7OOjCeFBtMkmpRY4VrVrFUCs+u3UmLzQ/QybaIFT1z66T2k1skyBzklc0bgQ+rIBi6Og66RtvGYmOig5yaa/WTrngFPWzVTtT1vAaK/Rnm9bf5gBvarJm+5nWEUtsB8F1Nfd+zARkzn+5hBAsY5XF0OjPi6tAjjBYHGcHbdJh01pF8A0YOTfNM7N6ese7RlWAcRxsP6gfQdWoI+DA4fRQo5yv3WgfmBY9yHveGFmy8F0qeX+MHRiwQI4no47TIetAwylTYCjvv8B9zzE412I/duu269P1TZNm664HHzOIUyVQ/iZpq4xHYty8PUO60Q+zmF5oXJ/3CChdKLBAriwjjtKp60CDKXDXvFCw3yPk+0tYACHU8OPukN0EzPjMN20jkerQKP6/TFp+u3S6To6dIqBI8j9xQIJpRMPFsByJxzUiRfaaQcBhtImGsdh6TAtXxkgtUbQpgCj+u0Gzs11dFhH5VH77TCgfpxegENPFhdwLyvHq/h5sQHC0hMCLFq6FB20KWAoXWrP90FCeBBgAIdr83FuWNp08B4Xnw9aPt30N4el41wivZSgsIqekGBxqWjV3o/LTcclhO2S6cWgo87qmzqlgaPx52IOvRO3lfuY6EQ+1/M+/ouXuwkV2S3Il5MO247DzEa84euwv78U9ETlzxONtprFIehyCcJx2b3HQRcbAC6kvYeJKj1O+qsMEJa2YHEBNCUkFyqkF0vwTvreg+MGtXVZsi7kOl/OtAWLY6aTLFx59+YJAo1L5bg7yXx5otCJ7cOT5rf4q0SXy7MeiZZeW3ri0Faz+DKljVYbTlhQ0JYuL23BYksraTv4t2TpxJohW9rSlk4WnWiw2PottrSlk0MnGiy2tKUtnRzagsWWtrSljWgLFlva0pY2oi1YbGlLW9qITjxYbJ2cW9rSyaAnRJyFBYyP3vR3L2NLtrSlL1868ZpFS1tNY0tbujz0hAMLQABjCxpb2tKlpY3MkIcffhh33XUXHnvsMVx99dW455578LSnPa06J4SAt7zlLfjN3/xNEBFe/epX47u/+7svRpszbc2TLW3p0tFGYPH6178ed9xxB26//Xa85z3vwd133413vvOd1Tnvfe978elPfxof+MAH8Nhjj+E7vuM78I3f+I146lOfelEa3lKraWzBY0tbOl4i5vVbCx999FG85CUvwcc+9jF47xFCwHOf+1x84AMfwDXXXJPPe/WrX43v+q7vwi233AIAeNOb3oTrr78er3rVqzZuzF/+5VnEeHJyLayja6+9Eo8++vjlbsZG9ERqK/DEau8Tqa3OEb7iK6448u8P1CzOnDmD6667Dt57AID3Hk95ylNw5syZCizOnDmD66+/Pv9/+vRpfO5znztUYy7kQS4HXXvtlZe7CRvTE6mtwBOrvU+ktl4IPSEdnFva0pYuPR0IFqdPn8bnP/95hBAAiCPzkUcewenTp5fO++xnP5v/P3PmDL7yK7/ymJu7pS1t6XLRgWBx7bXX4sYbb8QDDzwAAHjggQdw4403ViYIANxyyy24//77EWPEF77wBXzwgx/ES17ykovT6i1taUuXnA50cALApz71Kdx111344he/iCc96Um455578LVf+7W488478drXvhY33XQTQgh405vehN/6rd8CANx555343u/93ov+AFva0pYuDW0EFlva0pa2tHVwbmlLW9qItmCxpS1taSPagsWWtrSljWgLFlva0pY2oksKFg8//DC+93u/Fy95yUvwvd/7vfjv//2/L50TQsAb3/hGvPCFL8SLXvQi3H///ZeyiRVt0t6f//mfx7d/+7fjZS97Gb7ru74Lv/mbv3npG4rN2qr0p3/6p/gbf+Nv4J577rl0DWxo0/a+733vw0tf+lLcdttteOlLX4q/+Iu/uLQNxWZtffTRR/HqV78aL33pS3HLLbfgDW94A8ZxvORtveeee3DzzTfjhhtuwJ/8yZ9MnnPkMcaXkL7/+7+f3/3udzMz87vf/W7+/u///qVz/sN/+A/8yle+kkMI/Oijj/Lzn/98/sxnPnMpm5lpk/Z+5CMf4XPnzjEz83/9r/+Vn/3sZ/P58+cvaTuZN2srM/M4jvx93/d9/A/+wT/gf/pP/+mlbGJFm7T3wQcf5G/7tm/jRx55hJmZv/jFL/L+/v4lbSfzZm19y1vekvtzsVjwy1/+cv7VX/3VS9pOZubf//3f589+9rP8rd/6rfzJT35y8pyjjrFLplk8+uijeOihh3DbbbcBAG677TY89NBD+MIXvlCd9773vQ/f/d3fDeccrrnmGrzwhS/E+9///kvVzEO39/nPfz729vYAADfccAOYGY899tiJbCsAvO1tb8O3fMu3LKUYuJS0aXt/+Zd/Ga985Svx5Cc/GQBw1VVXYWdn50S2lYhw9uxZxBixWCwwDAOuu+66S9pWAHjOc56zFF3d0lHH2CUDi3Ub0trzLnRD2nHQpu219O53vxtf8zVfc8nD3Ddt6yc+8Ql89KMfxQ/8wA9c0va1tGl7P/WpT+Ezn/kMXvGKV+A7v/M7ce+994IvcVjQpm394R/+YTz88MN43vOel1/PfvazL2lbN6WjjrGtg/OY6Pd+7/fwsz/7s/iZn/mZy92USRqGAT/5kz+JN77xjVnwTzqFEPDJT34S9913H/7Vv/pX+MhHPoL3vOc9l7tZk/T+978fN9xwAz760Y/iIx/5CP7gD/7gsmjEF5MuGVg80TakbdpeAPjDP/xD/NiP/Rh+/ud/Hl/7tV97qZu6UVv//M//HJ/+9Kfx6le/GjfffDPe8Y534N/9u3+Hn/zJnzyR7QWA66+/HrfccgtmsxmuvPJKvOAFL8CDDz54Itv6rne9Cy972cvgnMNVV12Fm2++GR/72McuaVs3paOOsUsGFk+0DWmbtvfBBx/Ej/7oj+Ktb30rnvGMZ1zydgKbtfX666/Hxz72MXzoQx/Chz70Ifydv/N38D3f8z1485vffCLbC4h/4KMf/SiYGcMw4Hd/93fx9V//9SeyrU996lPxkY98BACwWCzwO7/zO/i6r/u6S9rWTenIY+xYXbEH0H/7b/+NX/7yl/OLX/xifvnLX86f+tSnmJn5Va96FT/44IPMLN76u+++m1/wghfwC17wAv43/+bfXMomHrq93/Vd38XPfe5z+WUve1l+feITnziRbbX01re+9bKuhmzS3hAC/9RP/RTfcsstfOutt/JP/dRPcQjhRLb1f/yP/8E/8AM/wLfddht/27d9G7/hDW/gYRgueVvf/OY38/Of/3y+8cYb+Zu+6Zv41ltvXWrrUcfYdiPZlra0pY1o6+Dc0pa2tBFtwWJLW9rSRrQFiy1taUsb0RYstrSlLW1EW7DY0pa2tBFtwWJLW9rSRrQFiy1taUsb0RYstrSlLW1E/3+BA1y07wgisAAAAABJRU5ErkJggg==\n", "text/plain": [ "<Figure size 432x288 with 1 Axes>" ] @@ -77,7 +77,7 @@ }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 128, "id": "b87ebf1b", "metadata": {}, "outputs": [ @@ -85,7 +85,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "Epoch: 9999 | Train loss: 1.292e-02 | Test loss: 3.677e-02 \n", + "Epoch: 9999 | Train loss: 1.783e-04 | Test loss: 9.595e-04 \n", "Training complete - saving.\n" ] } @@ -103,24 +103,24 @@ "\n", "loss_function = nn.MSELoss()#RMSLELoss()\n", "\n", - "LR = 1e-3\n", + "LR = 2e-3\n", "\n", "neurons = [128,64,32,16]\n", "layers = 4\n", - "model = create_mlp(2,1,neurons,layers,nn.Tanh,'model1',device=device)\n", + "model = create_mlp(2,1,neurons,layers,nn.SiLU,'model1',device=device)\n", "model, train_loss, test_loss = model_train_test([points,values,xgrid,ygrid], model, device, n_epochs = int(1e4), n_batches = 1,\n", " loss_function=loss_function, learning_rate=LR, verbose=True, return_losses=True)" ] }, { "cell_type": "code", - "execution_count": 6, + "execution_count": 133, "id": "5d081b13", "metadata": {}, "outputs": [ { "data": { - "image/png": "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\n", + "image/png": "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\n", "text/plain": [ "<Figure size 1080x432 with 1 Axes>" ] @@ -141,13 +141,13 @@ }, { "cell_type": "code", - "execution_count": 7, + "execution_count": 134, "id": "b7bea305", "metadata": {}, "outputs": [ { "data": { - "image/png": "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\n", + "image/png": "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\n", "text/plain": [ "<Figure size 864x576 with 3 Axes>" ] @@ -174,28 +174,18 @@ }, { "cell_type": "code", - "execution_count": 147, + "execution_count": 135, "id": "0f3279d9", "metadata": {}, "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "[0.04633753001689911, 0.035959386403959015]\n", - "[-0.022450417280197144, -0.025545306926706146]\n" - ] - }, { "data": { - "image/png": "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\n", + "image/png": "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\n", "text/plain": [ "<Figure size 576x576 with 2 Axes>" ] }, - "metadata": { - "needs_background": "light" - }, + "metadata": {}, "output_type": "display_data" } ], @@ -206,21 +196,18 @@ "mx = np.max([np.max(nn_err),np.nanmax(cubic_err)])\n", "mn = np.min([np.min(nn_err),np.nanmin(cubic_err)])\n", "\n", - "print([np.max(nn_err),np.nanmax(cubic_err)])\n", - "print([np.min(nn_err),np.nanmin(cubic_err)])\n", - "\n", "fig, ax = plt.subplots(ncols=2,figsize=(8,8))\n", "ax[0].imshow(nn_err.T, extent=(0,1,0,1), origin='lower', vmin=mn, vmax=mx)\n", "ax[0].set_title('NN Error')\n", "ax[1].imshow(cubic_err.T, extent=(0,1,0,1), origin='lower', vmin=mn,vmax=mx)\n", "ax[1].set_title('Cubic Error')\n", - "\n", + "[axes.grid(False) for axes in ax]\n", "plt.show()" ] }, { "cell_type": "code", - "execution_count": 148, + "execution_count": 136, "id": "d4776dbc", "metadata": {}, "outputs": [ @@ -228,7 +215,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "0.003074652924089962 0.003030204088841601\n" + "0.002419143339942683 0.0028046816488585874\n" ] } ], diff --git a/mock_data/mock_data_function_1/notebooks/.ipynb_checkpoints/Check_interpolators_nd-checkpoint.ipynb b/mock_data/mock_data_function_1/notebooks/.ipynb_checkpoints/Check_interpolators_nd-checkpoint.ipynb index fd2f5f956d39dda402dc65e67824b662eaa5c41a..522c53835f17613e44cbe0dc1e0442e07c5fc952 100644 --- a/mock_data/mock_data_function_1/notebooks/.ipynb_checkpoints/Check_interpolators_nd-checkpoint.ipynb +++ b/mock_data/mock_data_function_1/notebooks/.ipynb_checkpoints/Check_interpolators_nd-checkpoint.ipynb @@ -306,7 +306,7 @@ " rate.append(LR)\n", " print(f'Train loss: {train_losses[-1]}, Test loss: {test_losses[-1]}')\n", "\n", - " torch.save(mlp.state_dict(),model_data_dir.format(f'{this_size}_model.pth'))\n", + " torch.save(mlp.state_dict(),full_path.format(f'{this_size}_model.pth'))\n", " \n", " return mlp" ] @@ -368,8 +368,8 @@ " print('Grid dimension: ',n)\n", " xgrid, ygrid, xsamp, ysamp, xtest, ytest, N = get_data(int(1e6), n, sample_ranges)\n", "\n", - " np.save(model_data_dir.format(f'{n}_xtest.npy'),arr=xtest)\n", - " np.save(model_data_dir.format(f'{n}_ytest.npy'),arr=ytest)\n", + " np.save(full_path.format(f'{n}_xtest.npy'),arr=xtest)\n", + " np.save(full_path.format(f'{n}_ytest.npy'),arr=ytest)\n", " \n", " xdata_in = ()\n", " for i in range(xgrid.shape[1]):\n", @@ -387,8 +387,8 @@ " \n", " print('Linear + NN done')\n", " \n", - " np.save(model_data_dir.format(f'{n}_linear_out.npy'),arr=new_values)\n", - " np.save(model_data_dir.format(f'{n}_nearest_out.npy'),arr=new_values2)\n", + " np.save(full_path.format(f'{n}_linear_out.npy'),arr=new_values)\n", + " np.save(full_path.format(f'{n}_nearest_out.npy'),arr=new_values2)\n", " \n", " xtest_in = []\n", "\n", @@ -404,17 +404,17 @@ "\n", " map_values = map_coordinates(cp.array(y_in), cp.array(xtest_in), prefilter=True).get()\n", " print('map_coords done')\n", - " np.save(model_data_dir.format(f'{n}_map_out.npy'),arr=map_values)\n", + " np.save(full_path.format(f'{n}_map_out.npy'),arr=map_values)\n", " \n", " device = 'cuda:0'\n", " \n", " mlp = MLP(n).to(device)\n", " \n", " mlp = prepare_model(mlp,xsamp,np.log(ysamp),xgrid,np.log(ygrid)) # train\n", - "# mlp.load_state_dict(torch.load(model_data_dir.format(f'/{n}_model.pth'))\n", + "# mlp.load_state_dict(torch.load(full_path.format(f'/{n}_model.pth'))\n", "# mlp.eval()\n", - "# xsamp = np.load(model_data_dir.format(f'{n}_xsamp.npy'))\n", - "# ysamp = np.log(np.load(model_data_dir.format(f'{n}_ysamp.npy')))\n", + "# xsamp = np.load(full_path.format(f'{n}_xsamp.npy'))\n", + "# ysamp = np.log(np.load(full_path.format(f'{n}_ysamp.npy')))\n", " \n", " test_input = torch.Tensor(xtest)\n", " normed_input = norm_inputs(test_input, xsamp).float().to(device)\n", @@ -430,7 +430,7 @@ " out_unnorm = np.exp(unnorm(output, ysamp))\n", " out_truth = out_unnorm.flatten()\n", " \n", - " np.save(model_data_dir.format(f'{n}_network_out.npy'),arr=out_truth)\n", + " np.save(full_path.format(f'{n}_network_out.npy'),arr=out_truth)\n", " " ] }, diff --git a/mock_data/mock_data_function_1/notebooks/.ipynb_checkpoints/Interpolators_halving_size-checkpoint.ipynb b/mock_data/mock_data_function_1/notebooks/.ipynb_checkpoints/Interpolators_halving_size-checkpoint.ipynb new file mode 100644 index 0000000000000000000000000000000000000000..fd2f5f956d39dda402dc65e67824b662eaa5c41a --- /dev/null +++ b/mock_data/mock_data_function_1/notebooks/.ipynb_checkpoints/Interpolators_halving_size-checkpoint.ipynb @@ -0,0 +1,467 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": 1, + "id": "e186ab86", + "metadata": {}, + "outputs": [], + "source": [ + "import os\n", + "import torch\n", + "import torch.nn as nn\n", + "from torch.nn.init import xavier_uniform_\n", + "import matplotlib.pyplot as plt\n", + "from scipy.interpolate import RegularGridInterpolator\n", + "from cupyx.scipy.ndimage import map_coordinates\n", + "import cupy as cp\n", + "import numpy as np\n", + "\n", + "from EMRI_DET.nn.model_creation import create_mlp\n", + "from EMRI_DET.nn.model_train_test import model_train_test\n", + "from EMRI_DET.utilities import norm_inputs, norm, unnorm, unnorm_inputs\n", + "\n", + "from pathlib import Path\n", + "model_data_dir = 'downscale_data'\n", + "full_path = f'./_{model_data_dir}/{}'\n", + "Path(f'./_{model_data_dir}').mkdir(parents=True, exist_ok=True)\n" + ] + }, + { + "cell_type": "markdown", + "id": "0afef590", + "metadata": {}, + "source": [ + "# Interpolators vs. NNs: where do they meet up?\n", + "\n", + "Currently we have observed that for a dataset of $M=10^6$ points, a neural network operating with uniform samples in the prior range outperforms classic interpolation schemes operating over an $n$-dimensional grid (where $n=9$ is the number of parameters/dimensionaliy of the prior space). We expect that if we decrease $n$ and keep $M$ constant, eventually this will flip around and the interpolators will outperform the network, even if the network is retrained on this new, lower-dimension space." + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "id": "4864b4b8", + "metadata": {}, + "outputs": [], + "source": [ + "# mock data function \n", + "\n", + "\n", + "def fM(lM):\n", + " return 10 * np.exp(-2 * (lM - 6) ** 2) + 0.1\n", + "\n", + "\n", + "def fq(lq):\n", + " #lq = np.log10(q)\n", + " return 3 / (lq - 4.1)\n", + "\n", + "\n", + "def fa(a):\n", + " return 0.2*a + 0.9\n", + "\n", + "\n", + "def fe(e):\n", + " return 1.1 - 0.2 * e\n", + "\n", + "\n", + "def fY(Y):\n", + " return Y\n", + "\n", + "\n", + "def fTh(th, phase):\n", + " return 0.4*(np.sin(2 * np.pi * th + phase)) ** 2 + 0.6\n", + "\n", + "\n", + "def fPh(ph, phase):\n", + " return 0.4*(np.sin(2 * np.pi * 2 * ph + phase) + 1) ** 2 + 0.6\n", + "\n", + "\n", + "def ft(t):\n", + " return t**0.5/4**0.5\n", + "\n", + "\n", + "def func_out(M, q, a, e, Y, qS, phiS, qK, tpl):\n", + " return fM(M) * fq(q) * fa(a) * fe(e) * fY(Y) * fTh(qS, np.pi / 3) * fTh(qK, np.pi / 4) * fPh(phiS, 3 * np.pi / 8) * ft(tpl)" + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "id": "976dcc72", + "metadata": {}, + "outputs": [], + "source": [ + "def cartesian_product_recursive(arrays, out=None):\n", + " 'Cartesian product of arrays, i.e. write out all combinations of the set elements.'\n", + " arrays = [np.asarray(x) for x in arrays]\n", + " dtype = arrays[0].dtype\n", + "\n", + " n = np.prod([x.size for x in arrays])\n", + " if out is None:\n", + " out = np.zeros([n, len(arrays)], dtype=dtype)\n", + "\n", + " m = n // arrays[0].size\n", + " out[:, 0] = np.repeat(arrays[0], m)\n", + " if arrays[1:]:\n", + " cartesian_product_recursive(arrays[1:], out=out[0:m, 1:])\n", + " for j in range(1, arrays[0].size):\n", + " out[j * m : (j + 1) * m, 1:] = out[0:m, 1:]\n", + " return out\n", + "\n", + "\n", + "\n", + "def get_data(M, n, sample_ranges):\n", + " # For a given M and n, we set a grid size to keep the number of samples ~constant.\n", + " # We output these as 'grid' set and a 'sample' set.\n", + " # Unused dimensions are taken to be constant-valued, with a value between the min and max.\n", + " \n", + " N = np.round(np.exp(np.log(M)/n)).astype(np.int64) # dimension of the grid\n", + " \n", + " ranges_kept = sample_ranges[:n,:]\n", + " fill_samps = np.mean(sample_ranges[n:,:],axis=1).tolist() # list for input\n", + " \n", + " iterators = [np.linspace(nums[0],nums[1],N) for nums in ranges_kept]\n", + "\n", + " combinations = cartesian_product_recursive(iterators)\n", + " \n", + " xgrid = np.array(combinations)\n", + " ygrid = np.zeros(xgrid.shape[0])\n", + " \n", + " for i, combo in enumerate(combinations):\n", + " ygrid[i] = func_out(*combo.tolist(),*fill_samps)\n", + " \n", + " xsamp = np.zeros((M,n))\n", + " \n", + " for i in range(n):\n", + " xsamp[:,i] = np.random.uniform(ranges_kept[i,0], ranges_kept[i,1], M)\n", + " \n", + " ysamp = np.zeros(M)\n", + " \n", + " for i,row in enumerate(xsamp):\n", + " ysamp[i] = func_out(*row.tolist(), *fill_samps)\n", + " \n", + " \n", + " xtest = np.zeros((M,n))\n", + " \n", + " for i in range(n):\n", + " xtest[:,i] = np.random.uniform(ranges_kept[i,0], ranges_kept[i,1], M)\n", + " \n", + " ytest = np.zeros(M)\n", + " \n", + " for i,row in enumerate(xtest):\n", + " ytest[i] = func_out(*row.tolist(), *fill_samps)\n", + " \n", + " return xgrid, ygrid, xsamp, ysamp, xtest, ytest, N" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "id": "9583b6ea", + "metadata": {}, + "outputs": [], + "source": [ + "#Our prior space (params are M,q,a,e,Y,thetaS,phiS,thetaK,t)\n", + "sample_ranges=np.array([\n", + " [np.log10(8e4),np.log10(5e7)],\n", + " [np.log10(5e4),np.log10(5e7)],\n", + " [0.01,0.99],\n", + " [0.01,0.5],\n", + " [0.5,0.99],\n", + " [0,np.pi],\n", + " [0,2*np.pi],\n", + " [0,np.pi],\n", + " [0.5,4],\n", + "])\n", + "\n", + "#xg, yg, xs, ys = get_data(int(1e6),9, sample_ranges)" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "id": "15de4530", + "metadata": {}, + "outputs": [], + "source": [ + "# now we define our network model. \n", + "\n", + "class MLP(nn.Module):\n", + " def __init__(self, n_inputs):\n", + " super().__init__()\n", + " self.layers = nn.Sequential(\n", + " nn.Linear(n_inputs, 128),\n", + " nn.Tanh(),\n", + " nn.Linear(128, 64),\n", + " nn.Tanh(),\n", + " nn.Linear(64, 32),\n", + " nn.Tanh(),\n", + " nn.Linear(32, 16),\n", + " nn.Tanh(),\n", + " nn.Linear(16, 1),\n", + " )\n", + " self.layers.apply(self.init_weights)\n", + "\n", + " def forward(self, x):\n", + " return self.layers(x)\n", + "\n", + " def init_weights(self, m):\n", + " if isinstance(m, nn.Linear):\n", + " xavier_uniform_(m.weight)" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "id": "d166b8d9", + "metadata": {}, + "outputs": [], + "source": [ + "# we define our network train and test functions\n", + "\n", + "def prepare_model(mlp,xtrain,ytrain,xtest,ytest):\n", + " # model training. note that xtrain,ytrain correspond to the SAMPLES, while xtest,ytest are the GRID\n", + " device = \"cuda:0\"\n", + " \n", + " this_size = xtest.shape[1]\n", + " \n", + " xtest = torch.from_numpy(norm_inputs(xtest, xtrain)).to(device).float()\n", + " ytest = torch.from_numpy(norm(ytest, ytrain)).to(device).float()\n", + "\n", + " xtrain = torch.from_numpy(norm_inputs(xtrain, xtrain)).to(device).float()\n", + " ytrain = torch.from_numpy(norm(ytrain, ytrain)).to(device).float()\n", + "\n", + " ytrainsize = len(ytrain)\n", + " ytestsize = len(ytest)\n", + "\n", + " loss_function = nn.MSELoss()\n", + "\n", + " LR = 1e-3\n", + "\n", + " optimizer = torch.optim.Adam(mlp.parameters(), lr=LR)\n", + " scheduler = torch.optim.lr_scheduler.ExponentialLR(optimizer, gamma=0.97)\n", + " train_losses = []\n", + " test_losses = []\n", + " rate = []\n", + " # Run the training loop\n", + "\n", + " datasets = {\"train\": [xtrain, ytrain], \"test\": [xtest, ytest]}\n", + "\n", + " num_epochs = 5000\n", + " nbatch_train = 50\n", + " nbatch_test = 1\n", + " cutoff_LR = 4900\n", + " for epoch in range(num_epochs):\n", + " print(f'Starting epoch {epoch + 1}')\n", + "\n", + " for phase in ['train','test']:\n", + " if phase == 'train':\n", + " mlp.train(True)\n", + " shuffled_inds = torch.randperm(ytrainsize)\n", + "\n", + " # Set current loss value\n", + " current_loss = 0.0\n", + "\n", + " # Iterate over the DataLoader for training data\n", + " # Get and prepare inputs\n", + " inputs, targets = datasets[phase]\n", + " inputs = inputs[shuffled_inds]\n", + " targets = targets[shuffled_inds]\n", + "\n", + " targets = targets.reshape((targets.shape[0], 1))\n", + "\n", + " for i in range(nbatch_train):\n", + " for param in mlp.parameters():\n", + " param.grad = None\n", + " outputs = mlp(inputs[i * ytrainsize // nbatch_train: (i+1)*ytrainsize // nbatch_train])\n", + " loss = torch.sqrt(loss_function(outputs, targets[i * ytrainsize // nbatch_train: (i+1)*ytrainsize // nbatch_train]))\n", + " loss.backward()\n", + " optimizer.step()\n", + " current_loss += loss.item()\n", + "\n", + " train_losses.append(current_loss / nbatch_train)\n", + "\n", + " else:\n", + " with torch.no_grad():\n", + " mlp.train(False)\n", + " shuffled_inds = torch.randperm(ytestsize)\n", + " current_loss = 0.0\n", + " inputs, targets = datasets[phase]\n", + " inputs = inputs[shuffled_inds]\n", + " targets = targets[shuffled_inds]\n", + "\n", + " targets = targets.reshape((targets.shape[0], 1))\n", + "\n", + " for i in range(nbatch_test):\n", + " outputs = mlp(inputs[i * ytestsize // nbatch_test: (i+1)*ytestsize // nbatch_test])\n", + " loss = torch.sqrt(loss_function(outputs, targets[i * ytestsize // nbatch_test: (i+1)*ytestsize // nbatch_test]))\n", + " current_loss += loss.item()\n", + "\n", + " test_losses.append(current_loss / nbatch_test)\n", + "\n", + " if epoch >= cutoff_LR:\n", + " scheduler.step()\n", + " rate.append(scheduler.get_last_lr()[0])\n", + " else:\n", + " rate.append(LR)\n", + " print(f'Train loss: {train_losses[-1]}, Test loss: {test_losses[-1]}')\n", + "\n", + " torch.save(mlp.state_dict(),model_data_dir.format(f'{this_size}_model.pth'))\n", + " \n", + " return mlp" + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "id": "f2206ce9", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Grid dimension: 9\n", + "Linear + NN done\n", + "map_coords done\n" + ] + }, + { + "name": "stderr", + "output_type": "stream", + "text": [ + "OMP: Warning #96: Cannot form a team with 44 threads, using 42 instead.\n", + "OMP: Hint Consider unsetting KMP_DEVICE_THREAD_LIMIT (KMP_ALL_THREADS), KMP_TEAMS_THREAD_LIMIT, and OMP_THREAD_LIMIT (if any are set).\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Grid dimension: 8\n", + "Linear + NN done\n", + "map_coords done\n", + "Grid dimension: 7\n", + "Linear + NN done\n", + "map_coords done\n", + "Grid dimension: 6\n", + "Linear + NN done\n", + "map_coords done\n", + "Grid dimension: 5\n", + "Linear + NN done\n", + "map_coords done\n", + "Grid dimension: 4\n", + "Linear + NN done\n", + "map_coords done\n", + "Grid dimension: 3\n", + "Linear + NN done\n", + "map_coords done\n", + "Grid dimension: 2\n", + "Linear + NN done\n", + "map_coords done\n" + ] + } + ], + "source": [ + "#for n in np.arange(9,1,-1):\n", + "for n in np.arange(9,1,-1):\n", + " print('Grid dimension: ',n)\n", + " xgrid, ygrid, xsamp, ysamp, xtest, ytest, N = get_data(int(1e6), n, sample_ranges)\n", + "\n", + " np.save(model_data_dir.format(f'{n}_xtest.npy'),arr=xtest)\n", + " np.save(model_data_dir.format(f'{n}_ytest.npy'),arr=ytest)\n", + " \n", + " xdata_in = ()\n", + " for i in range(xgrid.shape[1]):\n", + " xdata_in += (np.unique(xgrid[:,i]),)\n", + " yresh = (N,)\n", + " for i in range(n-1):\n", + " yresh += (N,)\n", + " y_in = ygrid.reshape(yresh)\n", + "\n", + " interp = RegularGridInterpolator(xdata_in, y_in, method='linear',bounds_error=False)\n", + " nn_interp = RegularGridInterpolator(xdata_in, y_in, method='nearest',bounds_error=False)\n", + " \n", + " new_values = interp(xtest)\n", + " new_values2 = nn_interp(xtest)\n", + " \n", + " print('Linear + NN done')\n", + " \n", + " np.save(model_data_dir.format(f'{n}_linear_out.npy'),arr=new_values)\n", + " np.save(model_data_dir.format(f'{n}_nearest_out.npy'),arr=new_values2)\n", + " \n", + " xtest_in = []\n", + "\n", + " norms = [listy[1] - listy[0] for listy in sample_ranges]\n", + "\n", + " for i in range(ytest.size):\n", + " sub = []\n", + " for j in range(n):\n", + " sub.append((N-1)*(xtest[i, j] - sample_ranges[j][0])/norms[j])\n", + " xtest_in.append(sub)\n", + "\n", + " xtest_in = np.asarray(xtest_in).T\n", + "\n", + " map_values = map_coordinates(cp.array(y_in), cp.array(xtest_in), prefilter=True).get()\n", + " print('map_coords done')\n", + " np.save(model_data_dir.format(f'{n}_map_out.npy'),arr=map_values)\n", + " \n", + " device = 'cuda:0'\n", + " \n", + " mlp = MLP(n).to(device)\n", + " \n", + " mlp = prepare_model(mlp,xsamp,np.log(ysamp),xgrid,np.log(ygrid)) # train\n", + "# mlp.load_state_dict(torch.load(model_data_dir.format(f'/{n}_model.pth'))\n", + "# mlp.eval()\n", + "# xsamp = np.load(model_data_dir.format(f'{n}_xsamp.npy'))\n", + "# ysamp = np.log(np.load(model_data_dir.format(f'{n}_ysamp.npy')))\n", + " \n", + " test_input = torch.Tensor(xtest)\n", + " normed_input = norm_inputs(test_input, xsamp).float().to(device)\n", + "\n", + " with torch.no_grad():\n", + " out = []\n", + " for i in range(20):\n", + " output = mlp(normed_input[i * ytest.size // 20 : (i+1)*ytest.size // 20])\n", + " out.append(output.detach().cpu().numpy())\n", + "\n", + " output = np.concatenate(out)\n", + "\n", + " out_unnorm = np.exp(unnorm(output, ysamp))\n", + " out_truth = out_unnorm.flatten()\n", + " \n", + " np.save(model_data_dir.format(f'{n}_network_out.npy'),arr=out_truth)\n", + " " + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "f4481771", + "metadata": {}, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3 (ipykernel)", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.9.6" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +} diff --git a/mock_data/mock_data_function_1/notebooks/.ipynb_checkpoints/halving_size_plot-checkpoint.ipynb b/mock_data/mock_data_function_1/notebooks/.ipynb_checkpoints/halving_size_plot-checkpoint.ipynb new file mode 100644 index 0000000000000000000000000000000000000000..b178f2bed3f2020615276043177caa62e9ecb68d --- /dev/null +++ b/mock_data/mock_data_function_1/notebooks/.ipynb_checkpoints/halving_size_plot-checkpoint.ipynb @@ -0,0 +1,107 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": 1, + "id": "dd1f6d19", + "metadata": {}, + "outputs": [], + "source": [ + "import numpy as np\n", + "import matplotlib.pyplot as plt" + ] + }, + { + "cell_type": "code", + "execution_count": 22, + "id": "abc2e918", + "metadata": { + "scrolled": false + }, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "<Figure size 1440x720 with 8 Axes>" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots(nrows=2, ncols=4, figsize=(20,10), facecolor=\"w\")\n", + "ax = ax.flatten()\n", + "\n", + "i = 0\n", + "for n in range(2,10):\n", + " ydata = np.load(f'./_downscale_data/{n}_ytest.npy')\n", + " \n", + " lin = np.load(f'./_downscale_data/{n}_linear_out.npy')\n", + " near = np.load(f'./_downscale_data/{n}_nearest_out.npy')\n", + " mcoord = np.load(f'./_downscale_data/{n}_map_out.npy')\n", + " network = np.load(f'./_downscale_data/{n}_network_out.npy') \n", + "\n", + " # Absolute difference\n", + "# ax[i].hist(np.log10(abs(lin-ydata)), bins='auto', label='Linear Interp.',histtype='step',density=True, color=\"blue\")\n", + "# ax[i].hist(np.log10(abs(near-ydata)), bins='auto', label='Nearest Neighbour',histtype='step',density=True, color=\"orange\")\n", + "# if n != 2:\n", + "# ax[i].hist(np.log10(abs(mcoord-ydata)), bins='auto', label='n-Cubic Spine interp.',histtype='step',density=True, color=\"green\")\n", + "# else:\n", + "# temp = abs(mcoord-ydata)\n", + "# ax[i].hist(np.log10(temp[temp != 0]), bins='auto', label='n-Cubic Spine interp.',histtype='step',density=True, color=\"green\")\n", + " \n", + "# ax[i].hist(np.log10(abs(network-ydata)), bins='auto',label='NN', histtype='step', density=True, color=\"red\")\n", + " \n", + " # Relative difference\n", + " if n > 5:\n", + " ax[i].hist(np.log10(near/ydata), bins='auto', label='Nearest Neighbour',histtype='step',density=False, color='orange')\n", + " if n > 2:\n", + " temp = mcoord/ydata\n", + " ax[i].hist(np.log10(temp[temp != 0]), bins='auto', label='n-Cubic Spine interp.',histtype='step',density=False, color='green')\n", + " ax[i].hist(np.log10(lin/ydata), bins='auto', label='Linear Interp.',histtype='step',density=False, color='blue')\n", + " \n", + " ax[i].hist(np.log10(network/ydata), bins='auto',label='NN', histtype='step', density=False, color='red')\n", + "\n", + " ax[i].legend(loc='upper left')\n", + " ax[i].set_title(f'{n} dimensional parameter space')\n", + " i += 1\n", + "\n", + "fig.suptitle('Network vs. Interpolators - log10(ratio)',fontsize=24)\n", + "#plt.xlabel('log10(|diff|) between truth and interpolated')\n", + "\n", + "plt.show()\n" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "0e249162", + "metadata": {}, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3 (ipykernel)", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.9.6" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +} diff --git a/mock_data/mock_data_function_1/notebooks/Check_interpolators_nd.ipynb b/mock_data/mock_data_function_1/notebooks/Check_interpolators_nd.ipynb index fd2f5f956d39dda402dc65e67824b662eaa5c41a..522c53835f17613e44cbe0dc1e0442e07c5fc952 100644 --- a/mock_data/mock_data_function_1/notebooks/Check_interpolators_nd.ipynb +++ b/mock_data/mock_data_function_1/notebooks/Check_interpolators_nd.ipynb @@ -306,7 +306,7 @@ " rate.append(LR)\n", " print(f'Train loss: {train_losses[-1]}, Test loss: {test_losses[-1]}')\n", "\n", - " torch.save(mlp.state_dict(),model_data_dir.format(f'{this_size}_model.pth'))\n", + " torch.save(mlp.state_dict(),full_path.format(f'{this_size}_model.pth'))\n", " \n", " return mlp" ] @@ -368,8 +368,8 @@ " print('Grid dimension: ',n)\n", " xgrid, ygrid, xsamp, ysamp, xtest, ytest, N = get_data(int(1e6), n, sample_ranges)\n", "\n", - " np.save(model_data_dir.format(f'{n}_xtest.npy'),arr=xtest)\n", - " np.save(model_data_dir.format(f'{n}_ytest.npy'),arr=ytest)\n", + " np.save(full_path.format(f'{n}_xtest.npy'),arr=xtest)\n", + " np.save(full_path.format(f'{n}_ytest.npy'),arr=ytest)\n", " \n", " xdata_in = ()\n", " for i in range(xgrid.shape[1]):\n", @@ -387,8 +387,8 @@ " \n", " print('Linear + NN done')\n", " \n", - " np.save(model_data_dir.format(f'{n}_linear_out.npy'),arr=new_values)\n", - " np.save(model_data_dir.format(f'{n}_nearest_out.npy'),arr=new_values2)\n", + " np.save(full_path.format(f'{n}_linear_out.npy'),arr=new_values)\n", + " np.save(full_path.format(f'{n}_nearest_out.npy'),arr=new_values2)\n", " \n", " xtest_in = []\n", "\n", @@ -404,17 +404,17 @@ "\n", " map_values = map_coordinates(cp.array(y_in), cp.array(xtest_in), prefilter=True).get()\n", " print('map_coords done')\n", - " np.save(model_data_dir.format(f'{n}_map_out.npy'),arr=map_values)\n", + " np.save(full_path.format(f'{n}_map_out.npy'),arr=map_values)\n", " \n", " device = 'cuda:0'\n", " \n", " mlp = MLP(n).to(device)\n", " \n", " mlp = prepare_model(mlp,xsamp,np.log(ysamp),xgrid,np.log(ygrid)) # train\n", - "# mlp.load_state_dict(torch.load(model_data_dir.format(f'/{n}_model.pth'))\n", + "# mlp.load_state_dict(torch.load(full_path.format(f'/{n}_model.pth'))\n", "# mlp.eval()\n", - "# xsamp = np.load(model_data_dir.format(f'{n}_xsamp.npy'))\n", - "# ysamp = np.log(np.load(model_data_dir.format(f'{n}_ysamp.npy')))\n", + "# xsamp = np.load(full_path.format(f'{n}_xsamp.npy'))\n", + "# ysamp = np.log(np.load(full_path.format(f'{n}_ysamp.npy')))\n", " \n", " test_input = torch.Tensor(xtest)\n", " normed_input = norm_inputs(test_input, xsamp).float().to(device)\n", @@ -430,7 +430,7 @@ " out_unnorm = np.exp(unnorm(output, ysamp))\n", " out_truth = out_unnorm.flatten()\n", " \n", - " np.save(model_data_dir.format(f'{n}_network_out.npy'),arr=out_truth)\n", + " np.save(full_path.format(f'{n}_network_out.npy'),arr=out_truth)\n", " " ] }, diff --git a/mock_data/mock_data_function_1/notebooks/Interpolators_halving_size.ipynb b/mock_data/mock_data_function_1/notebooks/Interpolators_halving_size.ipynb new file mode 100644 index 0000000000000000000000000000000000000000..e6afb0b126da71e2ed26e094d48ef0d81a7d27dc --- /dev/null +++ b/mock_data/mock_data_function_1/notebooks/Interpolators_halving_size.ipynb @@ -0,0 +1,10833 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": 39, + "id": "e186ab86", + "metadata": {}, + "outputs": [], + "source": [ + "import os\n", + "import torch\n", + "import torch.nn as nn\n", + "from torch.nn.init import xavier_uniform_\n", + "import matplotlib.pyplot as plt\n", + "from scipy.interpolate import RegularGridInterpolator\n", + "from cupyx.scipy.ndimage import map_coordinates\n", + "import cupy as cp\n", + "import numpy as np\n", + "\n", + "from EMRI_DET.nn.model_creation import create_mlp\n", + "from EMRI_DET.nn.model_train_test import model_train_test\n", + "from EMRI_DET.utilities import norm_inputs, norm, unnorm, unnorm_inputs\n", + "\n", + "from pathlib import Path\n", + "model_data_dir = 'halving_data'\n", + "full_path = f'./_{model_data_dir}'+'/{}'\n", + "Path(f'./_{model_data_dir}').mkdir(parents=True, exist_ok=True)\n" + ] + }, + { + "cell_type": "markdown", + "id": "0afef590", + "metadata": {}, + "source": [ + "# Interpolators vs. NNs: where do they meet up?\n", + "\n", + "Currently we have observed that for a dataset of $M=10^6$ points, a neural network operating with uniform samples in the prior range outperforms classic interpolation schemes operating over an $n$-dimensional grid (where $n=9$ is the number of parameters/dimensionaliy of the prior space). We expect that if we decrease $n$ and keep $M$ constant, eventually this will flip around and the interpolators will outperform the network, even if the network is retrained on this new, lower-dimension space." + ] + }, + { + "cell_type": "code", + "execution_count": 40, + "id": "4864b4b8", + "metadata": {}, + "outputs": [], + "source": [ + "# mock data function \n", + "\n", + "\n", + "def fM(lM):\n", + " return 10 * np.exp(-2 * (lM - 6) ** 2) + 0.1\n", + "\n", + "\n", + "def fq(lq):\n", + " #lq = np.log10(q)\n", + " return 3 / (lq - 4.1)\n", + "\n", + "\n", + "def fa(a):\n", + " return 0.2*a + 0.9\n", + "\n", + "\n", + "def fe(e):\n", + " return 1.1 - 0.2 * e\n", + "\n", + "\n", + "def fY(Y):\n", + " return Y\n", + "\n", + "\n", + "def fTh(th, phase):\n", + " return 0.4*(np.sin(2 * np.pi * th + phase)) ** 2 + 0.6\n", + "\n", + "\n", + "def fPh(ph, phase):\n", + " return 0.4*(np.sin(2 * np.pi * 2 * ph + phase) + 1) ** 2 + 0.6\n", + "\n", + "\n", + "def ft(t):\n", + " return t**0.5/4**0.5\n", + "\n", + "\n", + "def func_out(M, q, a, e, Y, qS, phiS, qK, tpl):\n", + " return fM(M) * fq(q) * fa(a) * fe(e) * fY(Y) * fTh(qS, np.pi / 3) * fTh(qK, np.pi / 4) * fPh(phiS, 3 * np.pi / 8) * ft(tpl)" + ] + }, + { + "cell_type": "code", + "execution_count": 49, + "id": "6408c0bd", + "metadata": {}, + "outputs": [], + "source": [ + "def cartesian_product_recursive(arrays, out=None):\n", + " 'Cartesian product of arrays, i.e. write out all combinations of the set elements.'\n", + " arrays = [np.asarray(x) for x in arrays]\n", + " dtype = arrays[0].dtype\n", + "\n", + " n = np.prod([x.size for x in arrays])\n", + " if out is None:\n", + " out = np.zeros([n, len(arrays)], dtype=dtype)\n", + "\n", + " m = n // arrays[0].size\n", + " out[:, 0] = np.repeat(arrays[0], m)\n", + " if arrays[1:]:\n", + " cartesian_product_recursive(arrays[1:], out=out[0:m, 1:])\n", + " for j in range(1, arrays[0].size):\n", + " out[j * m : (j + 1) * m, 1:] = out[0:m, 1:]\n", + " return out\n", + "\n", + "\n", + "\n", + "def get_data(M, exponent, sample_ranges):\n", + " # For a given sample size M, grid size is found. We then have the sample_ranges `exponent` times, zooming\n", + " # in on the center of parameter space.\n", + " \n", + " N = np.round(np.exp(np.log(M)/9)).astype(np.int64) # dimension of the grid\n", + " \n", + " midpoints = np.mean(sample_ranges,axis=1)\n", + " sample_ranges = sample_ranges.copy()\n", + " for num in range(exponent):\n", + " sample_ranges[:,0] = 0.5*(sample_ranges[:,0] + midpoints[:])\n", + " sample_ranges[:,1] = 0.5*(sample_ranges[:,1] + midpoints[:])\n", + " print(sample_ranges)\n", + " iterators = [np.linspace(nums[0],nums[1],N) for nums in sample_ranges]\n", + "\n", + " combinations = cartesian_product_recursive(iterators)\n", + " \n", + " xgrid = np.array(combinations)\n", + " ygrid = np.zeros(xgrid.shape[0])\n", + " \n", + " for i, combo in enumerate(combinations):\n", + " ygrid[i] = func_out(*combo.tolist())\n", + " \n", + " xsamp = np.zeros((M,9))\n", + " \n", + " for i in range(9):\n", + " xsamp[:,i] = np.random.uniform(sample_ranges[i,0], sample_ranges[i,1], M)\n", + " \n", + " ysamp = np.zeros(M)\n", + " \n", + " for i,row in enumerate(xsamp):\n", + " ysamp[i] = func_out(*row.tolist())\n", + " \n", + " \n", + " xtest = np.zeros((M,9))\n", + " \n", + " for i in range(9):\n", + " xtest[:,i] = np.random.uniform(sample_ranges[i,0], sample_ranges[i,1], M)\n", + " \n", + " ytest = np.zeros(M)\n", + " \n", + " for i,row in enumerate(xtest):\n", + " ytest[i] = func_out(*row.tolist())\n", + " \n", + " revised_ranges = np.copy(sample_ranges)\n", + " \n", + " return xgrid, ygrid, xsamp, ysamp, xtest, ytest, N, revised_ranges" + ] + }, + { + "cell_type": "code", + "execution_count": 50, + "id": "9583b6ea", + "metadata": {}, + "outputs": [], + "source": [ + "#Our prior space (params are M,q,a,e,Y,thetaS,phiS,thetaK,t)\n", + "sample_ranges=np.array([\n", + " [np.log10(8e4),np.log10(5e7)],\n", + " [np.log10(5e4),np.log10(5e7)],\n", + " [0.01,0.99],\n", + " [0.01,0.5],\n", + " [0.5,0.99],\n", + " [0,np.pi],\n", + " [0,2*np.pi],\n", + " [0,np.pi],\n", + " [0.5,4],\n", + "])\n", + "\n", + "#xg, yg, xs, ys = get_data(int(1e6),9, sample_ranges)" + ] + }, + { + "cell_type": "code", + "execution_count": 55, + "id": "15de4530", + "metadata": {}, + "outputs": [], + "source": [ + "# now we define our network model. \n", + "\n", + "class MLP(nn.Module):\n", + " def __init__(self):\n", + " super().__init__()\n", + " self.layers = nn.Sequential(\n", + " nn.Linear(9, 128),\n", + " nn.SiLU(),\n", + " nn.Linear(128, 64),\n", + " nn.SiLU(),\n", + " nn.Linear(64, 32),\n", + " nn.SiLU(),\n", + " nn.Linear(32, 16),\n", + " nn.SiLU(),\n", + " nn.Linear(16, 1),\n", + " )\n", + " self.layers.apply(self.init_weights)\n", + "\n", + " def forward(self, x):\n", + " return self.layers(x)\n", + "\n", + " def init_weights(self, m):\n", + " if isinstance(m, nn.Linear):\n", + " xavier_uniform_(m.weight)" + ] + }, + { + "cell_type": "code", + "execution_count": 58, + "id": "d166b8d9", + "metadata": {}, + "outputs": [], + "source": [ + "# we define our network train and test functions\n", + "\n", + "def prepare_model(mlp,xtrain,ytrain,xtest,ytest):\n", + " # model training. note that xtrain,ytrain correspond to the SAMPLES, while xtest,ytest are the GRID\n", + " device = \"cuda:0\"\n", + " \n", + " xtest = torch.from_numpy(norm_inputs(xtest, xtrain)).to(device).float()\n", + " ytest = torch.from_numpy(norm(ytest, ytrain)).to(device).float()\n", + "\n", + " xtrain = torch.from_numpy(norm_inputs(xtrain, xtrain)).to(device).float()\n", + " ytrain = torch.from_numpy(norm(ytrain, ytrain)).to(device).float()\n", + "\n", + " ytrainsize = len(ytrain)\n", + " ytestsize = len(ytest)\n", + "\n", + " loss_function = nn.MSELoss()\n", + "\n", + " LR = 2e-3\n", + "\n", + " optimizer = torch.optim.Adam(mlp.parameters(), lr=LR)\n", + " scheduler = torch.optim.lr_scheduler.ExponentialLR(optimizer, gamma=0.97)\n", + " train_losses = []\n", + " test_losses = []\n", + " rate = []\n", + " # Run the training loop\n", + "\n", + " datasets = {\"train\": [xtrain, ytrain], \"test\": [xtest, ytest]}\n", + "\n", + " num_epochs = 5000\n", + " nbatch_train = 50\n", + " nbatch_test = 1\n", + " cutoff_LR = 4800\n", + " for epoch in range(num_epochs):\n", + " print(f'Starting epoch {epoch + 1}')\n", + "\n", + " for phase in ['train','test']:\n", + " if phase == 'train':\n", + " mlp.train(True)\n", + " shuffled_inds = torch.randperm(ytrainsize)\n", + "\n", + " # Set current loss value\n", + " current_loss = 0.0\n", + "\n", + " # Iterate over the DataLoader for training data\n", + " # Get and prepare inputs\n", + " inputs, targets = datasets[phase]\n", + " inputs = inputs[shuffled_inds]\n", + " targets = targets[shuffled_inds]\n", + "\n", + " targets = targets.reshape((targets.shape[0], 1))\n", + "\n", + " for i in range(nbatch_train):\n", + " for param in mlp.parameters():\n", + " param.grad = None\n", + " outputs = mlp(inputs[i * ytrainsize // nbatch_train: (i+1)*ytrainsize // nbatch_train])\n", + " loss = loss_function(outputs, targets[i * ytrainsize // nbatch_train: (i+1)*ytrainsize // nbatch_train])\n", + " loss.backward()\n", + " optimizer.step()\n", + " current_loss += loss.item()\n", + "\n", + " train_losses.append(current_loss / nbatch_train)\n", + "\n", + " else:\n", + " with torch.no_grad():\n", + " mlp.train(False)\n", + " shuffled_inds = torch.randperm(ytestsize)\n", + " current_loss = 0.0\n", + " inputs, targets = datasets[phase]\n", + " inputs = inputs[shuffled_inds]\n", + " targets = targets[shuffled_inds]\n", + "\n", + " targets = targets.reshape((targets.shape[0], 1))\n", + "\n", + " for i in range(nbatch_test):\n", + " outputs = mlp(inputs[i * ytestsize // nbatch_test: (i+1)*ytestsize // nbatch_test])\n", + " loss = loss_function(outputs, targets[i * ytestsize // nbatch_test: (i+1)*ytestsize // nbatch_test])\n", + " current_loss += loss.item()\n", + "\n", + " test_losses.append(current_loss / nbatch_test)\n", + "\n", + " if epoch >= cutoff_LR:\n", + " scheduler.step()\n", + " rate.append(scheduler.get_last_lr()[0])\n", + " else:\n", + " rate.append(LR)\n", + " print(f'Train loss: {train_losses[-1]}, Test loss: {test_losses[-1]}')\n", + "\n", + " torch.save(mlp.state_dict(),full_path.format(f'model.pth'))\n", + " \n", + " return mlp" + ] + }, + { + "cell_type": "code", + "execution_count": 59, + "id": "aa4f706f", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "[[4.90308999 7.69897 ]\n", + " [4.69897 7.69897 ]\n", + " [0.01 0.99 ]\n", + " [0.01 0.5 ]\n", + " [0.5 0.99 ]\n", + " [0. 3.14159265]\n", + " [0. 6.28318531]\n", + " [0. 3.14159265]\n", + " [0.5 4. ]]\n", + "Starting epoch 1\n", + "Train loss: 0.6269454747438431, Test loss: 0.5284115076065063\n", + "Starting epoch 2\n", + "Train loss: 0.3765950983762741, Test loss: 0.4013672173023224\n", + "Starting epoch 3\n", + "Train loss: 0.36587609350681305, Test loss: 0.4039730727672577\n", + "Starting epoch 4\n", + "Train loss: 0.3652468681335449, Test loss: 0.39670512080192566\n", + "Starting epoch 5\n", + "Train loss: 0.3649262034893036, Test loss: 0.4001621603965759\n", + "Starting epoch 6\n", + "Train loss: 0.3646185100078583, Test loss: 0.3952043354511261\n", + "Starting epoch 7\n", + "Train loss: 0.3644212627410889, Test loss: 0.39899498224258423\n", + "Starting epoch 8\n", + "Train loss: 0.3644523477554321, Test loss: 0.3976602256298065\n", + "Starting epoch 9\n", + "Train loss: 0.3645124119520187, Test loss: 0.3898366093635559\n", + "Starting epoch 10\n", + "Train loss: 0.36424205243587493, Test loss: 0.39315274357795715\n", + "Starting epoch 11\n", + "Train loss: 0.36411518514156344, Test loss: 0.3946133852005005\n", + "Starting epoch 12\n", + "Train loss: 0.36415678143501284, Test loss: 0.3856911063194275\n", + "Starting epoch 13\n", + "Train loss: 0.3641889160871506, Test loss: 0.39931097626686096\n", + "Starting epoch 14\n", + "Train loss: 0.36403424918651583, Test loss: 0.39666590094566345\n", + "Starting epoch 15\n", + "Train loss: 0.3640035218000412, Test loss: 0.39682015776634216\n", + "Starting epoch 16\n", + "Train loss: 0.36402188777923583, Test loss: 0.3935002088546753\n", + "Starting epoch 17\n", + "Train loss: 0.3638608258962631, Test loss: 0.394620418548584\n", + "Starting epoch 18\n", + "Train loss: 0.3638099902868271, Test loss: 0.3872130513191223\n", + "Starting epoch 19\n", + "Train loss: 0.36376561045646666, Test loss: 0.38870760798454285\n", + "Starting epoch 20\n", + "Train loss: 0.3637218630313873, Test loss: 0.3897445797920227\n", + "Starting epoch 21\n", + "Train loss: 0.36373305022716523, Test loss: 0.39735373854637146\n", + "Starting epoch 22\n", + "Train loss: 0.36374634683132173, Test loss: 0.39328858256340027\n", + "Starting epoch 23\n", + "Train loss: 0.3637572681903839, Test loss: 0.3934611976146698\n", + "Starting epoch 24\n", + "Train loss: 0.36363638401031495, Test loss: 0.3903030753135681\n", + "Starting epoch 25\n", + "Train loss: 0.36365164756774904, Test loss: 0.38866034150123596\n", + "Starting epoch 26\n", + "Train loss: 0.3635778832435608, Test loss: 0.39219385385513306\n", + "Starting epoch 27\n", + "Train loss: 0.36358909726142885, Test loss: 0.39557766914367676\n", + "Starting epoch 28\n", + "Train loss: 0.36358524978160855, Test loss: 0.3965103328227997\n", + "Starting epoch 29\n", + "Train loss: 0.36365454494953153, Test loss: 0.3884894549846649\n", + "Starting epoch 30\n", + "Train loss: 0.3636226361989975, Test loss: 0.3901088237762451\n", + "Starting epoch 31\n", + "Train loss: 0.3636034917831421, Test loss: 0.3908214569091797\n", + "Starting epoch 32\n", + "Train loss: 0.3636446106433868, Test loss: 0.39351120591163635\n", + "Starting epoch 33\n", + "Train loss: 0.3634932839870453, Test loss: 0.3904225528240204\n", + "Starting epoch 34\n", + "Train loss: 0.36357990264892576, Test loss: 0.39153966307640076\n", + "Starting epoch 35\n", + "Train loss: 0.36365136206150056, Test loss: 0.391300767660141\n", + "Starting epoch 36\n", + "Train loss: 0.363581211566925, Test loss: 0.38834667205810547\n", + "Starting epoch 37\n", + "Train loss: 0.36352218270301817, Test loss: 0.39041733741760254\n", + "Starting epoch 38\n", + "Train loss: 0.36363439619541166, Test loss: 0.39175882935523987\n", + "Starting epoch 39\n", + "Train loss: 0.3634684079885483, Test loss: 0.388094425201416\n", + "Starting epoch 40\n", + "Train loss: 0.3634840553998947, Test loss: 0.39112356305122375\n", + "Starting epoch 41\n", + "Train loss: 0.36346020638942717, Test loss: 0.39075276255607605\n", + "Starting epoch 42\n", + "Train loss: 0.36343589901924134, Test loss: 0.39647629857063293\n", + "Starting epoch 43\n", + "Train loss: 0.3635596871376038, Test loss: 0.38538452982902527\n", + "Starting epoch 44\n", + "Train loss: 0.36356828570365907, Test loss: 0.3951885402202606\n", + "Starting epoch 45\n", + "Train loss: 0.36349795043468475, Test loss: 0.3874099850654602\n", + "Starting epoch 46\n", + "Train loss: 0.3633770859241486, Test loss: 0.38855671882629395\n", + "Starting epoch 47\n", + "Train loss: 0.3634782075881958, Test loss: 0.39070457220077515\n", + "Starting epoch 48\n", + "Train loss: 0.3633826941251755, Test loss: 0.3919328451156616\n", + "Starting epoch 49\n", + "Train loss: 0.3633652251958847, Test loss: 0.38940519094467163\n", + "Starting epoch 50\n", + "Train loss: 0.3634448111057281, Test loss: 0.3861110210418701\n", + "Starting epoch 51\n", + "Train loss: 0.3634380453824997, Test loss: 0.391597181558609\n", + "Starting epoch 52\n", + "Train loss: 0.3633087921142578, Test loss: 0.3915936350822449\n", + "Starting epoch 53\n", + "Train loss: 0.36333955585956573, Test loss: 0.3903568387031555\n", + "Starting epoch 54\n", + "Train loss: 0.36348390877246856, Test loss: 0.3807176649570465\n", + "Starting epoch 55\n", + "Train loss: 0.3635167443752289, Test loss: 0.3862379193305969\n", + "Starting epoch 56\n", + "Train loss: 0.3633750128746033, Test loss: 0.39163821935653687\n", + "Starting epoch 57\n", + "Train loss: 0.3632934474945068, Test loss: 0.38970687985420227\n", + "Starting epoch 58\n", + "Train loss: 0.3634952735900879, Test loss: 0.39030253887176514\n", + "Starting epoch 59\n", + "Train loss: 0.36321195304393766, Test loss: 0.38953039050102234\n", + "Starting epoch 60\n", + "Train loss: 0.36325901448726655, Test loss: 0.3876489996910095\n", + "Starting epoch 61\n", + "Train loss: 0.3632402944564819, Test loss: 0.3881087899208069\n", + "Starting epoch 62\n", + "Train loss: 0.36332199096679685, Test loss: 0.38456621766090393\n", + "Starting epoch 63\n", + "Train loss: 0.36325289189815524, Test loss: 0.38578370213508606\n", + "Starting epoch 64\n", + "Train loss: 0.363210186958313, Test loss: 0.38136112689971924\n", + "Starting epoch 65\n", + "Train loss: 0.3632686239480972, Test loss: 0.3843691051006317\n", + "Starting epoch 66\n", + "Train loss: 0.36338461816310885, Test loss: 0.3849172592163086\n", + "Starting epoch 67\n", + "Train loss: 0.3633870106935501, Test loss: 0.38647153973579407\n", + "Starting epoch 68\n", + "Train loss: 0.3632158809900284, Test loss: 0.38801658153533936\n", + "Starting epoch 69\n", + "Train loss: 0.3632782286405563, Test loss: 0.3737984001636505\n", + "Starting epoch 70\n", + "Train loss: 0.36321691393852235, Test loss: 0.38373449444770813\n", + "Starting epoch 71\n", + "Train loss: 0.3631861412525177, Test loss: 0.38744238018989563\n", + "Starting epoch 72\n", + "Train loss: 0.3632031363248825, Test loss: 0.3847767114639282\n", + "Starting epoch 73\n", + "Train loss: 0.36324151456356046, Test loss: 0.39386335015296936\n", + "Starting epoch 74\n", + "Train loss: 0.36326436698436737, Test loss: 0.3876800835132599\n", + "Starting epoch 75\n", + "Train loss: 0.3633109426498413, Test loss: 0.3861430883407593\n", + "Starting epoch 76\n", + "Train loss: 0.36330102920532226, Test loss: 0.3844907581806183\n", + "Starting epoch 77\n", + "Train loss: 0.3631198567152023, Test loss: 0.3762993812561035\n", + "Starting epoch 78\n", + "Train loss: 0.363177136182785, Test loss: 0.3838859498500824\n", + "Starting epoch 79\n", + "Train loss: 0.36331279039382935, Test loss: 0.3843669891357422\n", + "Starting epoch 80\n", + "Train loss: 0.3631077021360397, Test loss: 0.3871324956417084\n", + "Starting epoch 81\n", + "Train loss: 0.36326520204544066, Test loss: 0.38068416714668274\n", + "Starting epoch 82\n", + "Train loss: 0.36312822103500364, Test loss: 0.38697195053100586\n", + "Starting epoch 83\n", + "Train loss: 0.36317186295986176, Test loss: 0.378451943397522\n", + "Starting epoch 84\n", + "Train loss: 0.36322825610637666, Test loss: 0.38248491287231445\n", + "Starting epoch 85\n", + "Train loss: 0.36321584761142733, Test loss: 0.3831181526184082\n", + "Starting epoch 86\n", + "Train loss: 0.36309535026550294, Test loss: 0.38899680972099304\n", + "Starting epoch 87\n", + "Train loss: 0.36320776522159576, Test loss: 0.38244539499282837\n", + "Starting epoch 88\n", + "Train loss: 0.3631939178705215, Test loss: 0.3775007426738739\n", + "Starting epoch 89\n", + "Train loss: 0.36318371295928953, Test loss: 0.38037392497062683\n", + "Starting epoch 90\n", + "Train loss: 0.36297671377658847, Test loss: 0.38145604729652405\n", + "Starting epoch 91\n", + "Train loss: 0.36292528092861176, Test loss: 0.37857404351234436\n", + "Starting epoch 92\n", + "Train loss: 0.3629248684644699, Test loss: 0.38857540488243103\n", + "Starting epoch 93\n", + "Train loss: 0.36316669046878813, Test loss: 0.37795305252075195\n", + "Starting epoch 94\n", + "Train loss: 0.3628710299730301, Test loss: 0.3738834857940674\n", + "Starting epoch 95\n", + "Train loss: 0.36274663388729095, Test loss: 0.3716483414173126\n", + "Starting epoch 96\n", + "Train loss: 0.3625561785697937, Test loss: 0.37074708938598633\n", + "Starting epoch 97\n", + "Train loss: 0.362571861743927, Test loss: 0.36867642402648926\n", + "Starting epoch 98\n", + "Train loss: 0.36228335082530977, Test loss: 0.3665842115879059\n", + "Starting epoch 99\n", + "Train loss: 0.3619890224933624, Test loss: 0.3476148545742035\n", + "Starting epoch 100\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.3615925127267838, Test loss: 0.3485625684261322\n", + "Starting epoch 101\n", + "Train loss: 0.360949804186821, Test loss: 0.3456147611141205\n", + "Starting epoch 102\n", + "Train loss: 0.3597677731513977, Test loss: 0.3517780900001526\n", + "Starting epoch 103\n", + "Train loss: 0.358520530462265, Test loss: 0.3670318126678467\n", + "Starting epoch 104\n", + "Train loss: 0.35772379279136657, Test loss: 0.3693113923072815\n", + "Starting epoch 105\n", + "Train loss: 0.35707773864269254, Test loss: 0.3689192533493042\n", + "Starting epoch 106\n", + "Train loss: 0.356054350733757, Test loss: 0.35626646876335144\n", + "Starting epoch 107\n", + "Train loss: 0.35419742882251737, Test loss: 0.3643738329410553\n", + "Starting epoch 108\n", + "Train loss: 0.3517108368873596, Test loss: 0.34108442068099976\n", + "Starting epoch 109\n", + "Train loss: 0.34856239438056946, Test loss: 0.3461962640285492\n", + "Starting epoch 110\n", + "Train loss: 0.3459764105081558, Test loss: 0.35031041502952576\n", + "Starting epoch 111\n", + "Train loss: 0.3438185805082321, Test loss: 0.35874438285827637\n", + "Starting epoch 112\n", + "Train loss: 0.34222314953804017, Test loss: 0.3674602806568146\n", + "Starting epoch 113\n", + "Train loss: 0.340915287733078, Test loss: 0.3607209622859955\n", + "Starting epoch 114\n", + "Train loss: 0.33990616858005523, Test loss: 0.3650568425655365\n", + "Starting epoch 115\n", + "Train loss: 0.3388424891233444, Test loss: 0.3538447320461273\n", + "Starting epoch 116\n", + "Train loss: 0.33792890429496764, Test loss: 0.3523561656475067\n", + "Starting epoch 117\n", + "Train loss: 0.3365205001831055, Test loss: 0.35455936193466187\n", + "Starting epoch 118\n", + "Train loss: 0.3356513339281082, Test loss: 0.34129443764686584\n", + "Starting epoch 119\n", + "Train loss: 0.3350427341461182, Test loss: 0.3435269296169281\n", + "Starting epoch 120\n", + "Train loss: 0.33450307488441466, Test loss: 0.330332487821579\n", + "Starting epoch 121\n", + "Train loss: 0.3340336638689041, Test loss: 0.32802361249923706\n", + "Starting epoch 122\n", + "Train loss: 0.33313876688480376, Test loss: 0.3177652060985565\n", + "Starting epoch 123\n", + "Train loss: 0.3319502919912338, Test loss: 0.30718594789505005\n", + "Starting epoch 124\n", + "Train loss: 0.33096094787120817, Test loss: 0.3020181953907013\n", + "Starting epoch 125\n", + "Train loss: 0.32926656544208527, Test loss: 0.28990015387535095\n", + "Starting epoch 126\n", + "Train loss: 0.327436158657074, Test loss: 0.3130899965763092\n", + "Starting epoch 127\n", + "Train loss: 0.32427490651607516, Test loss: 0.32735177874565125\n", + "Starting epoch 128\n", + "Train loss: 0.31834947526454926, Test loss: 0.27816635370254517\n", + "Starting epoch 129\n", + "Train loss: 0.31315451323986054, Test loss: 0.2699378728866577\n", + "Starting epoch 130\n", + "Train loss: 0.3105203354358673, Test loss: 0.27193304896354675\n", + "Starting epoch 131\n", + "Train loss: 0.30883385598659513, Test loss: 0.26706281304359436\n", + "Starting epoch 132\n", + "Train loss: 0.30730386316776276, Test loss: 0.2683414816856384\n", + "Starting epoch 133\n", + "Train loss: 0.30701243460178373, Test loss: 0.2637064754962921\n", + "Starting epoch 134\n", + "Train loss: 0.3051982343196869, Test loss: 0.26808422803878784\n", + "Starting epoch 135\n", + "Train loss: 0.30415072321891784, Test loss: 0.26493528485298157\n", + "Starting epoch 136\n", + "Train loss: 0.30303559243679046, Test loss: 0.2619319558143616\n", + "Starting epoch 137\n", + "Train loss: 0.3018393981456757, Test loss: 0.2698417901992798\n", + "Starting epoch 138\n", + "Train loss: 0.30107632398605344, Test loss: 0.26598772406578064\n", + "Starting epoch 139\n", + "Train loss: 0.29848443627357485, Test loss: 0.26501014828681946\n", + "Starting epoch 140\n", + "Train loss: 0.2965683615207672, Test loss: 0.26227492094039917\n", + "Starting epoch 141\n", + "Train loss: 0.2957805353403091, Test loss: 0.25835150480270386\n", + "Starting epoch 142\n", + "Train loss: 0.29465813338756563, Test loss: 0.2620556652545929\n", + "Starting epoch 143\n", + "Train loss: 0.29327681481838225, Test loss: 0.2559955418109894\n", + "Starting epoch 144\n", + "Train loss: 0.2929026138782501, Test loss: 0.2575532793998718\n", + "Starting epoch 145\n", + "Train loss: 0.2908797663450241, Test loss: 0.25934258103370667\n", + "Starting epoch 146\n", + "Train loss: 0.2910093778371811, Test loss: 0.24925024807453156\n", + "Starting epoch 147\n", + "Train loss: 0.28940152704715727, Test loss: 0.25451791286468506\n", + "Starting epoch 148\n", + "Train loss: 0.2876336318254471, Test loss: 0.2455560564994812\n", + "Starting epoch 149\n", + "Train loss: 0.28717496156692507, Test loss: 0.2441680133342743\n", + "Starting epoch 150\n", + "Train loss: 0.2873407232761383, Test loss: 0.25555479526519775\n", + "Starting epoch 151\n", + "Train loss: 0.2855211317539215, Test loss: 0.23133230209350586\n", + "Starting epoch 152\n", + "Train loss: 0.2849275368452072, Test loss: 0.23434993624687195\n", + "Starting epoch 153\n", + "Train loss: 0.28424556851387023, Test loss: 0.23823197185993195\n", + "Starting epoch 154\n", + "Train loss: 0.2834112709760666, Test loss: 0.23303920030593872\n", + "Starting epoch 155\n", + "Train loss: 0.2820774781703949, Test loss: 0.2295272946357727\n", + "Starting epoch 156\n", + "Train loss: 0.2817149406671524, Test loss: 0.21742184460163116\n", + "Starting epoch 157\n", + "Train loss: 0.2805761766433716, Test loss: 0.22222349047660828\n", + "Starting epoch 158\n", + "Train loss: 0.2789503717422485, Test loss: 0.21915017068386078\n", + "Starting epoch 159\n", + "Train loss: 0.2776974093914032, Test loss: 0.2119709998369217\n", + "Starting epoch 160\n", + "Train loss: 0.2766916364431381, Test loss: 0.21625111997127533\n", + "Starting epoch 161\n", + "Train loss: 0.2746465879678726, Test loss: 0.23054762184619904\n", + "Starting epoch 162\n", + "Train loss: 0.27253325045108795, Test loss: 0.2070249319076538\n", + "Starting epoch 163\n", + "Train loss: 0.2718994140625, Test loss: 0.20096664130687714\n", + "Starting epoch 164\n", + "Train loss: 0.268435914516449, Test loss: 0.207820862531662\n", + "Starting epoch 165\n", + "Train loss: 0.26696533024311064, Test loss: 0.2117065042257309\n", + "Starting epoch 166\n", + "Train loss: 0.26425688207149506, Test loss: 0.20515604317188263\n", + "Starting epoch 167\n", + "Train loss: 0.26234394490718843, Test loss: 0.198745459318161\n", + "Starting epoch 168\n", + "Train loss: 0.2596886545419693, Test loss: 0.22898657619953156\n", + "Starting epoch 169\n", + "Train loss: 0.2584247839450836, Test loss: 0.20092222094535828\n", + "Starting epoch 170\n", + "Train loss: 0.2569375044107437, Test loss: 0.20449469983577728\n", + "Starting epoch 171\n", + "Train loss: 0.2550519686937332, Test loss: 0.19834037125110626\n", + "Starting epoch 172\n", + "Train loss: 0.253098286986351, Test loss: 0.20250117778778076\n", + "Starting epoch 173\n", + "Train loss: 0.25211102664470675, Test loss: 0.19647932052612305\n", + "Starting epoch 174\n", + "Train loss: 0.2515099424123764, Test loss: 0.21112972497940063\n", + "Starting epoch 175\n", + "Train loss: 0.2504119625687599, Test loss: 0.19386066496372223\n", + "Starting epoch 176\n", + "Train loss: 0.24895010113716126, Test loss: 0.2130907028913498\n", + "Starting epoch 177\n", + "Train loss: 0.24912246465682983, Test loss: 0.1940905898809433\n", + "Starting epoch 178\n", + "Train loss: 0.24671282589435578, Test loss: 0.19885368645191193\n", + "Starting epoch 179\n", + "Train loss: 0.24714148461818694, Test loss: 0.19436392188072205\n", + "Starting epoch 180\n", + "Train loss: 0.24608005493879317, Test loss: 0.1929432451725006\n", + "Starting epoch 181\n", + "Train loss: 0.2455797716975212, Test loss: 0.19421835243701935\n", + "Starting epoch 182\n", + "Train loss: 0.24452740132808684, Test loss: 0.19389760494232178\n", + "Starting epoch 183\n", + "Train loss: 0.24365443080663682, Test loss: 0.19487427175045013\n", + "Starting epoch 184\n", + "Train loss: 0.24250098258256914, Test loss: 0.19073006510734558\n", + "Starting epoch 185\n", + "Train loss: 0.2424587419629097, Test loss: 0.1979159414768219\n", + "Starting epoch 186\n", + "Train loss: 0.2423720818758011, Test loss: 0.19988372921943665\n", + "Starting epoch 187\n", + "Train loss: 0.24155926167964936, Test loss: 0.19687522947788239\n", + "Starting epoch 188\n", + "Train loss: 0.2387770602107048, Test loss: 0.18796905875205994\n", + "Starting epoch 189\n", + "Train loss: 0.2373790943622589, Test loss: 0.19631412625312805\n", + "Starting epoch 190\n", + "Train loss: 0.2371345019340515, Test loss: 0.18998481333255768\n", + "Starting epoch 191\n", + "Train loss: 0.23618748664855957, Test loss: 0.1958928108215332\n", + "Starting epoch 192\n", + "Train loss: 0.23558850288391114, Test loss: 0.1968940943479538\n", + "Starting epoch 193\n", + "Train loss: 0.2348484778404236, Test loss: 0.19069167971611023\n", + "Starting epoch 194\n", + "Train loss: 0.2332456251978874, Test loss: 0.20346394181251526\n", + "Starting epoch 195\n", + "Train loss: 0.2312553867697716, Test loss: 0.18816417455673218\n", + "Starting epoch 196\n", + "Train loss: 0.23046134412288666, Test loss: 0.18882964551448822\n", + "Starting epoch 197\n", + "Train loss: 0.23092380732297899, Test loss: 0.18751730024814606\n", + "Starting epoch 198\n", + "Train loss: 0.22907230615615845, Test loss: 0.1912018358707428\n", + "Starting epoch 199\n", + "Train loss: 0.228268039226532, Test loss: 0.18831786513328552\n", + "Starting epoch 200\n", + "Train loss: 0.22888043344020845, Test loss: 0.19298307597637177\n", + "Starting epoch 201\n", + "Train loss: 0.2268107697367668, Test loss: 0.19016656279563904\n", + "Starting epoch 202\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.226170354783535, Test loss: 0.19757743179798126\n", + "Starting epoch 203\n", + "Train loss: 0.2263094651699066, Test loss: 0.1932748258113861\n", + "Starting epoch 204\n", + "Train loss: 0.2250983026623726, Test loss: 0.18528161942958832\n", + "Starting epoch 205\n", + "Train loss: 0.2240552070736885, Test loss: 0.1888120472431183\n", + "Starting epoch 206\n", + "Train loss: 0.22539281010627746, Test loss: 0.18865029513835907\n", + "Starting epoch 207\n", + "Train loss: 0.22416326105594636, Test loss: 0.19086670875549316\n", + "Starting epoch 208\n", + "Train loss: 0.2239014631509781, Test loss: 0.2122742235660553\n", + "Starting epoch 209\n", + "Train loss: 0.22274132013320924, Test loss: 0.1962028592824936\n", + "Starting epoch 210\n", + "Train loss: 0.22178714156150817, Test loss: 0.18812689185142517\n", + "Starting epoch 211\n", + "Train loss: 0.22248640090227126, Test loss: 0.19356156885623932\n", + "Starting epoch 212\n", + "Train loss: 0.22142658054828643, Test loss: 0.20196600258350372\n", + "Starting epoch 213\n", + "Train loss: 0.220247500538826, Test loss: 0.18714964389801025\n", + "Starting epoch 214\n", + "Train loss: 0.2204684191942215, Test loss: 0.19159483909606934\n", + "Starting epoch 215\n", + "Train loss: 0.21988464772701263, Test loss: 0.1968948096036911\n", + "Starting epoch 216\n", + "Train loss: 0.21826182901859284, Test loss: 0.18893302977085114\n", + "Starting epoch 217\n", + "Train loss: 0.21955573588609695, Test loss: 0.1920965313911438\n", + "Starting epoch 218\n", + "Train loss: 0.2159026536345482, Test loss: 0.1934739500284195\n", + "Starting epoch 219\n", + "Train loss: 0.2143270593881607, Test loss: 0.1905737668275833\n", + "Starting epoch 220\n", + "Train loss: 0.21389835566282273, Test loss: 0.19324296712875366\n", + "Starting epoch 221\n", + "Train loss: 0.21151452243328095, Test loss: 0.18695051968097687\n", + "Starting epoch 222\n", + "Train loss: 0.2108818197250366, Test loss: 0.21739336848258972\n", + "Starting epoch 223\n", + "Train loss: 0.2084437483549118, Test loss: 0.19282972812652588\n", + "Starting epoch 224\n", + "Train loss: 0.20771318465471267, Test loss: 0.18696852028369904\n", + "Starting epoch 225\n", + "Train loss: 0.20615222930908203, Test loss: 0.19475118815898895\n", + "Starting epoch 226\n", + "Train loss: 0.20596626430749893, Test loss: 0.20794351398944855\n", + "Starting epoch 227\n", + "Train loss: 0.20247509568929672, Test loss: 0.20480935275554657\n", + "Starting epoch 228\n", + "Train loss: 0.20246607899665833, Test loss: 0.18604542315006256\n", + "Starting epoch 229\n", + "Train loss: 0.19993168652057647, Test loss: 0.20623505115509033\n", + "Starting epoch 230\n", + "Train loss: 0.20176117599010468, Test loss: 0.1946200430393219\n", + "Starting epoch 231\n", + "Train loss: 0.19651556372642517, Test loss: 0.1870548129081726\n", + "Starting epoch 232\n", + "Train loss: 0.1940213891863823, Test loss: 0.18708297610282898\n", + "Starting epoch 233\n", + "Train loss: 0.1939871761202812, Test loss: 0.1923178881406784\n", + "Starting epoch 234\n", + "Train loss: 0.19208043158054353, Test loss: 0.19141574203968048\n", + "Starting epoch 235\n", + "Train loss: 0.19226059108972549, Test loss: 0.20162692666053772\n", + "Starting epoch 236\n", + "Train loss: 0.18880740225315093, Test loss: 0.19474506378173828\n", + "Starting epoch 237\n", + "Train loss: 0.1884711739420891, Test loss: 0.19052499532699585\n", + "Starting epoch 238\n", + "Train loss: 0.18785822868347168, Test loss: 0.19404906034469604\n", + "Starting epoch 239\n", + "Train loss: 0.18684806019067765, Test loss: 0.18999671936035156\n", + "Starting epoch 240\n", + "Train loss: 0.18590034633874894, Test loss: 0.19634442031383514\n", + "Starting epoch 241\n", + "Train loss: 0.18560695558786391, Test loss: 0.19241555035114288\n", + "Starting epoch 242\n", + "Train loss: 0.1842741674184799, Test loss: 0.1977449357509613\n", + "Starting epoch 243\n", + "Train loss: 0.18301920503377914, Test loss: 0.19183194637298584\n", + "Starting epoch 244\n", + "Train loss: 0.18272157222032548, Test loss: 0.19019274413585663\n", + "Starting epoch 245\n", + "Train loss: 0.1810527664422989, Test loss: 0.18875513970851898\n", + "Starting epoch 246\n", + "Train loss: 0.18160562545061112, Test loss: 0.20350609719753265\n", + "Starting epoch 247\n", + "Train loss: 0.18030618041753768, Test loss: 0.1917669028043747\n", + "Starting epoch 248\n", + "Train loss: 0.17955500841140748, Test loss: 0.18879148364067078\n", + "Starting epoch 249\n", + "Train loss: 0.17787596076726914, Test loss: 0.19158127903938293\n", + "Starting epoch 250\n", + "Train loss: 0.1783916038274765, Test loss: 0.19459550082683563\n", + "Starting epoch 251\n", + "Train loss: 0.17603518426418305, Test loss: 0.19500122964382172\n", + "Starting epoch 252\n", + "Train loss: 0.17501287668943405, Test loss: 0.19123691320419312\n", + "Starting epoch 253\n", + "Train loss: 0.17503537088632584, Test loss: 0.2067839652299881\n", + "Starting epoch 254\n", + "Train loss: 0.17465169221162796, Test loss: 0.20901896059513092\n", + "Starting epoch 255\n", + "Train loss: 0.17496959686279298, Test loss: 0.19958645105361938\n", + "Starting epoch 256\n", + "Train loss: 0.17415398925542833, Test loss: 0.1874791830778122\n", + "Starting epoch 257\n", + "Train loss: 0.1736619535088539, Test loss: 0.18778198957443237\n", + "Starting epoch 258\n", + "Train loss: 0.1741215717792511, Test loss: 0.19222326576709747\n", + "Starting epoch 259\n", + "Train loss: 0.17227981865406036, Test loss: 0.18782372772693634\n", + "Starting epoch 260\n", + "Train loss: 0.17370371729135514, Test loss: 0.18645474314689636\n", + "Starting epoch 261\n", + "Train loss: 0.17353013396263123, Test loss: 0.19302977621555328\n", + "Starting epoch 262\n", + "Train loss: 0.1729368296265602, Test loss: 0.18836218118667603\n", + "Starting epoch 263\n", + "Train loss: 0.17273843556642532, Test loss: 0.1850820779800415\n", + "Starting epoch 264\n", + "Train loss: 0.17207788318395614, Test loss: 0.1900264173746109\n", + "Starting epoch 265\n", + "Train loss: 0.17270654022693635, Test loss: 0.18756793439388275\n", + "Starting epoch 266\n", + "Train loss: 0.172070474922657, Test loss: 0.1888301521539688\n", + "Starting epoch 267\n", + "Train loss: 0.173347869515419, Test loss: 0.1852317899465561\n", + "Starting epoch 268\n", + "Train loss: 0.1719832420349121, Test loss: 0.1873391717672348\n", + "Starting epoch 269\n", + "Train loss: 0.17310505151748656, Test loss: 0.19378848373889923\n", + "Starting epoch 270\n", + "Train loss: 0.17168695092201233, Test loss: 0.1859746277332306\n", + "Starting epoch 271\n", + "Train loss: 0.1738535460829735, Test loss: 0.19274106621742249\n", + "Starting epoch 272\n", + "Train loss: 0.17137377411127092, Test loss: 0.18458017706871033\n", + "Starting epoch 273\n", + "Train loss: 0.17113147884607316, Test loss: 0.18703190982341766\n", + "Starting epoch 274\n", + "Train loss: 0.1716510945558548, Test loss: 0.19292591512203217\n", + "Starting epoch 275\n", + "Train loss: 0.17237029492855072, Test loss: 0.18362733721733093\n", + "Starting epoch 276\n", + "Train loss: 0.17243712544441223, Test loss: 0.18605509400367737\n", + "Starting epoch 277\n", + "Train loss: 0.17110251903533935, Test loss: 0.18839119374752045\n", + "Starting epoch 278\n", + "Train loss: 0.17201347708702086, Test loss: 0.187350332736969\n", + "Starting epoch 279\n", + "Train loss: 0.1724971503019333, Test loss: 0.1857621669769287\n", + "Starting epoch 280\n", + "Train loss: 0.17223135203123094, Test loss: 0.1869422197341919\n", + "Starting epoch 281\n", + "Train loss: 0.17255880892276765, Test loss: 0.18550829589366913\n", + "Starting epoch 282\n", + "Train loss: 0.1722767597436905, Test loss: 0.18420253694057465\n", + "Starting epoch 283\n", + "Train loss: 0.17177167564630508, Test loss: 0.19434227049350739\n", + "Starting epoch 284\n", + "Train loss: 0.17144422173500062, Test loss: 0.19593821465969086\n", + "Starting epoch 285\n", + "Train loss: 0.17192387074232102, Test loss: 0.1930590718984604\n", + "Starting epoch 286\n", + "Train loss: 0.17229140490293504, Test loss: 0.18871724605560303\n", + "Starting epoch 287\n", + "Train loss: 0.1717612498998642, Test loss: 0.194051131606102\n", + "Starting epoch 288\n", + "Train loss: 0.17141373932361603, Test loss: 0.18322749435901642\n", + "Starting epoch 289\n", + "Train loss: 0.17165308326482773, Test loss: 0.18401315808296204\n", + "Starting epoch 290\n", + "Train loss: 0.17099475622177124, Test loss: 0.19526781141757965\n", + "Starting epoch 291\n", + "Train loss: 0.1720897814631462, Test loss: 0.1918928474187851\n", + "Starting epoch 292\n", + "Train loss: 0.1715096914768219, Test loss: 0.18732716143131256\n", + "Starting epoch 293\n", + "Train loss: 0.1709428024291992, Test loss: 0.1894780993461609\n", + "Starting epoch 294\n", + "Train loss: 0.1727566021680832, Test loss: 0.185350239276886\n", + "Starting epoch 295\n", + "Train loss: 0.17108700841665267, Test loss: 0.18931671977043152\n", + "Starting epoch 296\n", + "Train loss: 0.1721758434176445, Test loss: 0.18571801483631134\n", + "Starting epoch 297\n", + "Train loss: 0.17158415764570237, Test loss: 0.18310542404651642\n", + "Starting epoch 298\n", + "Train loss: 0.17136700242757796, Test loss: 0.1883663833141327\n", + "Starting epoch 299\n", + "Train loss: 0.1719487887620926, Test loss: 0.18734322488307953\n", + "Starting epoch 300\n", + "Train loss: 0.17238999605178834, Test loss: 0.19066686928272247\n", + "Starting epoch 301\n", + "Train loss: 0.17148099213838577, Test loss: 0.18424846231937408\n", + "Starting epoch 302\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.1726206985116005, Test loss: 0.18244457244873047\n", + "Starting epoch 303\n", + "Train loss: 0.17137669563293456, Test loss: 0.1860610395669937\n", + "Starting epoch 304\n", + "Train loss: 0.17136358737945556, Test loss: 0.18499557673931122\n", + "Starting epoch 305\n", + "Train loss: 0.17148921489715577, Test loss: 0.18568895757198334\n", + "Starting epoch 306\n", + "Train loss: 0.1722726747393608, Test loss: 0.18673449754714966\n", + "Starting epoch 307\n", + "Train loss: 0.17086882948875426, Test loss: 0.18590937554836273\n", + "Starting epoch 308\n", + "Train loss: 0.17143023937940596, Test loss: 0.18446169793605804\n", + "Starting epoch 309\n", + "Train loss: 0.17210818707942963, Test loss: 0.19084210693836212\n", + "Starting epoch 310\n", + "Train loss: 0.17039875596761703, Test loss: 0.18437479436397552\n", + "Starting epoch 311\n", + "Train loss: 0.1705440977215767, Test loss: 0.19254913926124573\n", + "Starting epoch 312\n", + "Train loss: 0.17090477854013442, Test loss: 0.19732363522052765\n", + "Starting epoch 313\n", + "Train loss: 0.17129944980144501, Test loss: 0.19303855299949646\n", + "Starting epoch 314\n", + "Train loss: 0.17070741325616837, Test loss: 0.18860496580600739\n", + "Starting epoch 315\n", + "Train loss: 0.17201438486576082, Test loss: 0.18553651869297028\n", + "Starting epoch 316\n", + "Train loss: 0.17144148349761962, Test loss: 0.18691761791706085\n", + "Starting epoch 317\n", + "Train loss: 0.17169346541166305, Test loss: 0.18569807708263397\n", + "Starting epoch 318\n", + "Train loss: 0.17095892935991286, Test loss: 0.18943102657794952\n", + "Starting epoch 319\n", + "Train loss: 0.17170984506607057, Test loss: 0.18959937989711761\n", + "Starting epoch 320\n", + "Train loss: 0.17084970712661743, Test loss: 0.1900557577610016\n", + "Starting epoch 321\n", + "Train loss: 0.17162291795015336, Test loss: 0.18917962908744812\n", + "Starting epoch 322\n", + "Train loss: 0.17157352924346925, Test loss: 0.19080272316932678\n", + "Starting epoch 323\n", + "Train loss: 0.1710421434044838, Test loss: 0.18562042713165283\n", + "Starting epoch 324\n", + "Train loss: 0.17083582907915115, Test loss: 0.1953287571668625\n", + "Starting epoch 325\n", + "Train loss: 0.17124094009399415, Test loss: 0.1974295973777771\n", + "Starting epoch 326\n", + "Train loss: 0.1721392148733139, Test loss: 0.18370240926742554\n", + "Starting epoch 327\n", + "Train loss: 0.17096995085477829, Test loss: 0.1868620216846466\n", + "Starting epoch 328\n", + "Train loss: 0.1707344573736191, Test loss: 0.18702593445777893\n", + "Starting epoch 329\n", + "Train loss: 0.17211799144744874, Test loss: 0.1975221186876297\n", + "Starting epoch 330\n", + "Train loss: 0.17140806406736375, Test loss: 0.18698899447917938\n", + "Starting epoch 331\n", + "Train loss: 0.1705625656247139, Test loss: 0.18764640390872955\n", + "Starting epoch 332\n", + "Train loss: 0.17150658518075942, Test loss: 0.19443577527999878\n", + "Starting epoch 333\n", + "Train loss: 0.1707953169941902, Test loss: 0.1886487454175949\n", + "Starting epoch 334\n", + "Train loss: 0.17094042509794236, Test loss: 0.18783622980117798\n", + "Starting epoch 335\n", + "Train loss: 0.17057715922594072, Test loss: 0.18691116571426392\n", + "Starting epoch 336\n", + "Train loss: 0.17068623274564743, Test loss: 0.18609309196472168\n", + "Starting epoch 337\n", + "Train loss: 0.17265334695577622, Test loss: 0.18887607753276825\n", + "Starting epoch 338\n", + "Train loss: 0.17150324136018752, Test loss: 0.18672215938568115\n", + "Starting epoch 339\n", + "Train loss: 0.17008207619190216, Test loss: 0.18539071083068848\n", + "Starting epoch 340\n", + "Train loss: 0.17064897924661637, Test loss: 0.1890142261981964\n", + "Starting epoch 341\n", + "Train loss: 0.17101783573627471, Test loss: 0.18806703388690948\n", + "Starting epoch 342\n", + "Train loss: 0.17048245191574096, Test loss: 0.1840284764766693\n", + "Starting epoch 343\n", + "Train loss: 0.17056311696767806, Test loss: 0.18859004974365234\n", + "Starting epoch 344\n", + "Train loss: 0.1707879850268364, Test loss: 0.1851024478673935\n", + "Starting epoch 345\n", + "Train loss: 0.17134592205286026, Test loss: 0.1864549070596695\n", + "Starting epoch 346\n", + "Train loss: 0.17070788830518724, Test loss: 0.20096947252750397\n", + "Starting epoch 347\n", + "Train loss: 0.17126452594995498, Test loss: 0.1892734169960022\n", + "Starting epoch 348\n", + "Train loss: 0.17011507868766784, Test loss: 0.19218410551548004\n", + "Starting epoch 349\n", + "Train loss: 0.17034972906112672, Test loss: 0.18709351122379303\n", + "Starting epoch 350\n", + "Train loss: 0.17268934428691865, Test loss: 0.1909652054309845\n", + "Starting epoch 351\n", + "Train loss: 0.1704642304778099, Test loss: 0.1888791173696518\n", + "Starting epoch 352\n", + "Train loss: 0.17010183572769166, Test loss: 0.18746891617774963\n", + "Starting epoch 353\n", + "Train loss: 0.17130182057619095, Test loss: 0.18656663596630096\n", + "Starting epoch 354\n", + "Train loss: 0.17090913116931916, Test loss: 0.18733642995357513\n", + "Starting epoch 355\n", + "Train loss: 0.1707014873623848, Test loss: 0.18988047540187836\n", + "Starting epoch 356\n", + "Train loss: 0.17085427135229111, Test loss: 0.18553310632705688\n", + "Starting epoch 357\n", + "Train loss: 0.17035568118095398, Test loss: 0.1877690702676773\n", + "Starting epoch 358\n", + "Train loss: 0.17045283794403077, Test loss: 0.1958836168050766\n", + "Starting epoch 359\n", + "Train loss: 0.17038139939308167, Test loss: 0.19042226672172546\n", + "Starting epoch 360\n", + "Train loss: 0.17073548913002015, Test loss: 0.19846461713314056\n", + "Starting epoch 361\n", + "Train loss: 0.17042287409305573, Test loss: 0.18682710826396942\n", + "Starting epoch 362\n", + "Train loss: 0.1707505801320076, Test loss: 0.1889892816543579\n", + "Starting epoch 363\n", + "Train loss: 0.17098560333251953, Test loss: 0.18472091853618622\n", + "Starting epoch 364\n", + "Train loss: 0.16993928015232085, Test loss: 0.18573367595672607\n", + "Starting epoch 365\n", + "Train loss: 0.1707867392897606, Test loss: 0.18595075607299805\n", + "Starting epoch 366\n", + "Train loss: 0.17154475867748262, Test loss: 0.19007089734077454\n", + "Starting epoch 367\n", + "Train loss: 0.17094133913516998, Test loss: 0.18879377841949463\n", + "Starting epoch 368\n", + "Train loss: 0.1699771335721016, Test loss: 0.1984555423259735\n", + "Starting epoch 369\n", + "Train loss: 0.1702322906255722, Test loss: 0.18653932213783264\n", + "Starting epoch 370\n", + "Train loss: 0.17012394934892655, Test loss: 0.21055848896503448\n", + "Starting epoch 371\n", + "Train loss: 0.17010463237762452, Test loss: 0.18859103322029114\n", + "Starting epoch 372\n", + "Train loss: 0.1711147955060005, Test loss: 0.18659332394599915\n", + "Starting epoch 373\n", + "Train loss: 0.1706847381591797, Test loss: 0.18638841807842255\n", + "Starting epoch 374\n", + "Train loss: 0.17017956346273422, Test loss: 0.18888545036315918\n", + "Starting epoch 375\n", + "Train loss: 0.16955278784036637, Test loss: 0.19258062541484833\n", + "Starting epoch 376\n", + "Train loss: 0.16962166607379914, Test loss: 0.18992814421653748\n", + "Starting epoch 377\n", + "Train loss: 0.1712900471687317, Test loss: 0.20023097097873688\n", + "Starting epoch 378\n", + "Train loss: 0.1694560667872429, Test loss: 0.1908327341079712\n", + "Starting epoch 379\n", + "Train loss: 0.17048790454864501, Test loss: 0.20690859854221344\n", + "Starting epoch 380\n", + "Train loss: 0.16980720013380052, Test loss: 0.19028539955615997\n", + "Starting epoch 381\n", + "Train loss: 0.17121654003858566, Test loss: 0.18781714141368866\n", + "Starting epoch 382\n", + "Train loss: 0.17077478706836702, Test loss: 0.1890014261007309\n", + "Starting epoch 383\n", + "Train loss: 0.16984086900949477, Test loss: 0.19108866155147552\n", + "Starting epoch 384\n", + "Train loss: 0.16951833695173263, Test loss: 0.18658150732517242\n", + "Starting epoch 385\n", + "Train loss: 0.16994521349668504, Test loss: 0.18681561946868896\n", + "Starting epoch 386\n", + "Train loss: 0.16969992846250534, Test loss: 0.19606010615825653\n", + "Starting epoch 387\n", + "Train loss: 0.17178470075130461, Test loss: 0.1905061900615692\n", + "Starting epoch 388\n", + "Train loss: 0.16941530078649522, Test loss: 0.19096235930919647\n", + "Starting epoch 389\n", + "Train loss: 0.16986454904079437, Test loss: 0.18964743614196777\n", + "Starting epoch 390\n", + "Train loss: 0.16916147708892823, Test loss: 0.1881306916475296\n", + "Starting epoch 391\n", + "Train loss: 0.16962518364191057, Test loss: 0.19268257915973663\n", + "Starting epoch 392\n", + "Train loss: 0.17001355171203614, Test loss: 0.1870252788066864\n", + "Starting epoch 393\n", + "Train loss: 0.16977314919233322, Test loss: 0.1893605887889862\n", + "Starting epoch 394\n", + "Train loss: 0.17078170627355577, Test loss: 0.18894703686237335\n", + "Starting epoch 395\n", + "Train loss: 0.16997762501239777, Test loss: 0.19744127988815308\n", + "Starting epoch 396\n", + "Train loss: 0.1695681145787239, Test loss: 0.19543875753879547\n", + "Starting epoch 397\n", + "Train loss: 0.17116981893777847, Test loss: 0.1898685097694397\n", + "Starting epoch 398\n", + "Train loss: 0.16978763431310653, Test loss: 0.19336414337158203\n", + "Starting epoch 399\n", + "Train loss: 0.1698060476779938, Test loss: 0.1902143806219101\n", + "Starting epoch 400\n", + "Train loss: 0.16989856481552124, Test loss: 0.18639028072357178\n", + "Starting epoch 401\n", + "Train loss: 0.16916876941919326, Test loss: 0.19080232083797455\n", + "Starting epoch 402\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.1694898158311844, Test loss: 0.19997504353523254\n", + "Starting epoch 403\n", + "Train loss: 0.16911037832498552, Test loss: 0.19961464405059814\n", + "Starting epoch 404\n", + "Train loss: 0.16958735078573228, Test loss: 0.19277828931808472\n", + "Starting epoch 405\n", + "Train loss: 0.1692374262213707, Test loss: 0.1948060244321823\n", + "Starting epoch 406\n", + "Train loss: 0.16987987339496613, Test loss: 0.1913715898990631\n", + "Starting epoch 407\n", + "Train loss: 0.1702338570356369, Test loss: 0.19077181816101074\n", + "Starting epoch 408\n", + "Train loss: 0.17087660282850264, Test loss: 0.19069130718708038\n", + "Starting epoch 409\n", + "Train loss: 0.17050980180501937, Test loss: 0.18942345678806305\n", + "Starting epoch 410\n", + "Train loss: 0.16898825228214265, Test loss: 0.18846383690834045\n", + "Starting epoch 411\n", + "Train loss: 0.16859743058681487, Test loss: 0.19882498681545258\n", + "Starting epoch 412\n", + "Train loss: 0.16888077616691588, Test loss: 0.19553212821483612\n", + "Starting epoch 413\n", + "Train loss: 0.17034269452095033, Test loss: 0.206274151802063\n", + "Starting epoch 414\n", + "Train loss: 0.16967749387025832, Test loss: 0.18993180990219116\n", + "Starting epoch 415\n", + "Train loss: 0.16969330430030824, Test loss: 0.20225104689598083\n", + "Starting epoch 416\n", + "Train loss: 0.1692409771680832, Test loss: 0.19181497395038605\n", + "Starting epoch 417\n", + "Train loss: 0.1697179964184761, Test loss: 0.1962883323431015\n", + "Starting epoch 418\n", + "Train loss: 0.16907157748937607, Test loss: 0.18717719614505768\n", + "Starting epoch 419\n", + "Train loss: 0.16976594656705857, Test loss: 0.1979045569896698\n", + "Starting epoch 420\n", + "Train loss: 0.16925403624773025, Test loss: 0.19380821287631989\n", + "Starting epoch 421\n", + "Train loss: 0.1693799751996994, Test loss: 0.19541190564632416\n", + "Starting epoch 422\n", + "Train loss: 0.1689588287472725, Test loss: 0.1950978934764862\n", + "Starting epoch 423\n", + "Train loss: 0.16992175728082656, Test loss: 0.20567117631435394\n", + "Starting epoch 424\n", + "Train loss: 0.16942392259836198, Test loss: 0.18954885005950928\n", + "Starting epoch 425\n", + "Train loss: 0.16963367342948912, Test loss: 0.20618554949760437\n", + "Starting epoch 426\n", + "Train loss: 0.16971721529960632, Test loss: 0.1919717937707901\n", + "Starting epoch 427\n", + "Train loss: 0.16914659291505812, Test loss: 0.1906914860010147\n", + "Starting epoch 428\n", + "Train loss: 0.16966362446546554, Test loss: 0.18915702402591705\n", + "Starting epoch 429\n", + "Train loss: 0.16866327911615372, Test loss: 0.1926620453596115\n", + "Starting epoch 430\n", + "Train loss: 0.16991084724664687, Test loss: 0.19409731030464172\n", + "Starting epoch 431\n", + "Train loss: 0.16934476733207704, Test loss: 0.19636431336402893\n", + "Starting epoch 432\n", + "Train loss: 0.1688056281208992, Test loss: 0.1910889744758606\n", + "Starting epoch 433\n", + "Train loss: 0.1693705877661705, Test loss: 0.1948755979537964\n", + "Starting epoch 434\n", + "Train loss: 0.16931391537189483, Test loss: 0.1999613493680954\n", + "Starting epoch 435\n", + "Train loss: 0.1695335900783539, Test loss: 0.194841206073761\n", + "Starting epoch 436\n", + "Train loss: 0.1700305598974228, Test loss: 0.19535671174526215\n", + "Starting epoch 437\n", + "Train loss: 0.16968213886022568, Test loss: 0.19594474136829376\n", + "Starting epoch 438\n", + "Train loss: 0.16885388135910034, Test loss: 0.1922580748796463\n", + "Starting epoch 439\n", + "Train loss: 0.16920395731925963, Test loss: 0.19220267236232758\n", + "Starting epoch 440\n", + "Train loss: 0.1694079977273941, Test loss: 0.19381362199783325\n", + "Starting epoch 441\n", + "Train loss: 0.1689172062277794, Test loss: 0.193503737449646\n", + "Starting epoch 442\n", + "Train loss: 0.1686201027035713, Test loss: 0.19681327044963837\n", + "Starting epoch 443\n", + "Train loss: 0.16932022094726562, Test loss: 0.19814977049827576\n", + "Starting epoch 444\n", + "Train loss: 0.16890739887952805, Test loss: 0.20021498203277588\n", + "Starting epoch 445\n", + "Train loss: 0.17010682672262192, Test loss: 0.2117123305797577\n", + "Starting epoch 446\n", + "Train loss: 0.16973287254571914, Test loss: 0.19843752682209015\n", + "Starting epoch 447\n", + "Train loss: 0.16874275773763656, Test loss: 0.18946325778961182\n", + "Starting epoch 448\n", + "Train loss: 0.16886014521121978, Test loss: 0.19438347220420837\n", + "Starting epoch 449\n", + "Train loss: 0.16947703510522844, Test loss: 0.19106057286262512\n", + "Starting epoch 450\n", + "Train loss: 0.1689303570985794, Test loss: 0.19079484045505524\n", + "Starting epoch 451\n", + "Train loss: 0.16996497929096221, Test loss: 0.1940096616744995\n", + "Starting epoch 452\n", + "Train loss: 0.16916854530572892, Test loss: 0.19142676889896393\n", + "Starting epoch 453\n", + "Train loss: 0.16968849569559097, Test loss: 0.2145235687494278\n", + "Starting epoch 454\n", + "Train loss: 0.16960569500923156, Test loss: 0.20149320363998413\n", + "Starting epoch 455\n", + "Train loss: 0.16933610647916794, Test loss: 0.19524750113487244\n", + "Starting epoch 456\n", + "Train loss: 0.16953387916088103, Test loss: 0.20609687268733978\n", + "Starting epoch 457\n", + "Train loss: 0.1696011558175087, Test loss: 0.19212770462036133\n", + "Starting epoch 458\n", + "Train loss: 0.16911363750696182, Test loss: 0.1915046125650406\n", + "Starting epoch 459\n", + "Train loss: 0.16910911500453948, Test loss: 0.1994849443435669\n", + "Starting epoch 460\n", + "Train loss: 0.16853979885578155, Test loss: 0.1976967304944992\n", + "Starting epoch 461\n", + "Train loss: 0.16860851109027863, Test loss: 0.19648408889770508\n", + "Starting epoch 462\n", + "Train loss: 0.16867861330509185, Test loss: 0.19897958636283875\n", + "Starting epoch 463\n", + "Train loss: 0.1689225658774376, Test loss: 0.2091146856546402\n", + "Starting epoch 464\n", + "Train loss: 0.16916196942329406, Test loss: 0.19547875225543976\n", + "Starting epoch 465\n", + "Train loss: 0.1690514838695526, Test loss: 0.20818842947483063\n", + "Starting epoch 466\n", + "Train loss: 0.16942438006401062, Test loss: 0.1921643614768982\n", + "Starting epoch 467\n", + "Train loss: 0.16860277086496353, Test loss: 0.20315751433372498\n", + "Starting epoch 468\n", + "Train loss: 0.1688495969772339, Test loss: 0.19763830304145813\n", + "Starting epoch 469\n", + "Train loss: 0.16866845399141311, Test loss: 0.19396644830703735\n", + "Starting epoch 470\n", + "Train loss: 0.1691959458589554, Test loss: 0.19877910614013672\n", + "Starting epoch 471\n", + "Train loss: 0.16923220783472062, Test loss: 0.1964140683412552\n", + "Starting epoch 472\n", + "Train loss: 0.16877979516983033, Test loss: 0.19972383975982666\n", + "Starting epoch 473\n", + "Train loss: 0.1692442375421524, Test loss: 0.19583149254322052\n", + "Starting epoch 474\n", + "Train loss: 0.1690359055995941, Test loss: 0.20555201172828674\n", + "Starting epoch 475\n", + "Train loss: 0.16970975905656815, Test loss: 0.19809597730636597\n", + "Starting epoch 476\n", + "Train loss: 0.16908167839050292, Test loss: 0.19495148956775665\n", + "Starting epoch 477\n", + "Train loss: 0.1693318244814873, Test loss: 0.19385607540607452\n", + "Starting epoch 478\n", + "Train loss: 0.16996412813663483, Test loss: 0.19155463576316833\n", + "Starting epoch 479\n", + "Train loss: 0.16910285234451294, Test loss: 0.20525521039962769\n", + "Starting epoch 480\n", + "Train loss: 0.16868854105472564, Test loss: 0.20686769485473633\n", + "Starting epoch 481\n", + "Train loss: 0.16933598905801772, Test loss: 0.1934356987476349\n", + "Starting epoch 482\n", + "Train loss: 0.16931673556566237, Test loss: 0.19178155064582825\n", + "Starting epoch 483\n", + "Train loss: 0.16891455054283142, Test loss: 0.19935844838619232\n", + "Starting epoch 484\n", + "Train loss: 0.16847745358943939, Test loss: 0.2004496455192566\n", + "Starting epoch 485\n", + "Train loss: 0.16889021247625352, Test loss: 0.20358671247959137\n", + "Starting epoch 486\n", + "Train loss: 0.16924903750419618, Test loss: 0.21896667778491974\n", + "Starting epoch 487\n", + "Train loss: 0.16883585780858992, Test loss: 0.20147821307182312\n", + "Starting epoch 488\n", + "Train loss: 0.1688630139827728, Test loss: 0.1964028775691986\n", + "Starting epoch 489\n", + "Train loss: 0.1688866940140724, Test loss: 0.19105759263038635\n", + "Starting epoch 490\n", + "Train loss: 0.16864112317562102, Test loss: 0.19316360354423523\n", + "Starting epoch 491\n", + "Train loss: 0.1684379830956459, Test loss: 0.19306102395057678\n", + "Starting epoch 492\n", + "Train loss: 0.16893732607364653, Test loss: 0.21290551126003265\n", + "Starting epoch 493\n", + "Train loss: 0.16903932958841325, Test loss: 0.191924050450325\n", + "Starting epoch 494\n", + "Train loss: 0.16885657399892806, Test loss: 0.19468681514263153\n", + "Starting epoch 495\n", + "Train loss: 0.16933515548706055, Test loss: 0.2035944014787674\n", + "Starting epoch 496\n", + "Train loss: 0.16885664492845534, Test loss: 0.19699591398239136\n", + "Starting epoch 497\n", + "Train loss: 0.16895090758800507, Test loss: 0.19383950531482697\n", + "Starting epoch 498\n", + "Train loss: 0.16962749660015106, Test loss: 0.19862838089466095\n", + "Starting epoch 499\n", + "Train loss: 0.16937126994132995, Test loss: 0.20406825840473175\n", + "Starting epoch 500\n", + "Train loss: 0.16886860281229019, Test loss: 0.20004841685295105\n", + "Starting epoch 501\n", + "Train loss: 0.1685245555639267, Test loss: 0.2017078548669815\n", + "Starting epoch 502\n", + "Train loss: 0.16852885335683823, Test loss: 0.20094707608222961\n", + "Starting epoch 503\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.16846682727336884, Test loss: 0.19414518773555756\n", + "Starting epoch 504\n", + "Train loss: 0.16874067693948747, Test loss: 0.19661927223205566\n", + "Starting epoch 505\n", + "Train loss: 0.1693406680226326, Test loss: 0.19952493906021118\n", + "Starting epoch 506\n", + "Train loss: 0.16819078117609024, Test loss: 0.1953514665365219\n", + "Starting epoch 507\n", + "Train loss: 0.16850033402442932, Test loss: 0.1949404627084732\n", + "Starting epoch 508\n", + "Train loss: 0.168498252928257, Test loss: 0.1945393830537796\n", + "Starting epoch 509\n", + "Train loss: 0.16857060611248018, Test loss: 0.2005159705877304\n", + "Starting epoch 510\n", + "Train loss: 0.16906868666410446, Test loss: 0.19718162715435028\n", + "Starting epoch 511\n", + "Train loss: 0.16853864789009093, Test loss: 0.19784502685070038\n", + "Starting epoch 512\n", + "Train loss: 0.16868704706430435, Test loss: 0.2054259479045868\n", + "Starting epoch 513\n", + "Train loss: 0.1682452976703644, Test loss: 0.1955898553133011\n", + "Starting epoch 514\n", + "Train loss: 0.16844956159591676, Test loss: 0.2144092619419098\n", + "Starting epoch 515\n", + "Train loss: 0.16885142803192138, Test loss: 0.20444385707378387\n", + "Starting epoch 516\n", + "Train loss: 0.16784779816865922, Test loss: 0.20354369282722473\n", + "Starting epoch 517\n", + "Train loss: 0.16752301633358002, Test loss: 0.19615478813648224\n", + "Starting epoch 518\n", + "Train loss: 0.16759277492761612, Test loss: 0.2112082540988922\n", + "Starting epoch 519\n", + "Train loss: 0.1681793922185898, Test loss: 0.1935802847146988\n", + "Starting epoch 520\n", + "Train loss: 0.1676567679643631, Test loss: 0.2015613615512848\n", + "Starting epoch 521\n", + "Train loss: 0.1664031305909157, Test loss: 0.1960364431142807\n", + "Starting epoch 522\n", + "Train loss: 0.1665572389960289, Test loss: 0.19234086573123932\n", + "Starting epoch 523\n", + "Train loss: 0.16663130193948747, Test loss: 0.20318424701690674\n", + "Starting epoch 524\n", + "Train loss: 0.16674590110778809, Test loss: 0.1928710639476776\n", + "Starting epoch 525\n", + "Train loss: 0.16579206377267838, Test loss: 0.19506476819515228\n", + "Starting epoch 526\n", + "Train loss: 0.1656011387705803, Test loss: 0.19800904393196106\n", + "Starting epoch 527\n", + "Train loss: 0.16525678783655168, Test loss: 0.19851315021514893\n", + "Starting epoch 528\n", + "Train loss: 0.16591049045324324, Test loss: 0.19446192681789398\n", + "Starting epoch 529\n", + "Train loss: 0.1650242352485657, Test loss: 0.1976049244403839\n", + "Starting epoch 530\n", + "Train loss: 0.16477370172739028, Test loss: 0.18674075603485107\n", + "Starting epoch 531\n", + "Train loss: 0.1646990492939949, Test loss: 0.1903313249349594\n", + "Starting epoch 532\n", + "Train loss: 0.16528427898883818, Test loss: 0.19113557040691376\n", + "Starting epoch 533\n", + "Train loss: 0.16488151997327805, Test loss: 0.18745774030685425\n", + "Starting epoch 534\n", + "Train loss: 0.1644889909029007, Test loss: 0.188883438706398\n", + "Starting epoch 535\n", + "Train loss: 0.1652061641216278, Test loss: 0.18769241869449615\n", + "Starting epoch 536\n", + "Train loss: 0.16504257380962373, Test loss: 0.1934484839439392\n", + "Starting epoch 537\n", + "Train loss: 0.16484984934329985, Test loss: 0.19047360122203827\n", + "Starting epoch 538\n", + "Train loss: 0.16485207945108413, Test loss: 0.18776921927928925\n", + "Starting epoch 539\n", + "Train loss: 0.16458357632160187, Test loss: 0.18813550472259521\n", + "Starting epoch 540\n", + "Train loss: 0.16465455502271653, Test loss: 0.18796396255493164\n", + "Starting epoch 541\n", + "Train loss: 0.16463326305150985, Test loss: 0.18623381853103638\n", + "Starting epoch 542\n", + "Train loss: 0.16437036782503128, Test loss: 0.18703152239322662\n", + "Starting epoch 543\n", + "Train loss: 0.16477832436561585, Test loss: 0.19073638319969177\n", + "Starting epoch 544\n", + "Train loss: 0.16403819531202316, Test loss: 0.1909957379102707\n", + "Starting epoch 545\n", + "Train loss: 0.16488976091146468, Test loss: 0.19517925381660461\n", + "Starting epoch 546\n", + "Train loss: 0.16457963198423387, Test loss: 0.18767185509204865\n", + "Starting epoch 547\n", + "Train loss: 0.16388028085231782, Test loss: 0.18587924540042877\n", + "Starting epoch 548\n", + "Train loss: 0.16450824290513993, Test loss: 0.19012416899204254\n", + "Starting epoch 549\n", + "Train loss: 0.16413818478584288, Test loss: 0.19416335225105286\n", + "Starting epoch 550\n", + "Train loss: 0.1638542714715004, Test loss: 0.18862250447273254\n", + "Starting epoch 551\n", + "Train loss: 0.16455253601074218, Test loss: 0.19007079303264618\n", + "Starting epoch 552\n", + "Train loss: 0.16408680856227875, Test loss: 0.18673980236053467\n", + "Starting epoch 553\n", + "Train loss: 0.1639355516433716, Test loss: 0.19180133938789368\n", + "Starting epoch 554\n", + "Train loss: 0.1635944190621376, Test loss: 0.19987687468528748\n", + "Starting epoch 555\n", + "Train loss: 0.16366616696119307, Test loss: 0.18791121244430542\n", + "Starting epoch 556\n", + "Train loss: 0.16388145565986634, Test loss: 0.1870172917842865\n", + "Starting epoch 557\n", + "Train loss: 0.1647382864356041, Test loss: 0.1918330043554306\n", + "Starting epoch 558\n", + "Train loss: 0.16434544026851655, Test loss: 0.19191020727157593\n", + "Starting epoch 559\n", + "Train loss: 0.1641173070669174, Test loss: 0.1936737596988678\n", + "Starting epoch 560\n", + "Train loss: 0.16372631937265397, Test loss: 0.1844886690378189\n", + "Starting epoch 561\n", + "Train loss: 0.1637727975845337, Test loss: 0.1962832510471344\n", + "Starting epoch 562\n", + "Train loss: 0.1630113333463669, Test loss: 0.18592050671577454\n", + "Starting epoch 563\n", + "Train loss: 0.16343995809555054, Test loss: 0.184577077627182\n", + "Starting epoch 564\n", + "Train loss: 0.16372302711009978, Test loss: 0.1842525601387024\n", + "Starting epoch 565\n", + "Train loss: 0.16338506191968918, Test loss: 0.1880921721458435\n", + "Starting epoch 566\n", + "Train loss: 0.16352534353733061, Test loss: 0.18897394835948944\n", + "Starting epoch 567\n", + "Train loss: 0.16300972193479538, Test loss: 0.1853771209716797\n", + "Starting epoch 568\n", + "Train loss: 0.16349260866642, Test loss: 0.18614248931407928\n", + "Starting epoch 569\n", + "Train loss: 0.16334152787923814, Test loss: 0.1854103058576584\n", + "Starting epoch 570\n", + "Train loss: 0.16338702231645585, Test loss: 0.18471954762935638\n", + "Starting epoch 571\n", + "Train loss: 0.1631396621465683, Test loss: 0.1852698177099228\n", + "Starting epoch 572\n", + "Train loss: 0.16293839514255523, Test loss: 0.18907037377357483\n", + "Starting epoch 573\n", + "Train loss: 0.16318581640720367, Test loss: 0.19094617664813995\n", + "Starting epoch 574\n", + "Train loss: 0.16273848891258239, Test loss: 0.1862320750951767\n", + "Starting epoch 575\n", + "Train loss: 0.16347822457551955, Test loss: 0.18765424191951752\n", + "Starting epoch 576\n", + "Train loss: 0.16413963794708253, Test loss: 0.19156762957572937\n", + "Starting epoch 577\n", + "Train loss: 0.16362814754247665, Test loss: 0.18440885841846466\n", + "Starting epoch 578\n", + "Train loss: 0.1623674988746643, Test loss: 0.18311220407485962\n", + "Starting epoch 579\n", + "Train loss: 0.16225864887237548, Test loss: 0.18373414874076843\n", + "Starting epoch 580\n", + "Train loss: 0.16296748131513594, Test loss: 0.18832074105739594\n", + "Starting epoch 581\n", + "Train loss: 0.16291239053010942, Test loss: 0.1831580549478531\n", + "Starting epoch 582\n", + "Train loss: 0.16313182830810546, Test loss: 0.18268856406211853\n", + "Starting epoch 583\n", + "Train loss: 0.1624886080622673, Test loss: 0.18577204644680023\n", + "Starting epoch 584\n", + "Train loss: 0.16237639456987382, Test loss: 0.18601743876934052\n", + "Starting epoch 585\n", + "Train loss: 0.1627075916528702, Test loss: 0.18349918723106384\n", + "Starting epoch 586\n", + "Train loss: 0.16233968019485473, Test loss: 0.18494808673858643\n", + "Starting epoch 587\n", + "Train loss: 0.16201391369104384, Test loss: 0.1860337108373642\n", + "Starting epoch 588\n", + "Train loss: 0.1627282017469406, Test loss: 0.18746578693389893\n", + "Starting epoch 589\n", + "Train loss: 0.16216138809919356, Test loss: 0.1838478446006775\n", + "Starting epoch 590\n", + "Train loss: 0.16200567722320558, Test loss: 0.1823105663061142\n", + "Starting epoch 591\n", + "Train loss: 0.16217013865709304, Test loss: 0.18550580739974976\n", + "Starting epoch 592\n", + "Train loss: 0.16245590686798095, Test loss: 0.1858127862215042\n", + "Starting epoch 593\n", + "Train loss: 0.1617835545539856, Test loss: 0.18313004076480865\n", + "Starting epoch 594\n", + "Train loss: 0.16194497793912888, Test loss: 0.1824815571308136\n", + "Starting epoch 595\n", + "Train loss: 0.1630406141281128, Test loss: 0.1840108036994934\n", + "Starting epoch 596\n", + "Train loss: 0.16219058960676194, Test loss: 0.1837037056684494\n", + "Starting epoch 597\n", + "Train loss: 0.1621752128005028, Test loss: 0.1857423484325409\n", + "Starting epoch 598\n", + "Train loss: 0.16216665178537368, Test loss: 0.18426434695720673\n", + "Starting epoch 599\n", + "Train loss: 0.16195582181215287, Test loss: 0.18151862919330597\n", + "Starting epoch 600\n", + "Train loss: 0.16191237151622773, Test loss: 0.18628737330436707\n", + "Starting epoch 601\n", + "Train loss: 0.16174341797828673, Test loss: 0.18057437241077423\n", + "Starting epoch 602\n", + "Train loss: 0.1620568236708641, Test loss: 0.1816571205854416\n", + "Starting epoch 603\n", + "Train loss: 0.1623873570561409, Test loss: 0.18199728429317474\n", + "Starting epoch 604\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.16182003438472747, Test loss: 0.18533924221992493\n", + "Starting epoch 605\n", + "Train loss: 0.16192881375551224, Test loss: 0.1847291737794876\n", + "Starting epoch 606\n", + "Train loss: 0.1618121138215065, Test loss: 0.18491403758525848\n", + "Starting epoch 607\n", + "Train loss: 0.16193379670381547, Test loss: 0.18198807537555695\n", + "Starting epoch 608\n", + "Train loss: 0.16197190046310425, Test loss: 0.18657638132572174\n", + "Starting epoch 609\n", + "Train loss: 0.1614822182059288, Test loss: 0.1813846379518509\n", + "Starting epoch 610\n", + "Train loss: 0.16154830396175385, Test loss: 0.18380646407604218\n", + "Starting epoch 611\n", + "Train loss: 0.16163119733333586, Test loss: 0.18692904710769653\n", + "Starting epoch 612\n", + "Train loss: 0.1618915194272995, Test loss: 0.18689198791980743\n", + "Starting epoch 613\n", + "Train loss: 0.16164952635765076, Test loss: 0.18030275404453278\n", + "Starting epoch 614\n", + "Train loss: 0.16205763071775436, Test loss: 0.181634783744812\n", + "Starting epoch 615\n", + "Train loss: 0.16169953167438508, Test loss: 0.18490056693553925\n", + "Starting epoch 616\n", + "Train loss: 0.1623554790019989, Test loss: 0.18623539805412292\n", + "Starting epoch 617\n", + "Train loss: 0.1615043631196022, Test loss: 0.179594948887825\n", + "Starting epoch 618\n", + "Train loss: 0.16174718946218491, Test loss: 0.18171320855617523\n", + "Starting epoch 619\n", + "Train loss: 0.16164813578128814, Test loss: 0.18363343179225922\n", + "Starting epoch 620\n", + "Train loss: 0.16115461498498918, Test loss: 0.18110255897045135\n", + "Starting epoch 621\n", + "Train loss: 0.1614383900165558, Test loss: 0.18644949793815613\n", + "Starting epoch 622\n", + "Train loss: 0.16213718950748443, Test loss: 0.18088796734809875\n", + "Starting epoch 623\n", + "Train loss: 0.16126574635505675, Test loss: 0.18170766532421112\n", + "Starting epoch 624\n", + "Train loss: 0.16130798131227495, Test loss: 0.18118125200271606\n", + "Starting epoch 625\n", + "Train loss: 0.1620790535211563, Test loss: 0.1806475818157196\n", + "Starting epoch 626\n", + "Train loss: 0.16110272020101546, Test loss: 0.18213599920272827\n", + "Starting epoch 627\n", + "Train loss: 0.16150651335716248, Test loss: 0.1918867528438568\n", + "Starting epoch 628\n", + "Train loss: 0.16259206593036651, Test loss: 0.18164992332458496\n", + "Starting epoch 629\n", + "Train loss: 0.16118126690387727, Test loss: 0.17876790463924408\n", + "Starting epoch 630\n", + "Train loss: 0.16151294142007827, Test loss: 0.18167006969451904\n", + "Starting epoch 631\n", + "Train loss: 0.16137310177087782, Test loss: 0.17987583577632904\n", + "Starting epoch 632\n", + "Train loss: 0.16110699713230134, Test loss: 0.18407577276229858\n", + "Starting epoch 633\n", + "Train loss: 0.16111007183790207, Test loss: 0.18130692839622498\n", + "Starting epoch 634\n", + "Train loss: 0.16157290816307068, Test loss: 0.17995908856391907\n", + "Starting epoch 635\n", + "Train loss: 0.16094233423471452, Test loss: 0.1804324835538864\n", + "Starting epoch 636\n", + "Train loss: 0.16148990154266357, Test loss: 0.18793296813964844\n", + "Starting epoch 637\n", + "Train loss: 0.16089255303144456, Test loss: 0.1802871823310852\n", + "Starting epoch 638\n", + "Train loss: 0.16186053276062012, Test loss: 0.18891260027885437\n", + "Starting epoch 639\n", + "Train loss: 0.16120513558387756, Test loss: 0.18384304642677307\n", + "Starting epoch 640\n", + "Train loss: 0.16172830015420914, Test loss: 0.18436570465564728\n", + "Starting epoch 641\n", + "Train loss: 0.16080111503601074, Test loss: 0.1787639856338501\n", + "Starting epoch 642\n", + "Train loss: 0.16179285019636155, Test loss: 0.18874327838420868\n", + "Starting epoch 643\n", + "Train loss: 0.16121867895126343, Test loss: 0.18078680336475372\n", + "Starting epoch 644\n", + "Train loss: 0.16127301067113875, Test loss: 0.18186737596988678\n", + "Starting epoch 645\n", + "Train loss: 0.1602489936351776, Test loss: 0.17921476066112518\n", + "Starting epoch 646\n", + "Train loss: 0.16115788936614991, Test loss: 0.18348674476146698\n", + "Starting epoch 647\n", + "Train loss: 0.1611812937259674, Test loss: 0.1801442950963974\n", + "Starting epoch 648\n", + "Train loss: 0.1606442853808403, Test loss: 0.1847635954618454\n", + "Starting epoch 649\n", + "Train loss: 0.16054677456617356, Test loss: 0.17805911600589752\n", + "Starting epoch 650\n", + "Train loss: 0.1608068087697029, Test loss: 0.1795518547296524\n", + "Starting epoch 651\n", + "Train loss: 0.1602665913105011, Test loss: 0.18219056725502014\n", + "Starting epoch 652\n", + "Train loss: 0.1603730657696724, Test loss: 0.1767164170742035\n", + "Starting epoch 653\n", + "Train loss: 0.16061155438423158, Test loss: 0.1780107617378235\n", + "Starting epoch 654\n", + "Train loss: 0.16094835698604584, Test loss: 0.17854827642440796\n", + "Starting epoch 655\n", + "Train loss: 0.16073622226715087, Test loss: 0.17940470576286316\n", + "Starting epoch 656\n", + "Train loss: 0.16072943717241286, Test loss: 0.17853212356567383\n", + "Starting epoch 657\n", + "Train loss: 0.1605519488453865, Test loss: 0.1795371025800705\n", + "Starting epoch 658\n", + "Train loss: 0.16017508506774902, Test loss: 0.17643855512142181\n", + "Starting epoch 659\n", + "Train loss: 0.15987197041511536, Test loss: 0.17873795330524445\n", + "Starting epoch 660\n", + "Train loss: 0.16076190024614334, Test loss: 0.17770065367221832\n", + "Starting epoch 661\n", + "Train loss: 0.15962961554527283, Test loss: 0.17675729095935822\n", + "Starting epoch 662\n", + "Train loss: 0.16044797241687775, Test loss: 0.17717412114143372\n", + "Starting epoch 663\n", + "Train loss: 0.1598675885796547, Test loss: 0.1811511069536209\n", + "Starting epoch 664\n", + "Train loss: 0.15977220833301545, Test loss: 0.17661112546920776\n", + "Starting epoch 665\n", + "Train loss: 0.15977845311164857, Test loss: 0.18095341324806213\n", + "Starting epoch 666\n", + "Train loss: 0.15967604249715805, Test loss: 0.17899911105632782\n", + "Starting epoch 667\n", + "Train loss: 0.16031012684106827, Test loss: 0.1773764193058014\n", + "Starting epoch 668\n", + "Train loss: 0.1601187816262245, Test loss: 0.177215576171875\n", + "Starting epoch 669\n", + "Train loss: 0.15935453921556472, Test loss: 0.17532379925251007\n", + "Starting epoch 670\n", + "Train loss: 0.15968923449516295, Test loss: 0.17954972386360168\n", + "Starting epoch 671\n", + "Train loss: 0.15988343566656113, Test loss: 0.1755954772233963\n", + "Starting epoch 672\n", + "Train loss: 0.15907670676708222, Test loss: 0.1765255481004715\n", + "Starting epoch 673\n", + "Train loss: 0.1591021278500557, Test loss: 0.17626744508743286\n", + "Starting epoch 674\n", + "Train loss: 0.15936748653650284, Test loss: 0.17545051872730255\n", + "Starting epoch 675\n", + "Train loss: 0.1595002281665802, Test loss: 0.1784745305776596\n", + "Starting epoch 676\n", + "Train loss: 0.15877019315958024, Test loss: 0.178004652261734\n", + "Starting epoch 677\n", + "Train loss: 0.15976150840520859, Test loss: 0.17596441507339478\n", + "Starting epoch 678\n", + "Train loss: 0.15900494962930678, Test loss: 0.17529918253421783\n", + "Starting epoch 679\n", + "Train loss: 0.15936952352523803, Test loss: 0.17563624680042267\n", + "Starting epoch 680\n", + "Train loss: 0.15865642458200455, Test loss: 0.17710502445697784\n", + "Starting epoch 681\n", + "Train loss: 0.15884379893541337, Test loss: 0.17614637315273285\n", + "Starting epoch 682\n", + "Train loss: 0.15854426920413972, Test loss: 0.17695212364196777\n", + "Starting epoch 683\n", + "Train loss: 0.1591955116391182, Test loss: 0.17538635432720184\n", + "Starting epoch 684\n", + "Train loss: 0.15928123891353607, Test loss: 0.17930054664611816\n", + "Starting epoch 685\n", + "Train loss: 0.159405837059021, Test loss: 0.17480947077274323\n", + "Starting epoch 686\n", + "Train loss: 0.1583508223295212, Test loss: 0.17533278465270996\n", + "Starting epoch 687\n", + "Train loss: 0.1586959344148636, Test loss: 0.17732861638069153\n", + "Starting epoch 688\n", + "Train loss: 0.15820155501365663, Test loss: 0.17709240317344666\n", + "Starting epoch 689\n", + "Train loss: 0.15848714023828506, Test loss: 0.17623326182365417\n", + "Starting epoch 690\n", + "Train loss: 0.15853796780109405, Test loss: 0.18071787059307098\n", + "Starting epoch 691\n", + "Train loss: 0.15903353065252304, Test loss: 0.181341752409935\n", + "Starting epoch 692\n", + "Train loss: 0.15833432406187056, Test loss: 0.1763196736574173\n", + "Starting epoch 693\n", + "Train loss: 0.15859110534191131, Test loss: 0.17908556759357452\n", + "Starting epoch 694\n", + "Train loss: 0.15883199155330657, Test loss: 0.17488548159599304\n", + "Starting epoch 695\n", + "Train loss: 0.158236645758152, Test loss: 0.17884433269500732\n", + "Starting epoch 696\n", + "Train loss: 0.15804498434066772, Test loss: 0.17566753923892975\n", + "Starting epoch 697\n", + "Train loss: 0.1580254179239273, Test loss: 0.17661210894584656\n", + "Starting epoch 698\n", + "Train loss: 0.15820262283086778, Test loss: 0.17543600499629974\n", + "Starting epoch 699\n", + "Train loss: 0.15824722349643708, Test loss: 0.17796462774276733\n", + "Starting epoch 700\n", + "Train loss: 0.1583078533411026, Test loss: 0.1765119433403015\n", + "Starting epoch 701\n", + "Train loss: 0.15785971522331238, Test loss: 0.17596091330051422\n", + "Starting epoch 702\n", + "Train loss: 0.15822854936122893, Test loss: 0.17575332522392273\n", + "Starting epoch 703\n", + "Train loss: 0.15806237310171128, Test loss: 0.1750902235507965\n", + "Starting epoch 704\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.158183431327343, Test loss: 0.17582829296588898\n", + "Starting epoch 705\n", + "Train loss: 0.15790867567062378, Test loss: 0.17778590321540833\n", + "Starting epoch 706\n", + "Train loss: 0.15802417308092118, Test loss: 0.17661525309085846\n", + "Starting epoch 707\n", + "Train loss: 0.15779226541519165, Test loss: 0.17697423696517944\n", + "Starting epoch 708\n", + "Train loss: 0.15875499486923217, Test loss: 0.17712467908859253\n", + "Starting epoch 709\n", + "Train loss: 0.15819715112447738, Test loss: 0.1769641637802124\n", + "Starting epoch 710\n", + "Train loss: 0.15770487278699874, Test loss: 0.1764911562204361\n", + "Starting epoch 711\n", + "Train loss: 0.15794411420822144, Test loss: 0.18007473647594452\n", + "Starting epoch 712\n", + "Train loss: 0.15770682364702224, Test loss: 0.17579814791679382\n", + "Starting epoch 713\n", + "Train loss: 0.15781997114419938, Test loss: 0.17892853915691376\n", + "Starting epoch 714\n", + "Train loss: 0.15871124476194381, Test loss: 0.17692816257476807\n", + "Starting epoch 715\n", + "Train loss: 0.15749455362558365, Test loss: 0.1770819127559662\n", + "Starting epoch 716\n", + "Train loss: 0.157713560461998, Test loss: 0.1776408702135086\n", + "Starting epoch 717\n", + "Train loss: 0.15781322360038758, Test loss: 0.1769373118877411\n", + "Starting epoch 718\n", + "Train loss: 0.15738754570484162, Test loss: 0.1761171519756317\n", + "Starting epoch 719\n", + "Train loss: 0.15755039364099502, Test loss: 0.17762912809848785\n", + "Starting epoch 720\n", + "Train loss: 0.1572081884741783, Test loss: 0.17692722380161285\n", + "Starting epoch 721\n", + "Train loss: 0.157270687520504, Test loss: 0.175065815448761\n", + "Starting epoch 722\n", + "Train loss: 0.15725588262081147, Test loss: 0.1748971939086914\n", + "Starting epoch 723\n", + "Train loss: 0.15773274689912797, Test loss: 0.17508305609226227\n", + "Starting epoch 724\n", + "Train loss: 0.15733925104141236, Test loss: 0.17882299423217773\n", + "Starting epoch 725\n", + "Train loss: 0.15768029987812043, Test loss: 0.1761925369501114\n", + "Starting epoch 726\n", + "Train loss: 0.1570081290602684, Test loss: 0.178508922457695\n", + "Starting epoch 727\n", + "Train loss: 0.1576265075802803, Test loss: 0.17841774225234985\n", + "Starting epoch 728\n", + "Train loss: 0.15678511887788774, Test loss: 0.17580080032348633\n", + "Starting epoch 729\n", + "Train loss: 0.1570144525170326, Test loss: 0.17715361714363098\n", + "Starting epoch 730\n", + "Train loss: 0.15779257088899612, Test loss: 0.17679238319396973\n", + "Starting epoch 731\n", + "Train loss: 0.1572554224729538, Test loss: 0.1778978705406189\n", + "Starting epoch 732\n", + "Train loss: 0.15714868366718293, Test loss: 0.17783789336681366\n", + "Starting epoch 733\n", + "Train loss: 0.15708053588867188, Test loss: 0.17751546204090118\n", + "Starting epoch 734\n", + "Train loss: 0.15670244574546813, Test loss: 0.17880071699619293\n", + "Starting epoch 735\n", + "Train loss: 0.15706685572862625, Test loss: 0.17580609023571014\n", + "Starting epoch 736\n", + "Train loss: 0.15678432434797288, Test loss: 0.17717331647872925\n", + "Starting epoch 737\n", + "Train loss: 0.15686407566070557, Test loss: 0.18278972804546356\n", + "Starting epoch 738\n", + "Train loss: 0.1570545920729637, Test loss: 0.17893603444099426\n", + "Starting epoch 739\n", + "Train loss: 0.1573004913330078, Test loss: 0.17568624019622803\n", + "Starting epoch 740\n", + "Train loss: 0.1570126360654831, Test loss: 0.17741960287094116\n", + "Starting epoch 741\n", + "Train loss: 0.15647889524698258, Test loss: 0.17769910395145416\n", + "Starting epoch 742\n", + "Train loss: 0.156828892827034, Test loss: 0.17917731404304504\n", + "Starting epoch 743\n", + "Train loss: 0.156875923871994, Test loss: 0.17782413959503174\n", + "Starting epoch 744\n", + "Train loss: 0.15690837264060975, Test loss: 0.17833548784255981\n", + "Starting epoch 745\n", + "Train loss: 0.15716096609830857, Test loss: 0.17794892191886902\n", + "Starting epoch 746\n", + "Train loss: 0.15651264250278474, Test loss: 0.1769465059041977\n", + "Starting epoch 747\n", + "Train loss: 0.15645872861146926, Test loss: 0.17871242761611938\n", + "Starting epoch 748\n", + "Train loss: 0.15629384696483611, Test loss: 0.17900624871253967\n", + "Starting epoch 749\n", + "Train loss: 0.15675463795661926, Test loss: 0.17978084087371826\n", + "Starting epoch 750\n", + "Train loss: 0.15670695453882216, Test loss: 0.1788851022720337\n", + "Starting epoch 751\n", + "Train loss: 0.1564847606420517, Test loss: 0.1844462901353836\n", + "Starting epoch 752\n", + "Train loss: 0.15673744261264802, Test loss: 0.1810971200466156\n", + "Starting epoch 753\n", + "Train loss: 0.15719325363636016, Test loss: 0.18165740370750427\n", + "Starting epoch 754\n", + "Train loss: 0.15593231201171875, Test loss: 0.1829884648323059\n", + "Starting epoch 755\n", + "Train loss: 0.15616499841213227, Test loss: 0.17958863079547882\n", + "Starting epoch 756\n", + "Train loss: 0.15645647317171096, Test loss: 0.18044564127922058\n", + "Starting epoch 757\n", + "Train loss: 0.1558713811635971, Test loss: 0.18047837913036346\n", + "Starting epoch 758\n", + "Train loss: 0.15708172887563707, Test loss: 0.1790766417980194\n", + "Starting epoch 759\n", + "Train loss: 0.15620475828647615, Test loss: 0.1816694140434265\n", + "Starting epoch 760\n", + "Train loss: 0.15679713755846023, Test loss: 0.18023186922073364\n", + "Starting epoch 761\n", + "Train loss: 0.1561601686477661, Test loss: 0.1828075647354126\n", + "Starting epoch 762\n", + "Train loss: 0.15630848437547684, Test loss: 0.18002402782440186\n", + "Starting epoch 763\n", + "Train loss: 0.15588688254356384, Test loss: 0.18349963426589966\n", + "Starting epoch 764\n", + "Train loss: 0.15600742131471634, Test loss: 0.18029139935970306\n", + "Starting epoch 765\n", + "Train loss: 0.15578888982534408, Test loss: 0.18164728581905365\n", + "Starting epoch 766\n", + "Train loss: 0.15564021348953247, Test loss: 0.18182048201560974\n", + "Starting epoch 767\n", + "Train loss: 0.15598283499479293, Test loss: 0.18205119669437408\n", + "Starting epoch 768\n", + "Train loss: 0.15589061111211777, Test loss: 0.18899935483932495\n", + "Starting epoch 769\n", + "Train loss: 0.1557464611530304, Test loss: 0.18035128712654114\n", + "Starting epoch 770\n", + "Train loss: 0.15517526745796204, Test loss: 0.1817605197429657\n", + "Starting epoch 771\n", + "Train loss: 0.1556827852129936, Test loss: 0.18273724615573883\n", + "Starting epoch 772\n", + "Train loss: 0.15593234330415726, Test loss: 0.18829523026943207\n", + "Starting epoch 773\n", + "Train loss: 0.15573953419923783, Test loss: 0.18339847028255463\n", + "Starting epoch 774\n", + "Train loss: 0.15526160329580307, Test loss: 0.18390217423439026\n", + "Starting epoch 775\n", + "Train loss: 0.15544029742479323, Test loss: 0.1839882731437683\n", + "Starting epoch 776\n", + "Train loss: 0.1557706278562546, Test loss: 0.18294546008110046\n", + "Starting epoch 777\n", + "Train loss: 0.1552382692694664, Test loss: 0.1870020180940628\n", + "Starting epoch 778\n", + "Train loss: 0.15550502866506577, Test loss: 0.18469437956809998\n", + "Starting epoch 779\n", + "Train loss: 0.15516233563423157, Test loss: 0.18466545641422272\n", + "Starting epoch 780\n", + "Train loss: 0.1551545563340187, Test loss: 0.18572235107421875\n", + "Starting epoch 781\n", + "Train loss: 0.15572760850191117, Test loss: 0.18933050334453583\n", + "Starting epoch 782\n", + "Train loss: 0.15486220091581346, Test loss: 0.18366560339927673\n", + "Starting epoch 783\n", + "Train loss: 0.1546187022328377, Test loss: 0.1850610077381134\n", + "Starting epoch 784\n", + "Train loss: 0.1551326560974121, Test loss: 0.183871328830719\n", + "Starting epoch 785\n", + "Train loss: 0.15516918063163757, Test loss: 0.18390993773937225\n", + "Starting epoch 786\n", + "Train loss: 0.15489620774984358, Test loss: 0.18996666371822357\n", + "Starting epoch 787\n", + "Train loss: 0.15486974596977235, Test loss: 0.18488909304141998\n", + "Starting epoch 788\n", + "Train loss: 0.15481788784265518, Test loss: 0.1847730427980423\n", + "Starting epoch 789\n", + "Train loss: 0.15493398398160935, Test loss: 0.18682339787483215\n", + "Starting epoch 790\n", + "Train loss: 0.15479137152433395, Test loss: 0.18591415882110596\n", + "Starting epoch 791\n", + "Train loss: 0.15510978937149048, Test loss: 0.18541556596755981\n", + "Starting epoch 792\n", + "Train loss: 0.15477289855480195, Test loss: 0.1852128654718399\n", + "Starting epoch 793\n", + "Train loss: 0.1545557263493538, Test loss: 0.18419036269187927\n", + "Starting epoch 794\n", + "Train loss: 0.1546010845899582, Test loss: 0.18671908974647522\n", + "Starting epoch 795\n", + "Train loss: 0.1545565891265869, Test loss: 0.18531611561775208\n", + "Starting epoch 796\n", + "Train loss: 0.15462544322013855, Test loss: 0.1913522183895111\n", + "Starting epoch 797\n", + "Train loss: 0.15442328840494157, Test loss: 0.18493583798408508\n", + "Starting epoch 798\n", + "Train loss: 0.15401893049478532, Test loss: 0.185791477560997\n", + "Starting epoch 799\n", + "Train loss: 0.15453988343477248, Test loss: 0.18566888570785522\n", + "Starting epoch 800\n", + "Train loss: 0.15439439624547957, Test loss: 0.1858370453119278\n", + "Starting epoch 801\n", + "Train loss: 0.15465423136949538, Test loss: 0.18508237600326538\n", + "Starting epoch 802\n", + "Train loss: 0.1543620654940605, Test loss: 0.1855679452419281\n", + "Starting epoch 803\n", + "Train loss: 0.15419078439474107, Test loss: 0.1872600018978119\n", + "Starting epoch 804\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.15403182566165924, Test loss: 0.18614567816257477\n", + "Starting epoch 805\n", + "Train loss: 0.15393486231565476, Test loss: 0.18560685217380524\n", + "Starting epoch 806\n", + "Train loss: 0.15453119069337845, Test loss: 0.1849733591079712\n", + "Starting epoch 807\n", + "Train loss: 0.15429811507463456, Test loss: 0.1846090853214264\n", + "Starting epoch 808\n", + "Train loss: 0.1544571974873543, Test loss: 0.18517723679542542\n", + "Starting epoch 809\n", + "Train loss: 0.15415377914905548, Test loss: 0.1833035796880722\n", + "Starting epoch 810\n", + "Train loss: 0.1538264521956444, Test loss: 0.18545043468475342\n", + "Starting epoch 811\n", + "Train loss: 0.15353480964899063, Test loss: 0.18458662927150726\n", + "Starting epoch 812\n", + "Train loss: 0.15415227711200713, Test loss: 0.183198019862175\n", + "Starting epoch 813\n", + "Train loss: 0.1540687435865402, Test loss: 0.18791504204273224\n", + "Starting epoch 814\n", + "Train loss: 0.1535621777176857, Test loss: 0.18462049961090088\n", + "Starting epoch 815\n", + "Train loss: 0.15344598770141601, Test loss: 0.18565380573272705\n", + "Starting epoch 816\n", + "Train loss: 0.15419911861419677, Test loss: 0.18394143879413605\n", + "Starting epoch 817\n", + "Train loss: 0.15388002276420593, Test loss: 0.18407930433750153\n", + "Starting epoch 818\n", + "Train loss: 0.15351936221122742, Test loss: 0.1830548793077469\n", + "Starting epoch 819\n", + "Train loss: 0.15339540392160417, Test loss: 0.18263550102710724\n", + "Starting epoch 820\n", + "Train loss: 0.15351445943117142, Test loss: 0.186944380402565\n", + "Starting epoch 821\n", + "Train loss: 0.15342902958393098, Test loss: 0.18544471263885498\n", + "Starting epoch 822\n", + "Train loss: 0.15350386142730713, Test loss: 0.18285056948661804\n", + "Starting epoch 823\n", + "Train loss: 0.15389509826898576, Test loss: 0.18147876858711243\n", + "Starting epoch 824\n", + "Train loss: 0.1532948461174965, Test loss: 0.18251581490039825\n", + "Starting epoch 825\n", + "Train loss: 0.1538844844698906, Test loss: 0.18364784121513367\n", + "Starting epoch 826\n", + "Train loss: 0.153936066031456, Test loss: 0.1829390674829483\n", + "Starting epoch 827\n", + "Train loss: 0.15343904942274095, Test loss: 0.18254584074020386\n", + "Starting epoch 828\n", + "Train loss: 0.15319333851337433, Test loss: 0.18109746277332306\n", + "Starting epoch 829\n", + "Train loss: 0.1530449840426445, Test loss: 0.18136104941368103\n", + "Starting epoch 830\n", + "Train loss: 0.15344687730073928, Test loss: 0.18902276456356049\n", + "Starting epoch 831\n", + "Train loss: 0.15339966148138046, Test loss: 0.18190638720989227\n", + "Starting epoch 832\n", + "Train loss: 0.15294707030057908, Test loss: 0.18112923204898834\n", + "Starting epoch 833\n", + "Train loss: 0.1532418993115425, Test loss: 0.18106691539287567\n", + "Starting epoch 834\n", + "Train loss: 0.15299765288829803, Test loss: 0.18051889538764954\n", + "Starting epoch 835\n", + "Train loss: 0.15254301577806473, Test loss: 0.1812930405139923\n", + "Starting epoch 836\n", + "Train loss: 0.15262522369623185, Test loss: 0.1812898814678192\n", + "Starting epoch 837\n", + "Train loss: 0.15385435223579408, Test loss: 0.1813945472240448\n", + "Starting epoch 838\n", + "Train loss: 0.15304524600505828, Test loss: 0.18252508342266083\n", + "Starting epoch 839\n", + "Train loss: 0.15269990444183348, Test loss: 0.18280217051506042\n", + "Starting epoch 840\n", + "Train loss: 0.1531848919391632, Test loss: 0.18076340854167938\n", + "Starting epoch 841\n", + "Train loss: 0.15267266929149628, Test loss: 0.1805378794670105\n", + "Starting epoch 842\n", + "Train loss: 0.15247853964567185, Test loss: 0.17890667915344238\n", + "Starting epoch 843\n", + "Train loss: 0.15247229516506194, Test loss: 0.18373560905456543\n", + "Starting epoch 844\n", + "Train loss: 0.15242167562246323, Test loss: 0.17756184935569763\n", + "Starting epoch 845\n", + "Train loss: 0.15253809452056885, Test loss: 0.1818361133337021\n", + "Starting epoch 846\n", + "Train loss: 0.15310435950756074, Test loss: 0.18186083436012268\n", + "Starting epoch 847\n", + "Train loss: 0.15226980715990066, Test loss: 0.17918246984481812\n", + "Starting epoch 848\n", + "Train loss: 0.1528629407286644, Test loss: 0.17951756715774536\n", + "Starting epoch 849\n", + "Train loss: 0.15218637526035309, Test loss: 0.1843460649251938\n", + "Starting epoch 850\n", + "Train loss: 0.152684545814991, Test loss: 0.17973117530345917\n", + "Starting epoch 851\n", + "Train loss: 0.15232110589742662, Test loss: 0.17722758650779724\n", + "Starting epoch 852\n", + "Train loss: 0.15219103515148164, Test loss: 0.1771695762872696\n", + "Starting epoch 853\n", + "Train loss: 0.15241131395101548, Test loss: 0.1762767732143402\n", + "Starting epoch 854\n", + "Train loss: 0.1523800328373909, Test loss: 0.1782151162624359\n", + "Starting epoch 855\n", + "Train loss: 0.15275462925434113, Test loss: 0.17908038198947906\n", + "Starting epoch 856\n", + "Train loss: 0.15228213250637054, Test loss: 0.17688730359077454\n", + "Starting epoch 857\n", + "Train loss: 0.1518900015950203, Test loss: 0.177104189991951\n", + "Starting epoch 858\n", + "Train loss: 0.15193135738372804, Test loss: 0.18143117427825928\n", + "Starting epoch 859\n", + "Train loss: 0.15241081565618514, Test loss: 0.17505307495594025\n", + "Starting epoch 860\n", + "Train loss: 0.1520268738269806, Test loss: 0.18204008042812347\n", + "Starting epoch 861\n", + "Train loss: 0.1517037883400917, Test loss: 0.17474903166294098\n", + "Starting epoch 862\n", + "Train loss: 0.1517617878317833, Test loss: 0.17614306509494781\n", + "Starting epoch 863\n", + "Train loss: 0.15256436467170714, Test loss: 0.17454594373703003\n", + "Starting epoch 864\n", + "Train loss: 0.15219809204339982, Test loss: 0.17538651823997498\n", + "Starting epoch 865\n", + "Train loss: 0.15164065182209016, Test loss: 0.1746489703655243\n", + "Starting epoch 866\n", + "Train loss: 0.15152181327342987, Test loss: 0.17852577567100525\n", + "Starting epoch 867\n", + "Train loss: 0.15197062730789185, Test loss: 0.17737141251564026\n", + "Starting epoch 868\n", + "Train loss: 0.15213120073080064, Test loss: 0.17796139419078827\n", + "Starting epoch 869\n", + "Train loss: 0.15184614330530166, Test loss: 0.17417794466018677\n", + "Starting epoch 870\n", + "Train loss: 0.15129219204187394, Test loss: 0.1736692190170288\n", + "Starting epoch 871\n", + "Train loss: 0.15175350159406661, Test loss: 0.17714768648147583\n", + "Starting epoch 872\n", + "Train loss: 0.15139543652534485, Test loss: 0.17338334023952484\n", + "Starting epoch 873\n", + "Train loss: 0.15170849680900575, Test loss: 0.17218826711177826\n", + "Starting epoch 874\n", + "Train loss: 0.15127191960811615, Test loss: 0.171831414103508\n", + "Starting epoch 875\n", + "Train loss: 0.15092819273471833, Test loss: 0.17178256809711456\n", + "Starting epoch 876\n", + "Train loss: 0.1509759619832039, Test loss: 0.1707778126001358\n", + "Starting epoch 877\n", + "Train loss: 0.15160735160112382, Test loss: 0.1723155677318573\n", + "Starting epoch 878\n", + "Train loss: 0.1515893432497978, Test loss: 0.1720626950263977\n", + "Starting epoch 879\n", + "Train loss: 0.1507820564508438, Test loss: 0.17124752700328827\n", + "Starting epoch 880\n", + "Train loss: 0.15151018261909485, Test loss: 0.16991733014583588\n", + "Starting epoch 881\n", + "Train loss: 0.15068840444087983, Test loss: 0.17361070215702057\n", + "Starting epoch 882\n", + "Train loss: 0.15101482599973678, Test loss: 0.17342497408390045\n", + "Starting epoch 883\n", + "Train loss: 0.15154436469078064, Test loss: 0.17058473825454712\n", + "Starting epoch 884\n", + "Train loss: 0.15034719794988632, Test loss: 0.1717720925807953\n", + "Starting epoch 885\n", + "Train loss: 0.15015917241573334, Test loss: 0.16901426017284393\n", + "Starting epoch 886\n", + "Train loss: 0.14994985848665238, Test loss: 0.1686113327741623\n", + "Starting epoch 887\n", + "Train loss: 0.15043990135192872, Test loss: 0.17182250320911407\n", + "Starting epoch 888\n", + "Train loss: 0.1500939956307411, Test loss: 0.17009833455085754\n", + "Starting epoch 889\n", + "Train loss: 0.15012032985687257, Test loss: 0.17019490897655487\n", + "Starting epoch 890\n", + "Train loss: 0.14993320226669313, Test loss: 0.16761209070682526\n", + "Starting epoch 891\n", + "Train loss: 0.14961104214191437, Test loss: 0.16839458048343658\n", + "Starting epoch 892\n", + "Train loss: 0.14978093415498733, Test loss: 0.17497888207435608\n", + "Starting epoch 893\n", + "Train loss: 0.14985983252525328, Test loss: 0.16833806037902832\n", + "Starting epoch 894\n", + "Train loss: 0.14962367922067643, Test loss: 0.16806909441947937\n", + "Starting epoch 895\n", + "Train loss: 0.14924299240112304, Test loss: 0.16848313808441162\n", + "Starting epoch 896\n", + "Train loss: 0.14925175189971923, Test loss: 0.16790537536144257\n", + "Starting epoch 897\n", + "Train loss: 0.14941574692726134, Test loss: 0.17097708582878113\n", + "Starting epoch 898\n", + "Train loss: 0.14902330338954925, Test loss: 0.16679398715496063\n", + "Starting epoch 899\n", + "Train loss: 0.14883773118257523, Test loss: 0.1689378321170807\n", + "Starting epoch 900\n", + "Train loss: 0.1492244592308998, Test loss: 0.16687943041324615\n", + "Starting epoch 901\n", + "Train loss: 0.1482501122355461, Test loss: 0.1697612702846527\n", + "Starting epoch 902\n", + "Train loss: 0.14869109332561492, Test loss: 0.1653280109167099\n", + "Starting epoch 903\n", + "Train loss: 0.14827948421239853, Test loss: 0.16683430969715118\n", + "Starting epoch 904\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.148163044154644, Test loss: 0.1667039841413498\n", + "Starting epoch 905\n", + "Train loss: 0.1484239238500595, Test loss: 0.16841275990009308\n", + "Starting epoch 906\n", + "Train loss: 0.14785656154155732, Test loss: 0.16743656992912292\n", + "Starting epoch 907\n", + "Train loss: 0.14759473711252213, Test loss: 0.1669757068157196\n", + "Starting epoch 908\n", + "Train loss: 0.14774231344461441, Test loss: 0.1678246706724167\n", + "Starting epoch 909\n", + "Train loss: 0.14802570521831512, Test loss: 0.16878698766231537\n", + "Starting epoch 910\n", + "Train loss: 0.1480949753522873, Test loss: 0.1700417399406433\n", + "Starting epoch 911\n", + "Train loss: 0.1478197979927063, Test loss: 0.17120766639709473\n", + "Starting epoch 912\n", + "Train loss: 0.14747726738452913, Test loss: 0.16835835576057434\n", + "Starting epoch 913\n", + "Train loss: 0.1470561346411705, Test loss: 0.17283175885677338\n", + "Starting epoch 914\n", + "Train loss: 0.14707072585821152, Test loss: 0.17065219581127167\n", + "Starting epoch 915\n", + "Train loss: 0.1470457410812378, Test loss: 0.17214281857013702\n", + "Starting epoch 916\n", + "Train loss: 0.14705794662237168, Test loss: 0.1695377230644226\n", + "Starting epoch 917\n", + "Train loss: 0.1466624566912651, Test loss: 0.17550736665725708\n", + "Starting epoch 918\n", + "Train loss: 0.14687291115522386, Test loss: 0.17194977402687073\n", + "Starting epoch 919\n", + "Train loss: 0.1469936728477478, Test loss: 0.1698514223098755\n", + "Starting epoch 920\n", + "Train loss: 0.14699729919433593, Test loss: 0.17195454239845276\n", + "Starting epoch 921\n", + "Train loss: 0.14644166737794875, Test loss: 0.17500676214694977\n", + "Starting epoch 922\n", + "Train loss: 0.14664822041988373, Test loss: 0.16881540417671204\n", + "Starting epoch 923\n", + "Train loss: 0.14656826674938203, Test loss: 0.17128673195838928\n", + "Starting epoch 924\n", + "Train loss: 0.146870057284832, Test loss: 0.17642633616924286\n", + "Starting epoch 925\n", + "Train loss: 0.14622545063495637, Test loss: 0.17037270963191986\n", + "Starting epoch 926\n", + "Train loss: 0.1463041904568672, Test loss: 0.17363303899765015\n", + "Starting epoch 927\n", + "Train loss: 0.1462299981713295, Test loss: 0.18276938796043396\n", + "Starting epoch 928\n", + "Train loss: 0.14642820864915848, Test loss: 0.1728794425725937\n", + "Starting epoch 929\n", + "Train loss: 0.14583320260047913, Test loss: 0.17065834999084473\n", + "Starting epoch 930\n", + "Train loss: 0.1463873267173767, Test loss: 0.1886533945798874\n", + "Starting epoch 931\n", + "Train loss: 0.14669109642505646, Test loss: 0.17313209176063538\n", + "Starting epoch 932\n", + "Train loss: 0.14586580485105516, Test loss: 0.16969680786132812\n", + "Starting epoch 933\n", + "Train loss: 0.1453286150097847, Test loss: 0.16938282549381256\n", + "Starting epoch 934\n", + "Train loss: 0.1461893206834793, Test loss: 0.16933438181877136\n", + "Starting epoch 935\n", + "Train loss: 0.14532573133707047, Test loss: 0.16589513421058655\n", + "Starting epoch 936\n", + "Train loss: 0.14520068407058717, Test loss: 0.17683036625385284\n", + "Starting epoch 937\n", + "Train loss: 0.14609683334827422, Test loss: 0.17368851602077484\n", + "Starting epoch 938\n", + "Train loss: 0.14451414227485657, Test loss: 0.1667650192975998\n", + "Starting epoch 939\n", + "Train loss: 0.14445282578468321, Test loss: 0.1631922870874405\n", + "Starting epoch 940\n", + "Train loss: 0.14435013890266418, Test loss: 0.1640893816947937\n", + "Starting epoch 941\n", + "Train loss: 0.143919817507267, Test loss: 0.1633167266845703\n", + "Starting epoch 942\n", + "Train loss: 0.14297609090805052, Test loss: 0.16697660088539124\n", + "Starting epoch 943\n", + "Train loss: 0.14365106642246248, Test loss: 0.16178393363952637\n", + "Starting epoch 944\n", + "Train loss: 0.14296137183904647, Test loss: 0.16631431877613068\n", + "Starting epoch 945\n", + "Train loss: 0.1426262816786766, Test loss: 0.167317733168602\n", + "Starting epoch 946\n", + "Train loss: 0.14192384034395217, Test loss: 0.16802045702934265\n", + "Starting epoch 947\n", + "Train loss: 0.14207273662090303, Test loss: 0.16509012877941132\n", + "Starting epoch 948\n", + "Train loss: 0.14185331284999847, Test loss: 0.1797834187746048\n", + "Starting epoch 949\n", + "Train loss: 0.14171916723251343, Test loss: 0.17000384628772736\n", + "Starting epoch 950\n", + "Train loss: 0.1411633610725403, Test loss: 0.17207635939121246\n", + "Starting epoch 951\n", + "Train loss: 0.14073788553476332, Test loss: 0.1780034899711609\n", + "Starting epoch 952\n", + "Train loss: 0.14094052761793135, Test loss: 0.1725774109363556\n", + "Starting epoch 953\n", + "Train loss: 0.14057504534721374, Test loss: 0.17413926124572754\n", + "Starting epoch 954\n", + "Train loss: 0.13985313564538956, Test loss: 0.16750556230545044\n", + "Starting epoch 955\n", + "Train loss: 0.14011855125427247, Test loss: 0.1764967441558838\n", + "Starting epoch 956\n", + "Train loss: 0.1396419432759285, Test loss: 0.181986466050148\n", + "Starting epoch 957\n", + "Train loss: 0.1395201501250267, Test loss: 0.16480609774589539\n", + "Starting epoch 958\n", + "Train loss: 0.1389840704202652, Test loss: 0.1761760264635086\n", + "Starting epoch 959\n", + "Train loss: 0.13862037688493728, Test loss: 0.1741502285003662\n", + "Starting epoch 960\n", + "Train loss: 0.13787006884813308, Test loss: 0.1757761538028717\n", + "Starting epoch 961\n", + "Train loss: 0.13801406383514403, Test loss: 0.1854708343744278\n", + "Starting epoch 962\n", + "Train loss: 0.13801027864217758, Test loss: 0.17448584735393524\n", + "Starting epoch 963\n", + "Train loss: 0.13738477051258088, Test loss: 0.16906467080116272\n", + "Starting epoch 964\n", + "Train loss: 0.13774635672569274, Test loss: 0.17239654064178467\n", + "Starting epoch 965\n", + "Train loss: 0.13678600788116455, Test loss: 0.17119485139846802\n", + "Starting epoch 966\n", + "Train loss: 0.13683569699525833, Test loss: 0.1656903773546219\n", + "Starting epoch 967\n", + "Train loss: 0.1372031334042549, Test loss: 0.1658727377653122\n", + "Starting epoch 968\n", + "Train loss: 0.13670568019151688, Test loss: 0.16689440608024597\n", + "Starting epoch 969\n", + "Train loss: 0.13611359775066376, Test loss: 0.17139972746372223\n", + "Starting epoch 970\n", + "Train loss: 0.1366174551844597, Test loss: 0.1656612753868103\n", + "Starting epoch 971\n", + "Train loss: 0.13660117149353027, Test loss: 0.1818958967924118\n", + "Starting epoch 972\n", + "Train loss: 0.13701987206935884, Test loss: 0.17607851326465607\n", + "Starting epoch 973\n", + "Train loss: 0.13601728469133378, Test loss: 0.17147286236286163\n", + "Starting epoch 974\n", + "Train loss: 0.13637277305126191, Test loss: 0.16459982097148895\n", + "Starting epoch 975\n", + "Train loss: 0.13697579205036164, Test loss: 0.16696487367153168\n", + "Starting epoch 976\n", + "Train loss: 0.1357903450727463, Test loss: 0.16698910295963287\n", + "Starting epoch 977\n", + "Train loss: 0.13579561233520507, Test loss: 0.1676780879497528\n", + "Starting epoch 978\n", + "Train loss: 0.13586585938930512, Test loss: 0.1843281388282776\n", + "Starting epoch 979\n", + "Train loss: 0.13676699072122575, Test loss: 0.17119820415973663\n", + "Starting epoch 980\n", + "Train loss: 0.13546618163585664, Test loss: 0.17045684158802032\n", + "Starting epoch 981\n", + "Train loss: 0.1357137730717659, Test loss: 0.17480526864528656\n", + "Starting epoch 982\n", + "Train loss: 0.1361057072877884, Test loss: 0.16687752306461334\n", + "Starting epoch 983\n", + "Train loss: 0.13561264663934708, Test loss: 0.1712927222251892\n", + "Starting epoch 984\n", + "Train loss: 0.13583486676216125, Test loss: 0.1691386103630066\n", + "Starting epoch 985\n", + "Train loss: 0.13549711674451828, Test loss: 0.16838030517101288\n", + "Starting epoch 986\n", + "Train loss: 0.1353756245970726, Test loss: 0.17257164418697357\n", + "Starting epoch 987\n", + "Train loss: 0.13524686872959138, Test loss: 0.16707395017147064\n", + "Starting epoch 988\n", + "Train loss: 0.13606603056192398, Test loss: 0.16725914180278778\n", + "Starting epoch 989\n", + "Train loss: 0.1358884871006012, Test loss: 0.16943366825580597\n", + "Starting epoch 990\n", + "Train loss: 0.13536040991544723, Test loss: 0.1694674789905548\n", + "Starting epoch 991\n", + "Train loss: 0.1350603550672531, Test loss: 0.16414137184619904\n", + "Starting epoch 992\n", + "Train loss: 0.13547885537147522, Test loss: 0.1681455373764038\n", + "Starting epoch 993\n", + "Train loss: 0.13483263731002806, Test loss: 0.1663752943277359\n", + "Starting epoch 994\n", + "Train loss: 0.13490934938192367, Test loss: 0.16661545634269714\n", + "Starting epoch 995\n", + "Train loss: 0.13472729086875915, Test loss: 0.16450682282447815\n", + "Starting epoch 996\n", + "Train loss: 0.13483618527650834, Test loss: 0.16530035436153412\n", + "Starting epoch 997\n", + "Train loss: 0.13526238411664962, Test loss: 0.16946576535701752\n", + "Starting epoch 998\n", + "Train loss: 0.13464825928211213, Test loss: 0.16743266582489014\n", + "Starting epoch 999\n", + "Train loss: 0.13467494398355484, Test loss: 0.164796844124794\n", + "Starting epoch 1000\n", + "Train loss: 0.13482290685176848, Test loss: 0.1631363332271576\n", + "Starting epoch 1001\n", + "Train loss: 0.13415511310100556, Test loss: 0.16904528439044952\n", + "Starting epoch 1002\n", + "Train loss: 0.13453493714332582, Test loss: 0.1644889861345291\n", + "Starting epoch 1003\n", + "Train loss: 0.13441491723060608, Test loss: 0.16913630068302155\n", + "Starting epoch 1004\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.13376774460077287, Test loss: 0.16514889895915985\n", + "Starting epoch 1005\n", + "Train loss: 0.13403558403253554, Test loss: 0.17374403774738312\n", + "Starting epoch 1006\n", + "Train loss: 0.1338075366616249, Test loss: 0.16401812434196472\n", + "Starting epoch 1007\n", + "Train loss: 0.13424518167972566, Test loss: 0.16843381524085999\n", + "Starting epoch 1008\n", + "Train loss: 0.1328984159231186, Test loss: 0.16238072514533997\n", + "Starting epoch 1009\n", + "Train loss: 0.13299637466669081, Test loss: 0.16519396007061005\n", + "Starting epoch 1010\n", + "Train loss: 0.13236493855714798, Test loss: 0.15850885212421417\n", + "Starting epoch 1011\n", + "Train loss: 0.1312198594212532, Test loss: 0.15695539116859436\n", + "Starting epoch 1012\n", + "Train loss: 0.13167913287878036, Test loss: 0.16510191559791565\n", + "Starting epoch 1013\n", + "Train loss: 0.13084726482629777, Test loss: 0.15753403306007385\n", + "Starting epoch 1014\n", + "Train loss: 0.13003639072179796, Test loss: 0.15365031361579895\n", + "Starting epoch 1015\n", + "Train loss: 0.12962202936410905, Test loss: 0.1539866030216217\n", + "Starting epoch 1016\n", + "Train loss: 0.1293671876192093, Test loss: 0.15700143575668335\n", + "Starting epoch 1017\n", + "Train loss: 0.1299566814303398, Test loss: 0.16107624769210815\n", + "Starting epoch 1018\n", + "Train loss: 0.1287783908843994, Test loss: 0.15583942830562592\n", + "Starting epoch 1019\n", + "Train loss: 0.12743137568235396, Test loss: 0.15011684596538544\n", + "Starting epoch 1020\n", + "Train loss: 0.12720238775014878, Test loss: 0.14957383275032043\n", + "Starting epoch 1021\n", + "Train loss: 0.1269572839140892, Test loss: 0.14835162460803986\n", + "Starting epoch 1022\n", + "Train loss: 0.12691068798303604, Test loss: 0.1488005667924881\n", + "Starting epoch 1023\n", + "Train loss: 0.12695645302534103, Test loss: 0.15175935626029968\n", + "Starting epoch 1024\n", + "Train loss: 0.1259537823498249, Test loss: 0.14736710488796234\n", + "Starting epoch 1025\n", + "Train loss: 0.12552673295140265, Test loss: 0.1495881974697113\n", + "Starting epoch 1026\n", + "Train loss: 0.12555690303444864, Test loss: 0.14839142560958862\n", + "Starting epoch 1027\n", + "Train loss: 0.12491542711853981, Test loss: 0.1483641266822815\n", + "Starting epoch 1028\n", + "Train loss: 0.12536612167954445, Test loss: 0.14885666966438293\n", + "Starting epoch 1029\n", + "Train loss: 0.12434986174106598, Test loss: 0.146251380443573\n", + "Starting epoch 1030\n", + "Train loss: 0.12481644958257675, Test loss: 0.1479286104440689\n", + "Starting epoch 1031\n", + "Train loss: 0.1246460248529911, Test loss: 0.1508769690990448\n", + "Starting epoch 1032\n", + "Train loss: 0.12430201113224029, Test loss: 0.14644794166088104\n", + "Starting epoch 1033\n", + "Train loss: 0.12439876154065133, Test loss: 0.14531724154949188\n", + "Starting epoch 1034\n", + "Train loss: 0.12378449782729149, Test loss: 0.14877240359783173\n", + "Starting epoch 1035\n", + "Train loss: 0.1243552614748478, Test loss: 0.1421542465686798\n", + "Starting epoch 1036\n", + "Train loss: 0.12371716603636741, Test loss: 0.14739787578582764\n", + "Starting epoch 1037\n", + "Train loss: 0.12331763997673989, Test loss: 0.1473090648651123\n", + "Starting epoch 1038\n", + "Train loss: 0.12360805720090866, Test loss: 0.14430664479732513\n", + "Starting epoch 1039\n", + "Train loss: 0.12409818023443223, Test loss: 0.1435972899198532\n", + "Starting epoch 1040\n", + "Train loss: 0.12331635370850563, Test loss: 0.1450592428445816\n", + "Starting epoch 1041\n", + "Train loss: 0.12404737949371338, Test loss: 0.14630009233951569\n", + "Starting epoch 1042\n", + "Train loss: 0.12346553310751915, Test loss: 0.14428302645683289\n", + "Starting epoch 1043\n", + "Train loss: 0.12301335498690605, Test loss: 0.14143255352973938\n", + "Starting epoch 1044\n", + "Train loss: 0.12268982112407684, Test loss: 0.1428300440311432\n", + "Starting epoch 1045\n", + "Train loss: 0.12294131085276604, Test loss: 0.142765074968338\n", + "Starting epoch 1046\n", + "Train loss: 0.12306503996253014, Test loss: 0.14791318774223328\n", + "Starting epoch 1047\n", + "Train loss: 0.12296104654669762, Test loss: 0.1434955596923828\n", + "Starting epoch 1048\n", + "Train loss: 0.12322091653943062, Test loss: 0.13997237384319305\n", + "Starting epoch 1049\n", + "Train loss: 0.12305981874465942, Test loss: 0.1437879204750061\n", + "Starting epoch 1050\n", + "Train loss: 0.12328253120183945, Test loss: 0.14465832710266113\n", + "Starting epoch 1051\n", + "Train loss: 0.12261592462658882, Test loss: 0.1435922533273697\n", + "Starting epoch 1052\n", + "Train loss: 0.12267290741205215, Test loss: 0.14262424409389496\n", + "Starting epoch 1053\n", + "Train loss: 0.12283390641212463, Test loss: 0.13974323868751526\n", + "Starting epoch 1054\n", + "Train loss: 0.1228923763334751, Test loss: 0.1407404989004135\n", + "Starting epoch 1055\n", + "Train loss: 0.12305669918656349, Test loss: 0.1423315703868866\n", + "Starting epoch 1056\n", + "Train loss: 0.12288783460855485, Test loss: 0.14818516373634338\n", + "Starting epoch 1057\n", + "Train loss: 0.12252765282988548, Test loss: 0.14182105660438538\n", + "Starting epoch 1058\n", + "Train loss: 0.12276507288217545, Test loss: 0.1416722983121872\n", + "Starting epoch 1059\n", + "Train loss: 0.1225700867176056, Test loss: 0.14223352074623108\n", + "Starting epoch 1060\n", + "Train loss: 0.12270667463541031, Test loss: 0.1410994827747345\n", + "Starting epoch 1061\n", + "Train loss: 0.12215959519147873, Test loss: 0.14048990607261658\n", + "Starting epoch 1062\n", + "Train loss: 0.12208323106169701, Test loss: 0.14207890629768372\n", + "Starting epoch 1063\n", + "Train loss: 0.12269745439291001, Test loss: 0.1444937139749527\n", + "Starting epoch 1064\n", + "Train loss: 0.12221734136343003, Test loss: 0.13946665823459625\n", + "Starting epoch 1065\n", + "Train loss: 0.12217535093426704, Test loss: 0.14051960408687592\n", + "Starting epoch 1066\n", + "Train loss: 0.12230459526181221, Test loss: 0.1395007222890854\n", + "Starting epoch 1067\n", + "Train loss: 0.12254338994622231, Test loss: 0.13984400033950806\n", + "Starting epoch 1068\n", + "Train loss: 0.12234948739409447, Test loss: 0.14236338436603546\n", + "Starting epoch 1069\n", + "Train loss: 0.12254185795783996, Test loss: 0.1407565027475357\n", + "Starting epoch 1070\n", + "Train loss: 0.12287004709243775, Test loss: 0.13911186158657074\n", + "Starting epoch 1071\n", + "Train loss: 0.12194714218378067, Test loss: 0.14029328525066376\n", + "Starting epoch 1072\n", + "Train loss: 0.12188627526164054, Test loss: 0.14029265940189362\n", + "Starting epoch 1073\n", + "Train loss: 0.12224057167768479, Test loss: 0.13920828700065613\n", + "Starting epoch 1074\n", + "Train loss: 0.12185277745127678, Test loss: 0.13859187066555023\n", + "Starting epoch 1075\n", + "Train loss: 0.12231324255466461, Test loss: 0.14037197828292847\n", + "Starting epoch 1076\n", + "Train loss: 0.12225195825099945, Test loss: 0.1396687626838684\n", + "Starting epoch 1077\n", + "Train loss: 0.12239068299531937, Test loss: 0.1406034231185913\n", + "Starting epoch 1078\n", + "Train loss: 0.12218617454171181, Test loss: 0.13978311419487\n", + "Starting epoch 1079\n", + "Train loss: 0.12187314853072166, Test loss: 0.14064842462539673\n", + "Starting epoch 1080\n", + "Train loss: 0.12203539192676544, Test loss: 0.13978679478168488\n", + "Starting epoch 1081\n", + "Train loss: 0.12187922418117524, Test loss: 0.14065806567668915\n", + "Starting epoch 1082\n", + "Train loss: 0.12202485263347626, Test loss: 0.13907688856124878\n", + "Starting epoch 1083\n", + "Train loss: 0.1228077457845211, Test loss: 0.14194992184638977\n", + "Starting epoch 1084\n", + "Train loss: 0.12222039312124253, Test loss: 0.14180287718772888\n", + "Starting epoch 1085\n", + "Train loss: 0.1218537312746048, Test loss: 0.13979899883270264\n", + "Starting epoch 1086\n", + "Train loss: 0.12183201566338539, Test loss: 0.1398393213748932\n", + "Starting epoch 1087\n", + "Train loss: 0.12140804439783097, Test loss: 0.14072120189666748\n", + "Starting epoch 1088\n", + "Train loss: 0.12174391344189645, Test loss: 0.13923180103302002\n", + "Starting epoch 1089\n", + "Train loss: 0.12185578256845474, Test loss: 0.14429299533367157\n", + "Starting epoch 1090\n", + "Train loss: 0.12246060982346535, Test loss: 0.1474047154188156\n", + "Starting epoch 1091\n", + "Train loss: 0.1217084388434887, Test loss: 0.14010058343410492\n", + "Starting epoch 1092\n", + "Train loss: 0.12273768320679665, Test loss: 0.13955841958522797\n", + "Starting epoch 1093\n", + "Train loss: 0.12170530691742897, Test loss: 0.14021477103233337\n", + "Starting epoch 1094\n", + "Train loss: 0.12150283947587014, Test loss: 0.13792751729488373\n", + "Starting epoch 1095\n", + "Train loss: 0.12215218022465706, Test loss: 0.14041805267333984\n", + "Starting epoch 1096\n", + "Train loss: 0.12237241730093956, Test loss: 0.13987237215042114\n", + "Starting epoch 1097\n", + "Train loss: 0.1215892530977726, Test loss: 0.14085710048675537\n", + "Starting epoch 1098\n", + "Train loss: 0.12185326531529427, Test loss: 0.14244681596755981\n", + "Starting epoch 1099\n", + "Train loss: 0.12149014547467232, Test loss: 0.1400681436061859\n", + "Starting epoch 1100\n", + "Train loss: 0.12201826229691505, Test loss: 0.1428574174642563\n", + "Starting epoch 1101\n", + "Train loss: 0.12109823033213615, Test loss: 0.1376180499792099\n", + "Starting epoch 1102\n", + "Train loss: 0.1227036701142788, Test loss: 0.14411014318466187\n", + "Starting epoch 1103\n", + "Train loss: 0.12134160861372947, Test loss: 0.14171557128429413\n", + "Starting epoch 1104\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.12145571231842041, Test loss: 0.13785973191261292\n", + "Starting epoch 1105\n", + "Train loss: 0.12147428050637245, Test loss: 0.1381322592496872\n", + "Starting epoch 1106\n", + "Train loss: 0.12172526061534882, Test loss: 0.14048363268375397\n", + "Starting epoch 1107\n", + "Train loss: 0.1215949809551239, Test loss: 0.13873033225536346\n", + "Starting epoch 1108\n", + "Train loss: 0.12224166437983514, Test loss: 0.1451120376586914\n", + "Starting epoch 1109\n", + "Train loss: 0.1220216628909111, Test loss: 0.1379215568304062\n", + "Starting epoch 1110\n", + "Train loss: 0.12120642751455307, Test loss: 0.13696806132793427\n", + "Starting epoch 1111\n", + "Train loss: 0.12129907503724098, Test loss: 0.14293399453163147\n", + "Starting epoch 1112\n", + "Train loss: 0.12173706009984016, Test loss: 0.14365333318710327\n", + "Starting epoch 1113\n", + "Train loss: 0.121847093552351, Test loss: 0.13740374147891998\n", + "Starting epoch 1114\n", + "Train loss: 0.1215089650452137, Test loss: 0.14545950293540955\n", + "Starting epoch 1115\n", + "Train loss: 0.12126088663935661, Test loss: 0.13779468834400177\n", + "Starting epoch 1116\n", + "Train loss: 0.12205142185091972, Test loss: 0.1395784318447113\n", + "Starting epoch 1117\n", + "Train loss: 0.12105973213911056, Test loss: 0.13692106306552887\n", + "Starting epoch 1118\n", + "Train loss: 0.12137750819325448, Test loss: 0.1368221640586853\n", + "Starting epoch 1119\n", + "Train loss: 0.12186736017465591, Test loss: 0.1368716061115265\n", + "Starting epoch 1120\n", + "Train loss: 0.12159441098570824, Test loss: 0.13804234564304352\n", + "Starting epoch 1121\n", + "Train loss: 0.12188691794872283, Test loss: 0.14105965197086334\n", + "Starting epoch 1122\n", + "Train loss: 0.12130929619073867, Test loss: 0.13696177303791046\n", + "Starting epoch 1123\n", + "Train loss: 0.12112393617630005, Test loss: 0.13821172714233398\n", + "Starting epoch 1124\n", + "Train loss: 0.12206118419766426, Test loss: 0.14127230644226074\n", + "Starting epoch 1125\n", + "Train loss: 0.12126134559512139, Test loss: 0.14164011180400848\n", + "Starting epoch 1126\n", + "Train loss: 0.12129633396863937, Test loss: 0.1385437399148941\n", + "Starting epoch 1127\n", + "Train loss: 0.12132321447134017, Test loss: 0.14056584239006042\n", + "Starting epoch 1128\n", + "Train loss: 0.12227543488144875, Test loss: 0.14631259441375732\n", + "Starting epoch 1129\n", + "Train loss: 0.12132158324122429, Test loss: 0.13805310428142548\n", + "Starting epoch 1130\n", + "Train loss: 0.12167798936367034, Test loss: 0.14183203876018524\n", + "Starting epoch 1131\n", + "Train loss: 0.12100752010941505, Test loss: 0.14051927626132965\n", + "Starting epoch 1132\n", + "Train loss: 0.12176860630512237, Test loss: 0.13780595362186432\n", + "Starting epoch 1133\n", + "Train loss: 0.12156914740800857, Test loss: 0.14420020580291748\n", + "Starting epoch 1134\n", + "Train loss: 0.12098820328712463, Test loss: 0.13778936862945557\n", + "Starting epoch 1135\n", + "Train loss: 0.12185667291283607, Test loss: 0.13920235633850098\n", + "Starting epoch 1136\n", + "Train loss: 0.12149015069007874, Test loss: 0.13827326893806458\n", + "Starting epoch 1137\n", + "Train loss: 0.12167832851409913, Test loss: 0.14105795323848724\n", + "Starting epoch 1138\n", + "Train loss: 0.12177433416247369, Test loss: 0.1372978687286377\n", + "Starting epoch 1139\n", + "Train loss: 0.12077568978071213, Test loss: 0.1380738765001297\n", + "Starting epoch 1140\n", + "Train loss: 0.12106968909502029, Test loss: 0.14304892718791962\n", + "Starting epoch 1141\n", + "Train loss: 0.12169467389583588, Test loss: 0.14080530405044556\n", + "Starting epoch 1142\n", + "Train loss: 0.12120945170521737, Test loss: 0.14276526868343353\n", + "Starting epoch 1143\n", + "Train loss: 0.12178347676992417, Test loss: 0.13627062737941742\n", + "Starting epoch 1144\n", + "Train loss: 0.12122768491506576, Test loss: 0.1407669186592102\n", + "Starting epoch 1145\n", + "Train loss: 0.12144717633724213, Test loss: 0.13863235712051392\n", + "Starting epoch 1146\n", + "Train loss: 0.12154378965497017, Test loss: 0.13984644412994385\n", + "Starting epoch 1147\n", + "Train loss: 0.12160655096173287, Test loss: 0.1377655565738678\n", + "Starting epoch 1148\n", + "Train loss: 0.12145428344607354, Test loss: 0.1415892243385315\n", + "Starting epoch 1149\n", + "Train loss: 0.12117749914526939, Test loss: 0.14195860922336578\n", + "Starting epoch 1150\n", + "Train loss: 0.12127991512417793, Test loss: 0.1368289738893509\n", + "Starting epoch 1151\n", + "Train loss: 0.12123623609542847, Test loss: 0.13927127420902252\n", + "Starting epoch 1152\n", + "Train loss: 0.12114093124866486, Test loss: 0.1360926777124405\n", + "Starting epoch 1153\n", + "Train loss: 0.12074399694800377, Test loss: 0.14199216663837433\n", + "Starting epoch 1154\n", + "Train loss: 0.12134233579039573, Test loss: 0.14269469678401947\n", + "Starting epoch 1155\n", + "Train loss: 0.12092884421348572, Test loss: 0.13676704466342926\n", + "Starting epoch 1156\n", + "Train loss: 0.12188708052039146, Test loss: 0.1367558091878891\n", + "Starting epoch 1157\n", + "Train loss: 0.12125865623354912, Test loss: 0.13878417015075684\n", + "Starting epoch 1158\n", + "Train loss: 0.12070781543850899, Test loss: 0.13876573741436005\n", + "Starting epoch 1159\n", + "Train loss: 0.12170836687088013, Test loss: 0.13751435279846191\n", + "Starting epoch 1160\n", + "Train loss: 0.12121883675456047, Test loss: 0.14135347306728363\n", + "Starting epoch 1161\n", + "Train loss: 0.12109464719891548, Test loss: 0.1364954710006714\n", + "Starting epoch 1162\n", + "Train loss: 0.12166286215186119, Test loss: 0.13605670630931854\n", + "Starting epoch 1163\n", + "Train loss: 0.12134108930826187, Test loss: 0.14040511846542358\n", + "Starting epoch 1164\n", + "Train loss: 0.12057961195707322, Test loss: 0.1379510760307312\n", + "Starting epoch 1165\n", + "Train loss: 0.12076561838388443, Test loss: 0.13651636242866516\n", + "Starting epoch 1166\n", + "Train loss: 0.12160562187433242, Test loss: 0.13825175166130066\n", + "Starting epoch 1167\n", + "Train loss: 0.12184818357229232, Test loss: 0.1428174376487732\n", + "Starting epoch 1168\n", + "Train loss: 0.12102273538708687, Test loss: 0.1373734027147293\n", + "Starting epoch 1169\n", + "Train loss: 0.12109145686030388, Test loss: 0.13753686845302582\n", + "Starting epoch 1170\n", + "Train loss: 0.12114107206463814, Test loss: 0.13653020560741425\n", + "Starting epoch 1171\n", + "Train loss: 0.1208795015513897, Test loss: 0.1398748755455017\n", + "Starting epoch 1172\n", + "Train loss: 0.12150591358542442, Test loss: 0.13858245313167572\n", + "Starting epoch 1173\n", + "Train loss: 0.12093767628073693, Test loss: 0.14240571856498718\n", + "Starting epoch 1174\n", + "Train loss: 0.12159583494067192, Test loss: 0.13874536752700806\n", + "Starting epoch 1175\n", + "Train loss: 0.12140857636928558, Test loss: 0.13670609891414642\n", + "Starting epoch 1176\n", + "Train loss: 0.12104270130395889, Test loss: 0.13717815279960632\n", + "Starting epoch 1177\n", + "Train loss: 0.12203403100371361, Test loss: 0.13796944916248322\n", + "Starting epoch 1178\n", + "Train loss: 0.12138248085975648, Test loss: 0.13695839047431946\n", + "Starting epoch 1179\n", + "Train loss: 0.12096061035990716, Test loss: 0.13959035277366638\n", + "Starting epoch 1180\n", + "Train loss: 0.12083294987678528, Test loss: 0.13670873641967773\n", + "Starting epoch 1181\n", + "Train loss: 0.12090277269482613, Test loss: 0.1370653361082077\n", + "Starting epoch 1182\n", + "Train loss: 0.12106125101447106, Test loss: 0.13899517059326172\n", + "Starting epoch 1183\n", + "Train loss: 0.12053712695837021, Test loss: 0.136060893535614\n", + "Starting epoch 1184\n", + "Train loss: 0.1213914531469345, Test loss: 0.13811556994915009\n", + "Starting epoch 1185\n", + "Train loss: 0.12113184705376626, Test loss: 0.13812272250652313\n", + "Starting epoch 1186\n", + "Train loss: 0.12095858871936799, Test loss: 0.13870082795619965\n", + "Starting epoch 1187\n", + "Train loss: 0.12105348706245422, Test loss: 0.1351557970046997\n", + "Starting epoch 1188\n", + "Train loss: 0.1208814398944378, Test loss: 0.13701054453849792\n", + "Starting epoch 1189\n", + "Train loss: 0.12094866827130318, Test loss: 0.1433851718902588\n", + "Starting epoch 1190\n", + "Train loss: 0.12158679082989693, Test loss: 0.13782373070716858\n", + "Starting epoch 1191\n", + "Train loss: 0.12091946467757225, Test loss: 0.13577334582805634\n", + "Starting epoch 1192\n", + "Train loss: 0.1209359711408615, Test loss: 0.1376786231994629\n", + "Starting epoch 1193\n", + "Train loss: 0.12202261969447135, Test loss: 0.1363651156425476\n", + "Starting epoch 1194\n", + "Train loss: 0.12076823070645332, Test loss: 0.13771584630012512\n", + "Starting epoch 1195\n", + "Train loss: 0.12090495467185974, Test loss: 0.1404133141040802\n", + "Starting epoch 1196\n", + "Train loss: 0.12119431763887406, Test loss: 0.1354600191116333\n", + "Starting epoch 1197\n", + "Train loss: 0.12127105727791786, Test loss: 0.13659000396728516\n", + "Starting epoch 1198\n", + "Train loss: 0.12091465577483178, Test loss: 0.1350233107805252\n", + "Starting epoch 1199\n", + "Train loss: 0.12099265456199645, Test loss: 0.13743606209754944\n", + "Starting epoch 1200\n", + "Train loss: 0.12101323693990708, Test loss: 0.13888701796531677\n", + "Starting epoch 1201\n", + "Train loss: 0.12163012027740479, Test loss: 0.13711999356746674\n", + "Starting epoch 1202\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.120838353484869, Test loss: 0.14146456122398376\n", + "Starting epoch 1203\n", + "Train loss: 0.12100284889340401, Test loss: 0.13717125356197357\n", + "Starting epoch 1204\n", + "Train loss: 0.12056910797953606, Test loss: 0.1369127333164215\n", + "Starting epoch 1205\n", + "Train loss: 0.12062723934650421, Test loss: 0.14102469384670258\n", + "Starting epoch 1206\n", + "Train loss: 0.12104122415184974, Test loss: 0.13946467638015747\n", + "Starting epoch 1207\n", + "Train loss: 0.12102294921875, Test loss: 0.1367993950843811\n", + "Starting epoch 1208\n", + "Train loss: 0.12158467948436737, Test loss: 0.13603271543979645\n", + "Starting epoch 1209\n", + "Train loss: 0.12156140595674515, Test loss: 0.13860897719860077\n", + "Starting epoch 1210\n", + "Train loss: 0.12073223322629928, Test loss: 0.1359201818704605\n", + "Starting epoch 1211\n", + "Train loss: 0.1206938622891903, Test loss: 0.13651977479457855\n", + "Starting epoch 1212\n", + "Train loss: 0.12089804917573929, Test loss: 0.14466604590415955\n", + "Starting epoch 1213\n", + "Train loss: 0.12247686192393303, Test loss: 0.13782358169555664\n", + "Starting epoch 1214\n", + "Train loss: 0.12104405790567398, Test loss: 0.13700932264328003\n", + "Starting epoch 1215\n", + "Train loss: 0.12048729136586189, Test loss: 0.13827897608280182\n", + "Starting epoch 1216\n", + "Train loss: 0.12068433493375778, Test loss: 0.13637712597846985\n", + "Starting epoch 1217\n", + "Train loss: 0.12065253496170043, Test loss: 0.13701771199703217\n", + "Starting epoch 1218\n", + "Train loss: 0.12072966977953911, Test loss: 0.13641755282878876\n", + "Starting epoch 1219\n", + "Train loss: 0.12076361387968064, Test loss: 0.13620580732822418\n", + "Starting epoch 1220\n", + "Train loss: 0.12124882385134697, Test loss: 0.13956084847450256\n", + "Starting epoch 1221\n", + "Train loss: 0.12112958237528801, Test loss: 0.13730663061141968\n", + "Starting epoch 1222\n", + "Train loss: 0.12131357565522194, Test loss: 0.13625404238700867\n", + "Starting epoch 1223\n", + "Train loss: 0.12070394918322563, Test loss: 0.13804768025875092\n", + "Starting epoch 1224\n", + "Train loss: 0.12112667545676231, Test loss: 0.13793013989925385\n", + "Starting epoch 1225\n", + "Train loss: 0.12101539611816406, Test loss: 0.1401452273130417\n", + "Starting epoch 1226\n", + "Train loss: 0.12110535860061646, Test loss: 0.1385573297739029\n", + "Starting epoch 1227\n", + "Train loss: 0.12078710719943046, Test loss: 0.13870146870613098\n", + "Starting epoch 1228\n", + "Train loss: 0.12089278370141983, Test loss: 0.142944797873497\n", + "Starting epoch 1229\n", + "Train loss: 0.12140817746520043, Test loss: 0.1385338008403778\n", + "Starting epoch 1230\n", + "Train loss: 0.12058585345745086, Test loss: 0.13646192848682404\n", + "Starting epoch 1231\n", + "Train loss: 0.12066169887781143, Test loss: 0.1355256885290146\n", + "Starting epoch 1232\n", + "Train loss: 0.12144963681697846, Test loss: 0.1443382352590561\n", + "Starting epoch 1233\n", + "Train loss: 0.12118595361709594, Test loss: 0.13995005190372467\n", + "Starting epoch 1234\n", + "Train loss: 0.1211702510714531, Test loss: 0.14078012108802795\n", + "Starting epoch 1235\n", + "Train loss: 0.12073475599288941, Test loss: 0.13839854300022125\n", + "Starting epoch 1236\n", + "Train loss: 0.12105272069573403, Test loss: 0.1371650993824005\n", + "Starting epoch 1237\n", + "Train loss: 0.12049622789025306, Test loss: 0.13982418179512024\n", + "Starting epoch 1238\n", + "Train loss: 0.12100772336125373, Test loss: 0.1360250860452652\n", + "Starting epoch 1239\n", + "Train loss: 0.12066143795847893, Test loss: 0.13576039671897888\n", + "Starting epoch 1240\n", + "Train loss: 0.12060929387807846, Test loss: 0.13715864717960358\n", + "Starting epoch 1241\n", + "Train loss: 0.12079924672842025, Test loss: 0.13956798613071442\n", + "Starting epoch 1242\n", + "Train loss: 0.12121429696679115, Test loss: 0.13907422125339508\n", + "Starting epoch 1243\n", + "Train loss: 0.1214459054172039, Test loss: 0.13690216839313507\n", + "Starting epoch 1244\n", + "Train loss: 0.1208901758491993, Test loss: 0.13764864206314087\n", + "Starting epoch 1245\n", + "Train loss: 0.12095197394490242, Test loss: 0.13565826416015625\n", + "Starting epoch 1246\n", + "Train loss: 0.12118738859891892, Test loss: 0.1370471864938736\n", + "Starting epoch 1247\n", + "Train loss: 0.12189964443445206, Test loss: 0.1439361721277237\n", + "Starting epoch 1248\n", + "Train loss: 0.12080284982919692, Test loss: 0.1354314684867859\n", + "Starting epoch 1249\n", + "Train loss: 0.12097077056765557, Test loss: 0.1380508840084076\n", + "Starting epoch 1250\n", + "Train loss: 0.12109752714633942, Test loss: 0.14211532473564148\n", + "Starting epoch 1251\n", + "Train loss: 0.12085006073117256, Test loss: 0.1356135755777359\n", + "Starting epoch 1252\n", + "Train loss: 0.12066674143075944, Test loss: 0.1369647979736328\n", + "Starting epoch 1253\n", + "Train loss: 0.12064056023955345, Test loss: 0.13656538724899292\n", + "Starting epoch 1254\n", + "Train loss: 0.12064755395054817, Test loss: 0.13564105331897736\n", + "Starting epoch 1255\n", + "Train loss: 0.12071744367480278, Test loss: 0.13803941011428833\n", + "Starting epoch 1256\n", + "Train loss: 0.12051426574587822, Test loss: 0.1366981416940689\n", + "Starting epoch 1257\n", + "Train loss: 0.12076156735420226, Test loss: 0.136569544672966\n", + "Starting epoch 1258\n", + "Train loss: 0.12105354189872741, Test loss: 0.14420413970947266\n", + "Starting epoch 1259\n", + "Train loss: 0.1217434911429882, Test loss: 0.13703103363513947\n", + "Starting epoch 1260\n", + "Train loss: 0.12085046887397766, Test loss: 0.13620615005493164\n", + "Starting epoch 1261\n", + "Train loss: 0.1204320365190506, Test loss: 0.14117549359798431\n", + "Starting epoch 1262\n", + "Train loss: 0.12087816953659057, Test loss: 0.1355520337820053\n", + "Starting epoch 1263\n", + "Train loss: 0.12066713616251945, Test loss: 0.13824795186519623\n", + "Starting epoch 1264\n", + "Train loss: 0.12064923554658889, Test loss: 0.13584396243095398\n", + "Starting epoch 1265\n", + "Train loss: 0.12091498523950577, Test loss: 0.13575303554534912\n", + "Starting epoch 1266\n", + "Train loss: 0.12065885096788406, Test loss: 0.1423725038766861\n", + "Starting epoch 1267\n", + "Train loss: 0.1209659680724144, Test loss: 0.13553094863891602\n", + "Starting epoch 1268\n", + "Train loss: 0.12051469445228577, Test loss: 0.13457350432872772\n", + "Starting epoch 1269\n", + "Train loss: 0.12052518263459205, Test loss: 0.13872666656970978\n", + "Starting epoch 1270\n", + "Train loss: 0.12121908903121949, Test loss: 0.1369893103837967\n", + "Starting epoch 1271\n", + "Train loss: 0.12112831562757492, Test loss: 0.13672025501728058\n", + "Starting epoch 1272\n", + "Train loss: 0.12077367454767227, Test loss: 0.13887207210063934\n", + "Starting epoch 1273\n", + "Train loss: 0.12110977366566658, Test loss: 0.1375582218170166\n", + "Starting epoch 1274\n", + "Train loss: 0.12096869960427284, Test loss: 0.1356685906648636\n", + "Starting epoch 1275\n", + "Train loss: 0.12043831408023835, Test loss: 0.13793745636940002\n", + "Starting epoch 1276\n", + "Train loss: 0.12125621572136878, Test loss: 0.1371603012084961\n", + "Starting epoch 1277\n", + "Train loss: 0.12157043308019638, Test loss: 0.13695518672466278\n", + "Starting epoch 1278\n", + "Train loss: 0.12087330043315887, Test loss: 0.13681349158287048\n", + "Starting epoch 1279\n", + "Train loss: 0.12051288425922393, Test loss: 0.13703227043151855\n", + "Starting epoch 1280\n", + "Train loss: 0.1203566874563694, Test loss: 0.13548865914344788\n", + "Starting epoch 1281\n", + "Train loss: 0.12093723580241203, Test loss: 0.13847677409648895\n", + "Starting epoch 1282\n", + "Train loss: 0.12096379965543747, Test loss: 0.13782896101474762\n", + "Starting epoch 1283\n", + "Train loss: 0.12049737617373467, Test loss: 0.13603931665420532\n", + "Starting epoch 1284\n", + "Train loss: 0.12056446343660354, Test loss: 0.14221079647541046\n", + "Starting epoch 1285\n", + "Train loss: 0.1207563503086567, Test loss: 0.1375674605369568\n", + "Starting epoch 1286\n", + "Train loss: 0.12115773171186447, Test loss: 0.13568495213985443\n", + "Starting epoch 1287\n", + "Train loss: 0.1208443994820118, Test loss: 0.13550318777561188\n", + "Starting epoch 1288\n", + "Train loss: 0.1204143825173378, Test loss: 0.13566339015960693\n", + "Starting epoch 1289\n", + "Train loss: 0.12052393510937691, Test loss: 0.14310665428638458\n", + "Starting epoch 1290\n", + "Train loss: 0.12133467197418213, Test loss: 0.14142540097236633\n", + "Starting epoch 1291\n", + "Train loss: 0.12114878937602043, Test loss: 0.1368180513381958\n", + "Starting epoch 1292\n", + "Train loss: 0.1210230815410614, Test loss: 0.13689963519573212\n", + "Starting epoch 1293\n", + "Train loss: 0.12174984082579612, Test loss: 0.13953429460525513\n", + "Starting epoch 1294\n", + "Train loss: 0.12151720225811005, Test loss: 0.1428912878036499\n", + "Starting epoch 1295\n", + "Train loss: 0.12050129503011703, Test loss: 0.1367955505847931\n", + "Starting epoch 1296\n", + "Train loss: 0.12060604333877563, Test loss: 0.1449865847826004\n", + "Starting epoch 1297\n", + "Train loss: 0.12084199890494346, Test loss: 0.13580743968486786\n", + "Starting epoch 1298\n", + "Train loss: 0.12069080501794816, Test loss: 0.13611194491386414\n", + "Starting epoch 1299\n", + "Train loss: 0.12064536616206169, Test loss: 0.13652071356773376\n", + "Starting epoch 1300\n", + "Train loss: 0.12064963191747666, Test loss: 0.13675446808338165\n", + "Starting epoch 1301\n", + "Train loss: 0.12098412528634071, Test loss: 0.14149390161037445\n", + "Starting epoch 1302\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.12091241657733917, Test loss: 0.13924004137516022\n", + "Starting epoch 1303\n", + "Train loss: 0.12083302393555641, Test loss: 0.13873647153377533\n", + "Starting epoch 1304\n", + "Train loss: 0.1208796414732933, Test loss: 0.13763676583766937\n", + "Starting epoch 1305\n", + "Train loss: 0.12037241950631142, Test loss: 0.13811121881008148\n", + "Starting epoch 1306\n", + "Train loss: 0.12047268137335777, Test loss: 0.13859997689723969\n", + "Starting epoch 1307\n", + "Train loss: 0.12027598530054093, Test loss: 0.13551820814609528\n", + "Starting epoch 1308\n", + "Train loss: 0.12130687534809112, Test loss: 0.1395208090543747\n", + "Starting epoch 1309\n", + "Train loss: 0.12082116112112999, Test loss: 0.13670268654823303\n", + "Starting epoch 1310\n", + "Train loss: 0.1207766716182232, Test loss: 0.13687381148338318\n", + "Starting epoch 1311\n", + "Train loss: 0.12020348727703095, Test loss: 0.13569040596485138\n", + "Starting epoch 1312\n", + "Train loss: 0.12067459672689437, Test loss: 0.1373436599969864\n", + "Starting epoch 1313\n", + "Train loss: 0.1208046917617321, Test loss: 0.1429954469203949\n", + "Starting epoch 1314\n", + "Train loss: 0.12068522512912751, Test loss: 0.13760775327682495\n", + "Starting epoch 1315\n", + "Train loss: 0.12068765923380852, Test loss: 0.13593970239162445\n", + "Starting epoch 1316\n", + "Train loss: 0.12110899969935417, Test loss: 0.1361667662858963\n", + "Starting epoch 1317\n", + "Train loss: 0.12064239472150802, Test loss: 0.13822731375694275\n", + "Starting epoch 1318\n", + "Train loss: 0.12142160072922707, Test loss: 0.1363891363143921\n", + "Starting epoch 1319\n", + "Train loss: 0.12048447489738465, Test loss: 0.1371108591556549\n", + "Starting epoch 1320\n", + "Train loss: 0.12035660311579705, Test loss: 0.13543406128883362\n", + "Starting epoch 1321\n", + "Train loss: 0.12077385127544403, Test loss: 0.13656185567378998\n", + "Starting epoch 1322\n", + "Train loss: 0.1205257023870945, Test loss: 0.13717339932918549\n", + "Starting epoch 1323\n", + "Train loss: 0.12071104913949966, Test loss: 0.13511058688163757\n", + "Starting epoch 1324\n", + "Train loss: 0.12095920592546464, Test loss: 0.1348484307527542\n", + "Starting epoch 1325\n", + "Train loss: 0.12054912731051445, Test loss: 0.13611018657684326\n", + "Starting epoch 1326\n", + "Train loss: 0.12041939705610276, Test loss: 0.1392689049243927\n", + "Starting epoch 1327\n", + "Train loss: 0.12040262252092361, Test loss: 0.13507722318172455\n", + "Starting epoch 1328\n", + "Train loss: 0.1204539668560028, Test loss: 0.13631245493888855\n", + "Starting epoch 1329\n", + "Train loss: 0.12102331876754761, Test loss: 0.1440344601869583\n", + "Starting epoch 1330\n", + "Train loss: 0.12047790125012398, Test loss: 0.1372399926185608\n", + "Starting epoch 1331\n", + "Train loss: 0.12084952756762504, Test loss: 0.13777050375938416\n", + "Starting epoch 1332\n", + "Train loss: 0.12055905088782311, Test loss: 0.13688333332538605\n", + "Starting epoch 1333\n", + "Train loss: 0.12100274443626403, Test loss: 0.13477648794651031\n", + "Starting epoch 1334\n", + "Train loss: 0.12084962576627731, Test loss: 0.1388523131608963\n", + "Starting epoch 1335\n", + "Train loss: 0.12065161913633346, Test loss: 0.1404314935207367\n", + "Starting epoch 1336\n", + "Train loss: 0.12056343868374825, Test loss: 0.13632623851299286\n", + "Starting epoch 1337\n", + "Train loss: 0.12056624636054039, Test loss: 0.1350579857826233\n", + "Starting epoch 1338\n", + "Train loss: 0.12060620948672295, Test loss: 0.13716204464435577\n", + "Starting epoch 1339\n", + "Train loss: 0.12067511409521103, Test loss: 0.13872624933719635\n", + "Starting epoch 1340\n", + "Train loss: 0.1205928647518158, Test loss: 0.13639555871486664\n", + "Starting epoch 1341\n", + "Train loss: 0.12088397055864335, Test loss: 0.13680116832256317\n", + "Starting epoch 1342\n", + "Train loss: 0.12068705633282661, Test loss: 0.13652928173542023\n", + "Starting epoch 1343\n", + "Train loss: 0.12057030394673347, Test loss: 0.13574610650539398\n", + "Starting epoch 1344\n", + "Train loss: 0.1208455193042755, Test loss: 0.13766328990459442\n", + "Starting epoch 1345\n", + "Train loss: 0.12038674548268319, Test loss: 0.13564197719097137\n", + "Starting epoch 1346\n", + "Train loss: 0.12053958684206009, Test loss: 0.13564671576023102\n", + "Starting epoch 1347\n", + "Train loss: 0.12043206304311753, Test loss: 0.13712039589881897\n", + "Starting epoch 1348\n", + "Train loss: 0.12075709626078605, Test loss: 0.13994719088077545\n", + "Starting epoch 1349\n", + "Train loss: 0.12043715700507164, Test loss: 0.13674956560134888\n", + "Starting epoch 1350\n", + "Train loss: 0.12038759171962737, Test loss: 0.13834403455257416\n", + "Starting epoch 1351\n", + "Train loss: 0.12059404507279396, Test loss: 0.1361084282398224\n", + "Starting epoch 1352\n", + "Train loss: 0.12062872797250748, Test loss: 0.13595809042453766\n", + "Starting epoch 1353\n", + "Train loss: 0.12070746451616288, Test loss: 0.13575629889965057\n", + "Starting epoch 1354\n", + "Train loss: 0.1209071235358715, Test loss: 0.13654153048992157\n", + "Starting epoch 1355\n", + "Train loss: 0.12073741883039474, Test loss: 0.14093343913555145\n", + "Starting epoch 1356\n", + "Train loss: 0.12038714379072189, Test loss: 0.13841865956783295\n", + "Starting epoch 1357\n", + "Train loss: 0.12063498690724372, Test loss: 0.1406656801700592\n", + "Starting epoch 1358\n", + "Train loss: 0.12062075823545455, Test loss: 0.13616728782653809\n", + "Starting epoch 1359\n", + "Train loss: 0.1204006065428257, Test loss: 0.13772712647914886\n", + "Starting epoch 1360\n", + "Train loss: 0.1208703327178955, Test loss: 0.13553625345230103\n", + "Starting epoch 1361\n", + "Train loss: 0.12054889842867851, Test loss: 0.1388935148715973\n", + "Starting epoch 1362\n", + "Train loss: 0.12066675618290901, Test loss: 0.13636165857315063\n", + "Starting epoch 1363\n", + "Train loss: 0.1204876047372818, Test loss: 0.13570162653923035\n", + "Starting epoch 1364\n", + "Train loss: 0.12081346169114113, Test loss: 0.1361912041902542\n", + "Starting epoch 1365\n", + "Train loss: 0.12072432890534401, Test loss: 0.13752591609954834\n", + "Starting epoch 1366\n", + "Train loss: 0.12062060594558716, Test loss: 0.13565905392169952\n", + "Starting epoch 1367\n", + "Train loss: 0.12044285759329795, Test loss: 0.13786333799362183\n", + "Starting epoch 1368\n", + "Train loss: 0.12072895750403405, Test loss: 0.13462944328784943\n", + "Starting epoch 1369\n", + "Train loss: 0.12037791669368744, Test loss: 0.13643565773963928\n", + "Starting epoch 1370\n", + "Train loss: 0.12123147532343864, Test loss: 0.13662594556808472\n", + "Starting epoch 1371\n", + "Train loss: 0.12046519994735717, Test loss: 0.13706105947494507\n", + "Starting epoch 1372\n", + "Train loss: 0.1201309558749199, Test loss: 0.13787463307380676\n", + "Starting epoch 1373\n", + "Train loss: 0.1208356986939907, Test loss: 0.13920284807682037\n", + "Starting epoch 1374\n", + "Train loss: 0.12082102119922639, Test loss: 0.13583976030349731\n", + "Starting epoch 1375\n", + "Train loss: 0.12030014291405677, Test loss: 0.13516919314861298\n", + "Starting epoch 1376\n", + "Train loss: 0.12055062800645829, Test loss: 0.13759084045886993\n", + "Starting epoch 1377\n", + "Train loss: 0.1205149830877781, Test loss: 0.13902394473552704\n", + "Starting epoch 1378\n", + "Train loss: 0.12025961846113205, Test loss: 0.13705450296401978\n", + "Starting epoch 1379\n", + "Train loss: 0.12018253147602082, Test loss: 0.13554979860782623\n", + "Starting epoch 1380\n", + "Train loss: 0.12041966766119003, Test loss: 0.13624164462089539\n", + "Starting epoch 1381\n", + "Train loss: 0.12140702724456787, Test loss: 0.13618792593479156\n", + "Starting epoch 1382\n", + "Train loss: 0.12030860602855682, Test loss: 0.13582675158977509\n", + "Starting epoch 1383\n", + "Train loss: 0.12039236277341843, Test loss: 0.13688795268535614\n", + "Starting epoch 1384\n", + "Train loss: 0.12056860014796257, Test loss: 0.14302125573158264\n", + "Starting epoch 1385\n", + "Train loss: 0.12110022827982903, Test loss: 0.13817833364009857\n", + "Starting epoch 1386\n", + "Train loss: 0.12050550088286399, Test loss: 0.13598699867725372\n", + "Starting epoch 1387\n", + "Train loss: 0.12062376454472541, Test loss: 0.14018097519874573\n", + "Starting epoch 1388\n", + "Train loss: 0.1205835311114788, Test loss: 0.13645324110984802\n", + "Starting epoch 1389\n", + "Train loss: 0.12067984536290169, Test loss: 0.14000727236270905\n", + "Starting epoch 1390\n", + "Train loss: 0.12041022807359696, Test loss: 0.1378256380558014\n", + "Starting epoch 1391\n", + "Train loss: 0.12031074479222298, Test loss: 0.13702629506587982\n", + "Starting epoch 1392\n", + "Train loss: 0.12017198443412781, Test loss: 0.13556039333343506\n", + "Starting epoch 1393\n", + "Train loss: 0.12048652097582817, Test loss: 0.13720153272151947\n", + "Starting epoch 1394\n", + "Train loss: 0.12073310315608979, Test loss: 0.14197039604187012\n", + "Starting epoch 1395\n", + "Train loss: 0.12097869858145714, Test loss: 0.14202357828617096\n", + "Starting epoch 1396\n", + "Train loss: 0.12129245653748512, Test loss: 0.1363918036222458\n", + "Starting epoch 1397\n", + "Train loss: 0.12037577018141747, Test loss: 0.13641789555549622\n", + "Starting epoch 1398\n", + "Train loss: 0.12031079068779946, Test loss: 0.13868248462677002\n", + "Starting epoch 1399\n", + "Train loss: 0.12039998829364777, Test loss: 0.13803572952747345\n", + "Starting epoch 1400\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.12039016112685204, Test loss: 0.14095370471477509\n", + "Starting epoch 1401\n", + "Train loss: 0.12015105932950973, Test loss: 0.13528019189834595\n", + "Starting epoch 1402\n", + "Train loss: 0.12046802654862404, Test loss: 0.1360846310853958\n", + "Starting epoch 1403\n", + "Train loss: 0.12065372154116631, Test loss: 0.13601179420948029\n", + "Starting epoch 1404\n", + "Train loss: 0.12025086209177971, Test loss: 0.13617734611034393\n", + "Starting epoch 1405\n", + "Train loss: 0.1205544699728489, Test loss: 0.13596996665000916\n", + "Starting epoch 1406\n", + "Train loss: 0.12074584156274795, Test loss: 0.13656798005104065\n", + "Starting epoch 1407\n", + "Train loss: 0.12059229165315628, Test loss: 0.13555534183979034\n", + "Starting epoch 1408\n", + "Train loss: 0.12097951486706733, Test loss: 0.13631922006607056\n", + "Starting epoch 1409\n", + "Train loss: 0.12056293740868568, Test loss: 0.138677716255188\n", + "Starting epoch 1410\n", + "Train loss: 0.12049082085490227, Test loss: 0.13668979704380035\n", + "Starting epoch 1411\n", + "Train loss: 0.12009624764323235, Test loss: 0.13491106033325195\n", + "Starting epoch 1412\n", + "Train loss: 0.12042392522096634, Test loss: 0.13559715449810028\n", + "Starting epoch 1413\n", + "Train loss: 0.12070728197693825, Test loss: 0.13675656914710999\n", + "Starting epoch 1414\n", + "Train loss: 0.1212418720126152, Test loss: 0.13498219847679138\n", + "Starting epoch 1415\n", + "Train loss: 0.1204225604236126, Test loss: 0.13950619101524353\n", + "Starting epoch 1416\n", + "Train loss: 0.12006911143660545, Test loss: 0.13648481667041779\n", + "Starting epoch 1417\n", + "Train loss: 0.12094771906733513, Test loss: 0.14051559567451477\n", + "Starting epoch 1418\n", + "Train loss: 0.12041380897164344, Test loss: 0.1373811811208725\n", + "Starting epoch 1419\n", + "Train loss: 0.1201261229813099, Test loss: 0.13902005553245544\n", + "Starting epoch 1420\n", + "Train loss: 0.1205300609767437, Test loss: 0.13624845445156097\n", + "Starting epoch 1421\n", + "Train loss: 0.1205480870604515, Test loss: 0.13593754172325134\n", + "Starting epoch 1422\n", + "Train loss: 0.12030088037252426, Test loss: 0.13793723285198212\n", + "Starting epoch 1423\n", + "Train loss: 0.12010476246476173, Test loss: 0.1408115178346634\n", + "Starting epoch 1424\n", + "Train loss: 0.12070865258574486, Test loss: 0.136728897690773\n", + "Starting epoch 1425\n", + "Train loss: 0.12039232462644577, Test loss: 0.13609197735786438\n", + "Starting epoch 1426\n", + "Train loss: 0.12080480337142945, Test loss: 0.13771381974220276\n", + "Starting epoch 1427\n", + "Train loss: 0.12019689112901688, Test loss: 0.13571608066558838\n", + "Starting epoch 1428\n", + "Train loss: 0.12049511969089508, Test loss: 0.13690929114818573\n", + "Starting epoch 1429\n", + "Train loss: 0.12035527110099792, Test loss: 0.13608454167842865\n", + "Starting epoch 1430\n", + "Train loss: 0.12054039999842643, Test loss: 0.13620992004871368\n", + "Starting epoch 1431\n", + "Train loss: 0.12085060626268387, Test loss: 0.13749685883522034\n", + "Starting epoch 1432\n", + "Train loss: 0.11993717879056931, Test loss: 0.13652090728282928\n", + "Starting epoch 1433\n", + "Train loss: 0.12048644185066223, Test loss: 0.13642768561840057\n", + "Starting epoch 1434\n", + "Train loss: 0.12034344404935837, Test loss: 0.13749736547470093\n", + "Starting epoch 1435\n", + "Train loss: 0.12031763985753059, Test loss: 0.13594545423984528\n", + "Starting epoch 1436\n", + "Train loss: 0.12093843728303909, Test loss: 0.1426297426223755\n", + "Starting epoch 1437\n", + "Train loss: 0.12042699560523033, Test loss: 0.1362791508436203\n", + "Starting epoch 1438\n", + "Train loss: 0.1209727792441845, Test loss: 0.13593947887420654\n", + "Starting epoch 1439\n", + "Train loss: 0.12022392764687538, Test loss: 0.13865169882774353\n", + "Starting epoch 1440\n", + "Train loss: 0.12017286509275436, Test loss: 0.1369953602552414\n", + "Starting epoch 1441\n", + "Train loss: 0.12036002486944199, Test loss: 0.13703908026218414\n", + "Starting epoch 1442\n", + "Train loss: 0.1203779438138008, Test loss: 0.13590285181999207\n", + "Starting epoch 1443\n", + "Train loss: 0.12029570445418358, Test loss: 0.1353261023759842\n", + "Starting epoch 1444\n", + "Train loss: 0.12035134345293046, Test loss: 0.1351718008518219\n", + "Starting epoch 1445\n", + "Train loss: 0.120169767588377, Test loss: 0.13967594504356384\n", + "Starting epoch 1446\n", + "Train loss: 0.12038652434945106, Test loss: 0.13598449528217316\n", + "Starting epoch 1447\n", + "Train loss: 0.1206580550968647, Test loss: 0.13683074712753296\n", + "Starting epoch 1448\n", + "Train loss: 0.12107133015990257, Test loss: 0.13666971027851105\n", + "Starting epoch 1449\n", + "Train loss: 0.12001928091049194, Test loss: 0.13575942814350128\n", + "Starting epoch 1450\n", + "Train loss: 0.12044982850551605, Test loss: 0.13743740320205688\n", + "Starting epoch 1451\n", + "Train loss: 0.12074665442109107, Test loss: 0.13614237308502197\n", + "Starting epoch 1452\n", + "Train loss: 0.12071496248245239, Test loss: 0.145552396774292\n", + "Starting epoch 1453\n", + "Train loss: 0.12046051114797592, Test loss: 0.13573478162288666\n", + "Starting epoch 1454\n", + "Train loss: 0.12008674591779708, Test loss: 0.1388210654258728\n", + "Starting epoch 1455\n", + "Train loss: 0.12041782990097999, Test loss: 0.1366937756538391\n", + "Starting epoch 1456\n", + "Train loss: 0.12037062644958496, Test loss: 0.1350473314523697\n", + "Starting epoch 1457\n", + "Train loss: 0.12041976794600487, Test loss: 0.1416923999786377\n", + "Starting epoch 1458\n", + "Train loss: 0.12026369363069535, Test loss: 0.13660457730293274\n", + "Starting epoch 1459\n", + "Train loss: 0.12047918066382408, Test loss: 0.1418585181236267\n", + "Starting epoch 1460\n", + "Train loss: 0.12052879333496094, Test loss: 0.13879449665546417\n", + "Starting epoch 1461\n", + "Train loss: 0.12078572049736977, Test loss: 0.13820257782936096\n", + "Starting epoch 1462\n", + "Train loss: 0.12023800447583198, Test loss: 0.1361946165561676\n", + "Starting epoch 1463\n", + "Train loss: 0.12064702868461609, Test loss: 0.13504090905189514\n", + "Starting epoch 1464\n", + "Train loss: 0.12034445062279701, Test loss: 0.1368325650691986\n", + "Starting epoch 1465\n", + "Train loss: 0.12028238609433174, Test loss: 0.14114971458911896\n", + "Starting epoch 1466\n", + "Train loss: 0.12032667830586434, Test loss: 0.13484205305576324\n", + "Starting epoch 1467\n", + "Train loss: 0.12008261457085609, Test loss: 0.13507261872291565\n", + "Starting epoch 1468\n", + "Train loss: 0.12053996741771698, Test loss: 0.14187385141849518\n", + "Starting epoch 1469\n", + "Train loss: 0.12021910876035691, Test loss: 0.13622935116291046\n", + "Starting epoch 1470\n", + "Train loss: 0.12063249945640564, Test loss: 0.13706263899803162\n", + "Starting epoch 1471\n", + "Train loss: 0.12072112753987313, Test loss: 0.1365569531917572\n", + "Starting epoch 1472\n", + "Train loss: 0.12057173848152161, Test loss: 0.13740384578704834\n", + "Starting epoch 1473\n", + "Train loss: 0.12081433162093162, Test loss: 0.1360616832971573\n", + "Starting epoch 1474\n", + "Train loss: 0.12048050671815873, Test loss: 0.13550110161304474\n", + "Starting epoch 1475\n", + "Train loss: 0.1206814992427826, Test loss: 0.1353096067905426\n", + "Starting epoch 1476\n", + "Train loss: 0.12061751231551171, Test loss: 0.1376972198486328\n", + "Starting epoch 1477\n", + "Train loss: 0.120417008548975, Test loss: 0.13645875453948975\n", + "Starting epoch 1478\n", + "Train loss: 0.12049929335713387, Test loss: 0.13552020490169525\n", + "Starting epoch 1479\n", + "Train loss: 0.12013418599963188, Test loss: 0.13520066440105438\n", + "Starting epoch 1480\n", + "Train loss: 0.12071192979812623, Test loss: 0.13792793452739716\n", + "Starting epoch 1481\n", + "Train loss: 0.1199623018503189, Test loss: 0.13470566272735596\n", + "Starting epoch 1482\n", + "Train loss: 0.12038958206772804, Test loss: 0.13637089729309082\n", + "Starting epoch 1483\n", + "Train loss: 0.12004498541355133, Test loss: 0.13554568588733673\n", + "Starting epoch 1484\n", + "Train loss: 0.12013784438371658, Test loss: 0.136312335729599\n", + "Starting epoch 1485\n", + "Train loss: 0.1199906124174595, Test loss: 0.13518674671649933\n", + "Starting epoch 1486\n", + "Train loss: 0.12063758611679078, Test loss: 0.13620254397392273\n", + "Starting epoch 1487\n", + "Train loss: 0.12028070703148842, Test loss: 0.13624754548072815\n", + "Starting epoch 1488\n", + "Train loss: 0.12014840349555016, Test loss: 0.1389593482017517\n", + "Starting epoch 1489\n", + "Train loss: 0.12062398701906205, Test loss: 0.13618947565555573\n", + "Starting epoch 1490\n", + "Train loss: 0.12048663526773452, Test loss: 0.1415567249059677\n", + "Starting epoch 1491\n", + "Train loss: 0.12032791584730149, Test loss: 0.13551867008209229\n", + "Starting epoch 1492\n", + "Train loss: 0.12009481117129325, Test loss: 0.13529770076274872\n", + "Starting epoch 1493\n", + "Train loss: 0.12018832758069038, Test loss: 0.13847845792770386\n", + "Starting epoch 1494\n", + "Train loss: 0.12039058461785317, Test loss: 0.13716016709804535\n", + "Starting epoch 1495\n", + "Train loss: 0.12012693479657173, Test loss: 0.1354237198829651\n", + "Starting epoch 1496\n", + "Train loss: 0.12041818886995316, Test loss: 0.14073245227336884\n", + "Starting epoch 1497\n", + "Train loss: 0.12026149436831474, Test loss: 0.13532528281211853\n", + "Starting epoch 1498\n", + "Train loss: 0.12062097519636154, Test loss: 0.13539235293865204\n", + "Starting epoch 1499\n", + "Train loss: 0.12014803290367126, Test loss: 0.13605085015296936\n", + "Starting epoch 1500\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.12046393111348153, Test loss: 0.13596206903457642\n", + "Starting epoch 1501\n", + "Train loss: 0.12023654758930206, Test loss: 0.1360253393650055\n", + "Starting epoch 1502\n", + "Train loss: 0.12027890801429748, Test loss: 0.13858211040496826\n", + "Starting epoch 1503\n", + "Train loss: 0.12007680892944336, Test loss: 0.13505737483501434\n", + "Starting epoch 1504\n", + "Train loss: 0.12027729123830795, Test loss: 0.13761276006698608\n", + "Starting epoch 1505\n", + "Train loss: 0.12035753086209297, Test loss: 0.1358615905046463\n", + "Starting epoch 1506\n", + "Train loss: 0.1202017168700695, Test loss: 0.13578064739704132\n", + "Starting epoch 1507\n", + "Train loss: 0.12056099340319633, Test loss: 0.14143836498260498\n", + "Starting epoch 1508\n", + "Train loss: 0.1199941298365593, Test loss: 0.13584324717521667\n", + "Starting epoch 1509\n", + "Train loss: 0.12064793094992637, Test loss: 0.1372535675764084\n", + "Starting epoch 1510\n", + "Train loss: 0.12004356071352959, Test loss: 0.13602878153324127\n", + "Starting epoch 1511\n", + "Train loss: 0.12059300199151039, Test loss: 0.13786904513835907\n", + "Starting epoch 1512\n", + "Train loss: 0.12029457584023476, Test loss: 0.13890790939331055\n", + "Starting epoch 1513\n", + "Train loss: 0.12022836461663246, Test loss: 0.1357015073299408\n", + "Starting epoch 1514\n", + "Train loss: 0.12063898265361786, Test loss: 0.13580238819122314\n", + "Starting epoch 1515\n", + "Train loss: 0.12035865768790245, Test loss: 0.1383889615535736\n", + "Starting epoch 1516\n", + "Train loss: 0.12019498825073242, Test loss: 0.13837569952011108\n", + "Starting epoch 1517\n", + "Train loss: 0.12047159194946289, Test loss: 0.13709746301174164\n", + "Starting epoch 1518\n", + "Train loss: 0.12017828717827797, Test loss: 0.13528047502040863\n", + "Starting epoch 1519\n", + "Train loss: 0.12025870308279991, Test loss: 0.1350056529045105\n", + "Starting epoch 1520\n", + "Train loss: 0.12001861527562141, Test loss: 0.1410551369190216\n", + "Starting epoch 1521\n", + "Train loss: 0.12024434372782707, Test loss: 0.13568072021007538\n", + "Starting epoch 1522\n", + "Train loss: 0.12000417575240135, Test loss: 0.13526153564453125\n", + "Starting epoch 1523\n", + "Train loss: 0.12041775599122047, Test loss: 0.1374385803937912\n", + "Starting epoch 1524\n", + "Train loss: 0.12037122309207916, Test loss: 0.13730908930301666\n", + "Starting epoch 1525\n", + "Train loss: 0.12008930176496506, Test loss: 0.13554789125919342\n", + "Starting epoch 1526\n", + "Train loss: 0.12023232713341712, Test loss: 0.13989047706127167\n", + "Starting epoch 1527\n", + "Train loss: 0.12021319687366486, Test loss: 0.13513559103012085\n", + "Starting epoch 1528\n", + "Train loss: 0.12057169571518898, Test loss: 0.13643842935562134\n", + "Starting epoch 1529\n", + "Train loss: 0.12009455054998398, Test loss: 0.13561494648456573\n", + "Starting epoch 1530\n", + "Train loss: 0.12028465807437896, Test loss: 0.1405559927225113\n", + "Starting epoch 1531\n", + "Train loss: 0.1204848437011242, Test loss: 0.13775911927223206\n", + "Starting epoch 1532\n", + "Train loss: 0.12001579597592354, Test loss: 0.13620701432228088\n", + "Starting epoch 1533\n", + "Train loss: 0.12039339765906334, Test loss: 0.1368979513645172\n", + "Starting epoch 1534\n", + "Train loss: 0.12054193496704102, Test loss: 0.13468238711357117\n", + "Starting epoch 1535\n", + "Train loss: 0.11998557090759278, Test loss: 0.13566476106643677\n", + "Starting epoch 1536\n", + "Train loss: 0.12077489867806435, Test loss: 0.13656018674373627\n", + "Starting epoch 1537\n", + "Train loss: 0.12008593663573265, Test loss: 0.13652101159095764\n", + "Starting epoch 1538\n", + "Train loss: 0.12050897434353829, Test loss: 0.1431400626897812\n", + "Starting epoch 1539\n", + "Train loss: 0.12031797870993614, Test loss: 0.13538405299186707\n", + "Starting epoch 1540\n", + "Train loss: 0.12046491384506225, Test loss: 0.13721175491809845\n", + "Starting epoch 1541\n", + "Train loss: 0.11993826165795327, Test loss: 0.13594889640808105\n", + "Starting epoch 1542\n", + "Train loss: 0.1207861328125, Test loss: 0.13719283044338226\n", + "Starting epoch 1543\n", + "Train loss: 0.1202872234582901, Test loss: 0.13730251789093018\n", + "Starting epoch 1544\n", + "Train loss: 0.12005777955055237, Test loss: 0.13637053966522217\n", + "Starting epoch 1545\n", + "Train loss: 0.11998818188905716, Test loss: 0.13522963225841522\n", + "Starting epoch 1546\n", + "Train loss: 0.11997850373387337, Test loss: 0.13457827270030975\n", + "Starting epoch 1547\n", + "Train loss: 0.12027979642152786, Test loss: 0.1383780837059021\n", + "Starting epoch 1548\n", + "Train loss: 0.1204130682349205, Test loss: 0.13995909690856934\n", + "Starting epoch 1549\n", + "Train loss: 0.12014610677957535, Test loss: 0.1365131139755249\n", + "Starting epoch 1550\n", + "Train loss: 0.11992121070623397, Test loss: 0.13548295199871063\n", + "Starting epoch 1551\n", + "Train loss: 0.1200032602250576, Test loss: 0.13432000577449799\n", + "Starting epoch 1552\n", + "Train loss: 0.11976678550243378, Test loss: 0.1341603547334671\n", + "Starting epoch 1553\n", + "Train loss: 0.11985046967864037, Test loss: 0.13484422862529755\n", + "Starting epoch 1554\n", + "Train loss: 0.12023375898599625, Test loss: 0.13375431299209595\n", + "Starting epoch 1555\n", + "Train loss: 0.1197493588924408, Test loss: 0.13301818072795868\n", + "Starting epoch 1556\n", + "Train loss: 0.11910739943385124, Test loss: 0.13568268716335297\n", + "Starting epoch 1557\n", + "Train loss: 0.11838721200823783, Test loss: 0.14259073138237\n", + "Starting epoch 1558\n", + "Train loss: 0.11880164206027985, Test loss: 0.146325945854187\n", + "Starting epoch 1559\n", + "Train loss: 0.11784987971186638, Test loss: 0.14022882282733917\n", + "Starting epoch 1560\n", + "Train loss: 0.11810209915041923, Test loss: 0.16377490758895874\n", + "Starting epoch 1561\n", + "Train loss: 0.11730685576796532, Test loss: 0.15158608555793762\n", + "Starting epoch 1562\n", + "Train loss: 0.11692250609397888, Test loss: 0.1541978120803833\n", + "Starting epoch 1563\n", + "Train loss: 0.11726491004228592, Test loss: 0.15613152086734772\n", + "Starting epoch 1564\n", + "Train loss: 0.11720127314329147, Test loss: 0.150183767080307\n", + "Starting epoch 1565\n", + "Train loss: 0.11738607972860336, Test loss: 0.15410950779914856\n", + "Starting epoch 1566\n", + "Train loss: 0.1169783467054367, Test loss: 0.1573193371295929\n", + "Starting epoch 1567\n", + "Train loss: 0.11704792305827141, Test loss: 0.1541769802570343\n", + "Starting epoch 1568\n", + "Train loss: 0.11713172331452369, Test loss: 0.15420417487621307\n", + "Starting epoch 1569\n", + "Train loss: 0.11685422956943511, Test loss: 0.1623350828886032\n", + "Starting epoch 1570\n", + "Train loss: 0.11680835083127022, Test loss: 0.15719640254974365\n", + "Starting epoch 1571\n", + "Train loss: 0.1166985210776329, Test loss: 0.1518874317407608\n", + "Starting epoch 1572\n", + "Train loss: 0.11695843160152436, Test loss: 0.15411657094955444\n", + "Starting epoch 1573\n", + "Train loss: 0.1168484142422676, Test loss: 0.16350410878658295\n", + "Starting epoch 1574\n", + "Train loss: 0.11667484730482101, Test loss: 0.14814284443855286\n", + "Starting epoch 1575\n", + "Train loss: 0.11690065532922744, Test loss: 0.1578773409128189\n", + "Starting epoch 1576\n", + "Train loss: 0.11677723228931428, Test loss: 0.14905965328216553\n", + "Starting epoch 1577\n", + "Train loss: 0.11689622700214386, Test loss: 0.15463927388191223\n", + "Starting epoch 1578\n", + "Train loss: 0.1166431587934494, Test loss: 0.15275610983371735\n", + "Starting epoch 1579\n", + "Train loss: 0.11699127018451691, Test loss: 0.15020602941513062\n", + "Starting epoch 1580\n", + "Train loss: 0.11704445198178291, Test loss: 0.14678961038589478\n", + "Starting epoch 1581\n", + "Train loss: 0.11665154948830604, Test loss: 0.14869928359985352\n", + "Starting epoch 1582\n", + "Train loss: 0.11658100321888924, Test loss: 0.14955449104309082\n", + "Starting epoch 1583\n", + "Train loss: 0.11657865941524506, Test loss: 0.15537108480930328\n", + "Starting epoch 1584\n", + "Train loss: 0.1168563337624073, Test loss: 0.14711275696754456\n", + "Starting epoch 1585\n", + "Train loss: 0.1167568975687027, Test loss: 0.15845605731010437\n", + "Starting epoch 1586\n", + "Train loss: 0.1166544097661972, Test loss: 0.15708507597446442\n", + "Starting epoch 1587\n", + "Train loss: 0.11734849914908409, Test loss: 0.15675902366638184\n", + "Starting epoch 1588\n", + "Train loss: 0.11685158684849739, Test loss: 0.15405486524105072\n", + "Starting epoch 1589\n", + "Train loss: 0.11665386378765107, Test loss: 0.15613554418087006\n", + "Starting epoch 1590\n", + "Train loss: 0.11650794833898544, Test loss: 0.15790115296840668\n", + "Starting epoch 1591\n", + "Train loss: 0.11665393441915511, Test loss: 0.14862807095050812\n", + "Starting epoch 1592\n", + "Train loss: 0.11649618059396744, Test loss: 0.1526058167219162\n", + "Starting epoch 1593\n", + "Train loss: 0.11665326729416847, Test loss: 0.14594511687755585\n", + "Starting epoch 1594\n", + "Train loss: 0.11676721870899201, Test loss: 0.16051632165908813\n", + "Starting epoch 1595\n", + "Train loss: 0.11689959421753883, Test loss: 0.1510775238275528\n", + "Starting epoch 1596\n", + "Train loss: 0.11659931316971779, Test loss: 0.1585330218076706\n", + "Starting epoch 1597\n", + "Train loss: 0.1166253174841404, Test loss: 0.15460678935050964\n", + "Starting epoch 1598\n", + "Train loss: 0.11656560510396957, Test loss: 0.1473802626132965\n", + "Starting epoch 1599\n", + "Train loss: 0.11640792638063431, Test loss: 0.15546831488609314\n", + "Starting epoch 1600\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.11674309477210044, Test loss: 0.15212281048297882\n", + "Starting epoch 1601\n", + "Train loss: 0.1165670931339264, Test loss: 0.14270417392253876\n", + "Starting epoch 1602\n", + "Train loss: 0.11693968519568443, Test loss: 0.15023986995220184\n", + "Starting epoch 1603\n", + "Train loss: 0.11646769002079964, Test loss: 0.1502395123243332\n", + "Starting epoch 1604\n", + "Train loss: 0.11636195182800294, Test loss: 0.1514829695224762\n", + "Starting epoch 1605\n", + "Train loss: 0.11626966282725335, Test loss: 0.14433665573596954\n", + "Starting epoch 1606\n", + "Train loss: 0.11648281410336495, Test loss: 0.1475703865289688\n", + "Starting epoch 1607\n", + "Train loss: 0.11666763722896575, Test loss: 0.14554548263549805\n", + "Starting epoch 1608\n", + "Train loss: 0.11615578532218933, Test loss: 0.14331506192684174\n", + "Starting epoch 1609\n", + "Train loss: 0.11646099627017975, Test loss: 0.14422166347503662\n", + "Starting epoch 1610\n", + "Train loss: 0.11618852362036705, Test loss: 0.15454961359500885\n", + "Starting epoch 1611\n", + "Train loss: 0.11600187450647353, Test loss: 0.14868642389774323\n", + "Starting epoch 1612\n", + "Train loss: 0.11664525210857392, Test loss: 0.1425919532775879\n", + "Starting epoch 1613\n", + "Train loss: 0.11557554692029953, Test loss: 0.1463070809841156\n", + "Starting epoch 1614\n", + "Train loss: 0.1150375933945179, Test loss: 0.15503354370594025\n", + "Starting epoch 1615\n", + "Train loss: 0.11447558924555779, Test loss: 0.13739833235740662\n", + "Starting epoch 1616\n", + "Train loss: 0.113645761013031, Test loss: 0.1342042237520218\n", + "Starting epoch 1617\n", + "Train loss: 0.11407086193561554, Test loss: 0.13393184542655945\n", + "Starting epoch 1618\n", + "Train loss: 0.11317916363477706, Test loss: 0.1383870542049408\n", + "Starting epoch 1619\n", + "Train loss: 0.11274078026413918, Test loss: 0.1406925618648529\n", + "Starting epoch 1620\n", + "Train loss: 0.11241484850645066, Test loss: 0.14092442393302917\n", + "Starting epoch 1621\n", + "Train loss: 0.1125134852528572, Test loss: 0.13694100081920624\n", + "Starting epoch 1622\n", + "Train loss: 0.11261175557971001, Test loss: 0.13882842659950256\n", + "Starting epoch 1623\n", + "Train loss: 0.11211903184652329, Test loss: 0.1364034116268158\n", + "Starting epoch 1624\n", + "Train loss: 0.11225719287991524, Test loss: 0.14261923730373383\n", + "Starting epoch 1625\n", + "Train loss: 0.1119482558965683, Test loss: 0.14697672426700592\n", + "Starting epoch 1626\n", + "Train loss: 0.11273556917905808, Test loss: 0.14095698297023773\n", + "Starting epoch 1627\n", + "Train loss: 0.11197294041514397, Test loss: 0.13507097959518433\n", + "Starting epoch 1628\n", + "Train loss: 0.1122355504333973, Test loss: 0.147775337100029\n", + "Starting epoch 1629\n", + "Train loss: 0.11286209776997566, Test loss: 0.14622782170772552\n", + "Starting epoch 1630\n", + "Train loss: 0.11235419616103172, Test loss: 0.13270169496536255\n", + "Starting epoch 1631\n", + "Train loss: 0.11194438248872757, Test loss: 0.13647723197937012\n", + "Starting epoch 1632\n", + "Train loss: 0.11180065870285034, Test loss: 0.1407596617937088\n", + "Starting epoch 1633\n", + "Train loss: 0.11189653888344765, Test loss: 0.13670215010643005\n", + "Starting epoch 1634\n", + "Train loss: 0.11196846842765808, Test loss: 0.14368703961372375\n", + "Starting epoch 1635\n", + "Train loss: 0.11185796275734901, Test loss: 0.1341654658317566\n", + "Starting epoch 1636\n", + "Train loss: 0.11162891179323196, Test loss: 0.139460951089859\n", + "Starting epoch 1637\n", + "Train loss: 0.11223962590098381, Test loss: 0.14169053733348846\n", + "Starting epoch 1638\n", + "Train loss: 0.11154171273112297, Test loss: 0.13726158440113068\n", + "Starting epoch 1639\n", + "Train loss: 0.1117304089665413, Test loss: 0.1413484662771225\n", + "Starting epoch 1640\n", + "Train loss: 0.11168639793992043, Test loss: 0.1343436986207962\n", + "Starting epoch 1641\n", + "Train loss: 0.1114966145157814, Test loss: 0.1403805911540985\n", + "Starting epoch 1642\n", + "Train loss: 0.11187555611133576, Test loss: 0.14086446166038513\n", + "Starting epoch 1643\n", + "Train loss: 0.11233493059873581, Test loss: 0.14471103250980377\n", + "Starting epoch 1644\n", + "Train loss: 0.11170223832130433, Test loss: 0.13952113687992096\n", + "Starting epoch 1645\n", + "Train loss: 0.11129550367593766, Test loss: 0.13453078269958496\n", + "Starting epoch 1646\n", + "Train loss: 0.11119901731610299, Test loss: 0.131743922829628\n", + "Starting epoch 1647\n", + "Train loss: 0.11118376314640045, Test loss: 0.1305808424949646\n", + "Starting epoch 1648\n", + "Train loss: 0.11139758229255677, Test loss: 0.14363031089305878\n", + "Starting epoch 1649\n", + "Train loss: 0.11146572306752205, Test loss: 0.12994986772537231\n", + "Starting epoch 1650\n", + "Train loss: 0.11136253640055656, Test loss: 0.12896475195884705\n", + "Starting epoch 1651\n", + "Train loss: 0.111794101446867, Test loss: 0.13902495801448822\n", + "Starting epoch 1652\n", + "Train loss: 0.11155385285615921, Test loss: 0.12996569275856018\n", + "Starting epoch 1653\n", + "Train loss: 0.11121192991733551, Test loss: 0.1328871101140976\n", + "Starting epoch 1654\n", + "Train loss: 0.11114880964159965, Test loss: 0.12716323137283325\n", + "Starting epoch 1655\n", + "Train loss: 0.1112974514067173, Test loss: 0.12758192420005798\n", + "Starting epoch 1656\n", + "Train loss: 0.11093576937913895, Test loss: 0.12925536930561066\n", + "Starting epoch 1657\n", + "Train loss: 0.1112850821018219, Test loss: 0.12534038722515106\n", + "Starting epoch 1658\n", + "Train loss: 0.11095298826694489, Test loss: 0.13213345408439636\n", + "Starting epoch 1659\n", + "Train loss: 0.11097698017954827, Test loss: 0.12618346512317657\n", + "Starting epoch 1660\n", + "Train loss: 0.11117166504263878, Test loss: 0.13852913677692413\n", + "Starting epoch 1661\n", + "Train loss: 0.1110063786804676, Test loss: 0.13202489912509918\n", + "Starting epoch 1662\n", + "Train loss: 0.11088428735733032, Test loss: 0.12475771456956863\n", + "Starting epoch 1663\n", + "Train loss: 0.11100034281611443, Test loss: 0.13011787831783295\n", + "Starting epoch 1664\n", + "Train loss: 0.11100760251283645, Test loss: 0.1268354207277298\n", + "Starting epoch 1665\n", + "Train loss: 0.1107462702691555, Test loss: 0.13187189400196075\n", + "Starting epoch 1666\n", + "Train loss: 0.11059963434934617, Test loss: 0.12888720631599426\n", + "Starting epoch 1667\n", + "Train loss: 0.11105762541294098, Test loss: 0.1329336166381836\n", + "Starting epoch 1668\n", + "Train loss: 0.11094450250267983, Test loss: 0.12369589507579803\n", + "Starting epoch 1669\n", + "Train loss: 0.11093153432011604, Test loss: 0.13425403833389282\n", + "Starting epoch 1670\n", + "Train loss: 0.1104358896613121, Test loss: 0.13429602980613708\n", + "Starting epoch 1671\n", + "Train loss: 0.11079016759991646, Test loss: 0.13457715511322021\n", + "Starting epoch 1672\n", + "Train loss: 0.11118928045034408, Test loss: 0.1295943260192871\n", + "Starting epoch 1673\n", + "Train loss: 0.1106146889925003, Test loss: 0.13119325041770935\n", + "Starting epoch 1674\n", + "Train loss: 0.11099382549524307, Test loss: 0.12361655384302139\n", + "Starting epoch 1675\n", + "Train loss: 0.11069442242383958, Test loss: 0.12320949137210846\n", + "Starting epoch 1676\n", + "Train loss: 0.11075139001011848, Test loss: 0.1443464607000351\n", + "Starting epoch 1677\n", + "Train loss: 0.11049602434039116, Test loss: 0.13275213539600372\n", + "Starting epoch 1678\n", + "Train loss: 0.11119676381349564, Test loss: 0.1374763548374176\n", + "Starting epoch 1679\n", + "Train loss: 0.1104669938981533, Test loss: 0.13412775099277496\n", + "Starting epoch 1680\n", + "Train loss: 0.11036298483610153, Test loss: 0.13704116642475128\n", + "Starting epoch 1681\n", + "Train loss: 0.11017289116978646, Test loss: 0.12661850452423096\n", + "Starting epoch 1682\n", + "Train loss: 0.11000337392091751, Test loss: 0.12867115437984467\n", + "Starting epoch 1683\n", + "Train loss: 0.11033852532505989, Test loss: 0.13226747512817383\n", + "Starting epoch 1684\n", + "Train loss: 0.10964749217033386, Test loss: 0.12632551789283752\n", + "Starting epoch 1685\n", + "Train loss: 0.10957735732197761, Test loss: 0.1276141107082367\n", + "Starting epoch 1686\n", + "Train loss: 0.10923940405249595, Test loss: 0.1356353759765625\n", + "Starting epoch 1687\n", + "Train loss: 0.10794686019420624, Test loss: 0.12710624933242798\n", + "Starting epoch 1688\n", + "Train loss: 0.10533381178975106, Test loss: 0.12712611258029938\n", + "Starting epoch 1689\n", + "Train loss: 0.09959293261170388, Test loss: 0.17020642757415771\n", + "Starting epoch 1690\n", + "Train loss: 0.09267987117171288, Test loss: 0.13566653430461884\n", + "Starting epoch 1691\n", + "Train loss: 0.08721114039421081, Test loss: 0.16118057072162628\n", + "Starting epoch 1692\n", + "Train loss: 0.0845362351834774, Test loss: 0.12498892098665237\n", + "Starting epoch 1693\n", + "Train loss: 0.08333651408553124, Test loss: 0.13583795726299286\n", + "Starting epoch 1694\n", + "Train loss: 0.08159226968884468, Test loss: 0.12848451733589172\n", + "Starting epoch 1695\n", + "Train loss: 0.0809943874180317, Test loss: 0.12898722290992737\n", + "Starting epoch 1696\n", + "Train loss: 0.08032398626208305, Test loss: 0.111542709171772\n", + "Starting epoch 1697\n", + "Train loss: 0.08026940837502479, Test loss: 0.11182851344347\n", + "Starting epoch 1698\n", + "Train loss: 0.08030900776386261, Test loss: 0.11578219383955002\n", + "Starting epoch 1699\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.07893518656492234, Test loss: 0.11581934988498688\n", + "Starting epoch 1700\n", + "Train loss: 0.07855947315692902, Test loss: 0.11517677456140518\n", + "Starting epoch 1701\n", + "Train loss: 0.07803865566849709, Test loss: 0.11105945706367493\n", + "Starting epoch 1702\n", + "Train loss: 0.07865985780954361, Test loss: 0.13534507155418396\n", + "Starting epoch 1703\n", + "Train loss: 0.07791358292102814, Test loss: 0.11712225526571274\n", + "Starting epoch 1704\n", + "Train loss: 0.07737534791231156, Test loss: 0.1281667798757553\n", + "Starting epoch 1705\n", + "Train loss: 0.07790255472064019, Test loss: 0.10829020291566849\n", + "Starting epoch 1706\n", + "Train loss: 0.07782225131988525, Test loss: 0.1255517303943634\n", + "Starting epoch 1707\n", + "Train loss: 0.07713214471936226, Test loss: 0.12053300440311432\n", + "Starting epoch 1708\n", + "Train loss: 0.07728551417589187, Test loss: 0.11048202216625214\n", + "Starting epoch 1709\n", + "Train loss: 0.07695078998804092, Test loss: 0.1114087626338005\n", + "Starting epoch 1710\n", + "Train loss: 0.07663584813475609, Test loss: 0.12073423713445663\n", + "Starting epoch 1711\n", + "Train loss: 0.07753059566020966, Test loss: 0.11844781041145325\n", + "Starting epoch 1712\n", + "Train loss: 0.07695677116513253, Test loss: 0.10808390378952026\n", + "Starting epoch 1713\n", + "Train loss: 0.07627080872654915, Test loss: 0.10953063517808914\n", + "Starting epoch 1714\n", + "Train loss: 0.07709303259849548, Test loss: 0.10276758670806885\n", + "Starting epoch 1715\n", + "Train loss: 0.07704515129327774, Test loss: 0.10211518406867981\n", + "Starting epoch 1716\n", + "Train loss: 0.07646834701299668, Test loss: 0.1064673513174057\n", + "Starting epoch 1717\n", + "Train loss: 0.07602152302861213, Test loss: 0.1064675971865654\n", + "Starting epoch 1718\n", + "Train loss: 0.07622570514678956, Test loss: 0.11773980408906937\n", + "Starting epoch 1719\n", + "Train loss: 0.07639257252216339, Test loss: 0.10273859649896622\n", + "Starting epoch 1720\n", + "Train loss: 0.07582869753241539, Test loss: 0.1190701350569725\n", + "Starting epoch 1721\n", + "Train loss: 0.07617711931467057, Test loss: 0.11609575152397156\n", + "Starting epoch 1722\n", + "Train loss: 0.0756478214263916, Test loss: 0.11299698799848557\n", + "Starting epoch 1723\n", + "Train loss: 0.07629902690649032, Test loss: 0.10489962249994278\n", + "Starting epoch 1724\n", + "Train loss: 0.07679247617721557, Test loss: 0.10085546225309372\n", + "Starting epoch 1725\n", + "Train loss: 0.07519360736012459, Test loss: 0.10661066323518753\n", + "Starting epoch 1726\n", + "Train loss: 0.07509175136685371, Test loss: 0.11535178124904633\n", + "Starting epoch 1727\n", + "Train loss: 0.07510182216763496, Test loss: 0.11356548219919205\n", + "Starting epoch 1728\n", + "Train loss: 0.07408454656600952, Test loss: 0.10578680038452148\n", + "Starting epoch 1729\n", + "Train loss: 0.07317094296216965, Test loss: 0.09776926785707474\n", + "Starting epoch 1730\n", + "Train loss: 0.07295583426952362, Test loss: 0.11303385347127914\n", + "Starting epoch 1731\n", + "Train loss: 0.07025509476661682, Test loss: 0.1056397333741188\n", + "Starting epoch 1732\n", + "Train loss: 0.06866979569196702, Test loss: 0.09064941108226776\n", + "Starting epoch 1733\n", + "Train loss: 0.06732282072305679, Test loss: 0.09051742404699326\n", + "Starting epoch 1734\n", + "Train loss: 0.06546893447637558, Test loss: 0.08483171463012695\n", + "Starting epoch 1735\n", + "Train loss: 0.0649481648951769, Test loss: 0.10618491470813751\n", + "Starting epoch 1736\n", + "Train loss: 0.06349081113934517, Test loss: 0.10029003769159317\n", + "Starting epoch 1737\n", + "Train loss: 0.0626018337160349, Test loss: 0.09565829485654831\n", + "Starting epoch 1738\n", + "Train loss: 0.06295211613178253, Test loss: 0.1027940958738327\n", + "Starting epoch 1739\n", + "Train loss: 0.06072384215891361, Test loss: 0.07991862297058105\n", + "Starting epoch 1740\n", + "Train loss: 0.06061082027852535, Test loss: 0.08393567055463791\n", + "Starting epoch 1741\n", + "Train loss: 0.06006407357752323, Test loss: 0.07922448962926865\n", + "Starting epoch 1742\n", + "Train loss: 0.05977490775287151, Test loss: 0.09118757396936417\n", + "Starting epoch 1743\n", + "Train loss: 0.058875145092606544, Test loss: 0.08641355484724045\n", + "Starting epoch 1744\n", + "Train loss: 0.05861762069165707, Test loss: 0.08316630125045776\n", + "Starting epoch 1745\n", + "Train loss: 0.058375963866710665, Test loss: 0.0967693030834198\n", + "Starting epoch 1746\n", + "Train loss: 0.05812728442251682, Test loss: 0.07723825424909592\n", + "Starting epoch 1747\n", + "Train loss: 0.05753509044647217, Test loss: 0.09118131548166275\n", + "Starting epoch 1748\n", + "Train loss: 0.05751654252409935, Test loss: 0.095088891685009\n", + "Starting epoch 1749\n", + "Train loss: 0.05591733880341053, Test loss: 0.09519242495298386\n", + "Starting epoch 1750\n", + "Train loss: 0.05576207838952541, Test loss: 0.08875490725040436\n", + "Starting epoch 1751\n", + "Train loss: 0.05493428848683834, Test loss: 0.08092939853668213\n", + "Starting epoch 1752\n", + "Train loss: 0.05441487014293671, Test loss: 0.08096548169851303\n", + "Starting epoch 1753\n", + "Train loss: 0.054537034034729, Test loss: 0.08940810710191727\n", + "Starting epoch 1754\n", + "Train loss: 0.05364564768970013, Test loss: 0.08556384593248367\n", + "Starting epoch 1755\n", + "Train loss: 0.054349315389990806, Test loss: 0.07837984710931778\n", + "Starting epoch 1756\n", + "Train loss: 0.05319175437092781, Test loss: 0.07507570087909698\n", + "Starting epoch 1757\n", + "Train loss: 0.05327784851193428, Test loss: 0.08208952099084854\n", + "Starting epoch 1758\n", + "Train loss: 0.05370045378804207, Test loss: 0.07294155657291412\n", + "Starting epoch 1759\n", + "Train loss: 0.05296619974076748, Test loss: 0.0889030322432518\n", + "Starting epoch 1760\n", + "Train loss: 0.05303824111819267, Test loss: 0.09066814184188843\n", + "Starting epoch 1761\n", + "Train loss: 0.05283397540450096, Test loss: 0.07797323912382126\n", + "Starting epoch 1762\n", + "Train loss: 0.05280796095728874, Test loss: 0.08242589980363846\n", + "Starting epoch 1763\n", + "Train loss: 0.052472768872976305, Test loss: 0.07878001034259796\n", + "Starting epoch 1764\n", + "Train loss: 0.05231437437236309, Test loss: 0.09469433128833771\n", + "Starting epoch 1765\n", + "Train loss: 0.05216370679438114, Test loss: 0.07417436689138412\n", + "Starting epoch 1766\n", + "Train loss: 0.05268285445868969, Test loss: 0.08480064570903778\n", + "Starting epoch 1767\n", + "Train loss: 0.05198727749288082, Test loss: 0.07294025272130966\n", + "Starting epoch 1768\n", + "Train loss: 0.0518391777575016, Test loss: 0.09219758957624435\n", + "Starting epoch 1769\n", + "Train loss: 0.05196948669850826, Test loss: 0.09061920642852783\n", + "Starting epoch 1770\n", + "Train loss: 0.05159660808742046, Test loss: 0.08893212676048279\n", + "Starting epoch 1771\n", + "Train loss: 0.05172562807798386, Test loss: 0.08111871033906937\n", + "Starting epoch 1772\n", + "Train loss: 0.052094902098178866, Test loss: 0.07535938918590546\n", + "Starting epoch 1773\n", + "Train loss: 0.05187216706573963, Test loss: 0.07234576344490051\n", + "Starting epoch 1774\n", + "Train loss: 0.05128732845187187, Test loss: 0.07453405857086182\n", + "Starting epoch 1775\n", + "Train loss: 0.05140480123460293, Test loss: 0.07235682755708694\n", + "Starting epoch 1776\n", + "Train loss: 0.05138490565121174, Test loss: 0.07134301960468292\n", + "Starting epoch 1777\n", + "Train loss: 0.051357572823762895, Test loss: 0.08222646266222\n", + "Starting epoch 1778\n", + "Train loss: 0.05172449007630348, Test loss: 0.07791154086589813\n", + "Starting epoch 1779\n", + "Train loss: 0.05115801878273487, Test loss: 0.08800241351127625\n", + "Starting epoch 1780\n", + "Train loss: 0.05094582632184028, Test loss: 0.07397373765707016\n", + "Starting epoch 1781\n", + "Train loss: 0.05111303962767124, Test loss: 0.08470065146684647\n", + "Starting epoch 1782\n", + "Train loss: 0.050732662826776506, Test loss: 0.07287557423114777\n", + "Starting epoch 1783\n", + "Train loss: 0.050853393599390985, Test loss: 0.07271262258291245\n", + "Starting epoch 1784\n", + "Train loss: 0.051026956140995026, Test loss: 0.08624856173992157\n", + "Starting epoch 1785\n", + "Train loss: 0.050882255211472514, Test loss: 0.07399197667837143\n", + "Starting epoch 1786\n", + "Train loss: 0.05074858129024506, Test loss: 0.07090982794761658\n", + "Starting epoch 1787\n", + "Train loss: 0.051068407744169236, Test loss: 0.07052581757307053\n", + "Starting epoch 1788\n", + "Train loss: 0.05078314535319805, Test loss: 0.08058387041091919\n", + "Starting epoch 1789\n", + "Train loss: 0.05038471095263958, Test loss: 0.09000159800052643\n", + "Starting epoch 1790\n", + "Train loss: 0.05047070935368538, Test loss: 0.07494382560253143\n", + "Starting epoch 1791\n", + "Train loss: 0.05070013277232647, Test loss: 0.07807411998510361\n", + "Starting epoch 1792\n", + "Train loss: 0.05057052530348301, Test loss: 0.08593181520700455\n", + "Starting epoch 1793\n", + "Train loss: 0.04988313086330891, Test loss: 0.0938742458820343\n", + "Starting epoch 1794\n", + "Train loss: 0.05037168353796005, Test loss: 0.0822572261095047\n", + "Starting epoch 1795\n", + "Train loss: 0.05030250772833824, Test loss: 0.08670920133590698\n", + "Starting epoch 1796\n", + "Train loss: 0.05017186626791954, Test loss: 0.06985318660736084\n", + "Starting epoch 1797\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.05031939283013344, Test loss: 0.07280831038951874\n", + "Starting epoch 1798\n", + "Train loss: 0.049936127439141276, Test loss: 0.07143031805753708\n", + "Starting epoch 1799\n", + "Train loss: 0.05038512080907822, Test loss: 0.08557024598121643\n", + "Starting epoch 1800\n", + "Train loss: 0.049817077592015264, Test loss: 0.07503828406333923\n", + "Starting epoch 1801\n", + "Train loss: 0.050378338545560834, Test loss: 0.08368786424398422\n", + "Starting epoch 1802\n", + "Train loss: 0.049439595714211464, Test loss: 0.06969894468784332\n", + "Starting epoch 1803\n", + "Train loss: 0.05012215226888657, Test loss: 0.08474048972129822\n", + "Starting epoch 1804\n", + "Train loss: 0.04984707400202751, Test loss: 0.07571408897638321\n", + "Starting epoch 1805\n", + "Train loss: 0.050239021629095076, Test loss: 0.08687729388475418\n", + "Starting epoch 1806\n", + "Train loss: 0.04990849204361439, Test loss: 0.07318869233131409\n", + "Starting epoch 1807\n", + "Train loss: 0.04964451923966408, Test loss: 0.07146459817886353\n", + "Starting epoch 1808\n", + "Train loss: 0.049746012091636656, Test loss: 0.09580620378255844\n", + "Starting epoch 1809\n", + "Train loss: 0.050379182323813436, Test loss: 0.08027469366788864\n", + "Starting epoch 1810\n", + "Train loss: 0.04912651866674423, Test loss: 0.07232421636581421\n", + "Starting epoch 1811\n", + "Train loss: 0.04969143077731133, Test loss: 0.07299432903528214\n", + "Starting epoch 1812\n", + "Train loss: 0.04904382817447186, Test loss: 0.09031473845243454\n", + "Starting epoch 1813\n", + "Train loss: 0.049510106593370434, Test loss: 0.08614854514598846\n", + "Starting epoch 1814\n", + "Train loss: 0.049650682508945464, Test loss: 0.07077116519212723\n", + "Starting epoch 1815\n", + "Train loss: 0.04913757152855396, Test loss: 0.08580689132213593\n", + "Starting epoch 1816\n", + "Train loss: 0.04936315558850765, Test loss: 0.08810552954673767\n", + "Starting epoch 1817\n", + "Train loss: 0.04963691435754299, Test loss: 0.08356745541095734\n", + "Starting epoch 1818\n", + "Train loss: 0.048872927054762844, Test loss: 0.08928565680980682\n", + "Starting epoch 1819\n", + "Train loss: 0.04905723661184311, Test loss: 0.07097230851650238\n", + "Starting epoch 1820\n", + "Train loss: 0.04886439643800258, Test loss: 0.0914919301867485\n", + "Starting epoch 1821\n", + "Train loss: 0.04861641973257065, Test loss: 0.07643449306488037\n", + "Starting epoch 1822\n", + "Train loss: 0.04888561211526394, Test loss: 0.0728345513343811\n", + "Starting epoch 1823\n", + "Train loss: 0.04883846081793308, Test loss: 0.07510673254728317\n", + "Starting epoch 1824\n", + "Train loss: 0.04830372288823128, Test loss: 0.07368580251932144\n", + "Starting epoch 1825\n", + "Train loss: 0.048675807937979695, Test loss: 0.07137423008680344\n", + "Starting epoch 1826\n", + "Train loss: 0.04837139919400215, Test loss: 0.07872277498245239\n", + "Starting epoch 1827\n", + "Train loss: 0.04790018104016781, Test loss: 0.08545596152544022\n", + "Starting epoch 1828\n", + "Train loss: 0.048057422861456874, Test loss: 0.07035014033317566\n", + "Starting epoch 1829\n", + "Train loss: 0.04756544068455696, Test loss: 0.06763110309839249\n", + "Starting epoch 1830\n", + "Train loss: 0.046192761212587356, Test loss: 0.0634259507060051\n", + "Starting epoch 1831\n", + "Train loss: 0.044347723722457884, Test loss: 0.06082724407315254\n", + "Starting epoch 1832\n", + "Train loss: 0.04486418060958385, Test loss: 0.06715810298919678\n", + "Starting epoch 1833\n", + "Train loss: 0.044649554193019865, Test loss: 0.06088043749332428\n", + "Starting epoch 1834\n", + "Train loss: 0.04366676941514015, Test loss: 0.08104386925697327\n", + "Starting epoch 1835\n", + "Train loss: 0.043895396441221236, Test loss: 0.07758144289255142\n", + "Starting epoch 1836\n", + "Train loss: 0.04333030007779598, Test loss: 0.07343171536922455\n", + "Starting epoch 1837\n", + "Train loss: 0.043765967264771465, Test loss: 0.06896915286779404\n", + "Starting epoch 1838\n", + "Train loss: 0.042429708763957026, Test loss: 0.062089044600725174\n", + "Starting epoch 1839\n", + "Train loss: 0.04308285482227802, Test loss: 0.0631263479590416\n", + "Starting epoch 1840\n", + "Train loss: 0.042980520352721215, Test loss: 0.06322992593050003\n", + "Starting epoch 1841\n", + "Train loss: 0.0429382736235857, Test loss: 0.06096827983856201\n", + "Starting epoch 1842\n", + "Train loss: 0.04290293000638485, Test loss: 0.058979298919439316\n", + "Starting epoch 1843\n", + "Train loss: 0.04262166291475296, Test loss: 0.061555374413728714\n", + "Starting epoch 1844\n", + "Train loss: 0.042391585931181905, Test loss: 0.06306560337543488\n", + "Starting epoch 1845\n", + "Train loss: 0.04225565306842327, Test loss: 0.07311160117387772\n", + "Starting epoch 1846\n", + "Train loss: 0.04234776571393013, Test loss: 0.06647884100675583\n", + "Starting epoch 1847\n", + "Train loss: 0.04249917320907116, Test loss: 0.05906754359602928\n", + "Starting epoch 1848\n", + "Train loss: 0.04245627366006374, Test loss: 0.05832820385694504\n", + "Starting epoch 1849\n", + "Train loss: 0.042099080979824066, Test loss: 0.08099135756492615\n", + "Starting epoch 1850\n", + "Train loss: 0.04169776618480683, Test loss: 0.06664618849754333\n", + "Starting epoch 1851\n", + "Train loss: 0.04168556362390518, Test loss: 0.059212394058704376\n", + "Starting epoch 1852\n", + "Train loss: 0.04191537894308567, Test loss: 0.06983485817909241\n", + "Starting epoch 1853\n", + "Train loss: 0.041807743832468984, Test loss: 0.06414823234081268\n", + "Starting epoch 1854\n", + "Train loss: 0.041915333420038225, Test loss: 0.061633091419935226\n", + "Starting epoch 1855\n", + "Train loss: 0.041755208149552345, Test loss: 0.06652864068746567\n", + "Starting epoch 1856\n", + "Train loss: 0.041666617542505266, Test loss: 0.06809818744659424\n", + "Starting epoch 1857\n", + "Train loss: 0.04177802927792072, Test loss: 0.06137015298008919\n", + "Starting epoch 1858\n", + "Train loss: 0.041774657070636746, Test loss: 0.07624385505914688\n", + "Starting epoch 1859\n", + "Train loss: 0.04168532781302929, Test loss: 0.05783018469810486\n", + "Starting epoch 1860\n", + "Train loss: 0.04165995955467224, Test loss: 0.0665426254272461\n", + "Starting epoch 1861\n", + "Train loss: 0.04162552751600743, Test loss: 0.07599304616451263\n", + "Starting epoch 1862\n", + "Train loss: 0.04149510227143764, Test loss: 0.06588569283485413\n", + "Starting epoch 1863\n", + "Train loss: 0.04109880335628986, Test loss: 0.07453429698944092\n", + "Starting epoch 1864\n", + "Train loss: 0.041548939570784565, Test loss: 0.07176082581281662\n", + "Starting epoch 1865\n", + "Train loss: 0.04094580337405205, Test loss: 0.07093556970357895\n", + "Starting epoch 1866\n", + "Train loss: 0.04112445801496506, Test loss: 0.0602845624089241\n", + "Starting epoch 1867\n", + "Train loss: 0.041469057351350785, Test loss: 0.05750831589102745\n", + "Starting epoch 1868\n", + "Train loss: 0.04099199891090393, Test loss: 0.06057927757501602\n", + "Starting epoch 1869\n", + "Train loss: 0.04095262803137303, Test loss: 0.08050108700990677\n", + "Starting epoch 1870\n", + "Train loss: 0.04080803088843823, Test loss: 0.07260160893201828\n", + "Starting epoch 1871\n", + "Train loss: 0.04094062864780426, Test loss: 0.06006539613008499\n", + "Starting epoch 1872\n", + "Train loss: 0.04079399138689041, Test loss: 0.0635182112455368\n", + "Starting epoch 1873\n", + "Train loss: 0.0408877956867218, Test loss: 0.07885871082544327\n", + "Starting epoch 1874\n", + "Train loss: 0.04065374054014683, Test loss: 0.055711425840854645\n", + "Starting epoch 1875\n", + "Train loss: 0.040800605565309525, Test loss: 0.0734555795788765\n", + "Starting epoch 1876\n", + "Train loss: 0.04078435458242893, Test loss: 0.06420619785785675\n", + "Starting epoch 1877\n", + "Train loss: 0.04077854111790657, Test loss: 0.05737517774105072\n", + "Starting epoch 1878\n", + "Train loss: 0.04022541254758835, Test loss: 0.06110524386167526\n", + "Starting epoch 1879\n", + "Train loss: 0.04004470899701119, Test loss: 0.07739400118589401\n", + "Starting epoch 1880\n", + "Train loss: 0.037747508212924004, Test loss: 0.07130111753940582\n", + "Starting epoch 1881\n", + "Train loss: 0.03576098661869764, Test loss: 0.06540507078170776\n", + "Starting epoch 1882\n", + "Train loss: 0.034170336723327636, Test loss: 0.07106290757656097\n", + "Starting epoch 1883\n", + "Train loss: 0.03264734800904989, Test loss: 0.058315861970186234\n", + "Starting epoch 1884\n", + "Train loss: 0.03170620419085026, Test loss: 0.06823182106018066\n", + "Starting epoch 1885\n", + "Train loss: 0.03087764620780945, Test loss: 0.059688396751880646\n", + "Starting epoch 1886\n", + "Train loss: 0.030044879354536532, Test loss: 0.06213729828596115\n", + "Starting epoch 1887\n", + "Train loss: 0.029417962357401847, Test loss: 0.06358521431684494\n", + "Starting epoch 1888\n", + "Train loss: 0.028959463872015475, Test loss: 0.060796186327934265\n", + "Starting epoch 1889\n", + "Train loss: 0.02836841680109501, Test loss: 0.05682162940502167\n", + "Starting epoch 1890\n", + "Train loss: 0.027919573709368706, Test loss: 0.06260661035776138\n", + "Starting epoch 1891\n", + "Train loss: 0.02758781272917986, Test loss: 0.05432625114917755\n", + "Starting epoch 1892\n", + "Train loss: 0.027306592240929604, Test loss: 0.06759623438119888\n", + "Starting epoch 1893\n", + "Train loss: 0.026835125237703324, Test loss: 0.062520332634449\n", + "Starting epoch 1894\n", + "Train loss: 0.02675978846848011, Test loss: 0.05768334120512009\n", + "Starting epoch 1895\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.026466577872633935, Test loss: 0.06523439288139343\n", + "Starting epoch 1896\n", + "Train loss: 0.02657249480485916, Test loss: 0.05956991761922836\n", + "Starting epoch 1897\n", + "Train loss: 0.026187828853726388, Test loss: 0.05808895453810692\n", + "Starting epoch 1898\n", + "Train loss: 0.02581321705132723, Test loss: 0.057499077171087265\n", + "Starting epoch 1899\n", + "Train loss: 0.025826020501554012, Test loss: 0.06162785738706589\n", + "Starting epoch 1900\n", + "Train loss: 0.02579596497118473, Test loss: 0.057970985770225525\n", + "Starting epoch 1901\n", + "Train loss: 0.025425999127328396, Test loss: 0.06361131370067596\n", + "Starting epoch 1902\n", + "Train loss: 0.025376397669315338, Test loss: 0.05993163585662842\n", + "Starting epoch 1903\n", + "Train loss: 0.02518328282982111, Test loss: 0.06071762368083\n", + "Starting epoch 1904\n", + "Train loss: 0.025027818717062472, Test loss: 0.05880197882652283\n", + "Starting epoch 1905\n", + "Train loss: 0.025053195394575595, Test loss: 0.05880309268832207\n", + "Starting epoch 1906\n", + "Train loss: 0.024871369563043116, Test loss: 0.056384339928627014\n", + "Starting epoch 1907\n", + "Train loss: 0.024707077294588087, Test loss: 0.05438343435525894\n", + "Starting epoch 1908\n", + "Train loss: 0.024707734957337378, Test loss: 0.05425751581788063\n", + "Starting epoch 1909\n", + "Train loss: 0.024644218906760215, Test loss: 0.061313092708587646\n", + "Starting epoch 1910\n", + "Train loss: 0.02447779167443514, Test loss: 0.055645063519477844\n", + "Starting epoch 1911\n", + "Train loss: 0.024469102434813976, Test loss: 0.05691491439938545\n", + "Starting epoch 1912\n", + "Train loss: 0.02437082778662443, Test loss: 0.05818954482674599\n", + "Starting epoch 1913\n", + "Train loss: 0.024301865063607694, Test loss: 0.0581323504447937\n", + "Starting epoch 1914\n", + "Train loss: 0.02427196305245161, Test loss: 0.06209557130932808\n", + "Starting epoch 1915\n", + "Train loss: 0.024210257083177568, Test loss: 0.0551280714571476\n", + "Starting epoch 1916\n", + "Train loss: 0.024191825315356255, Test loss: 0.060365259647369385\n", + "Starting epoch 1917\n", + "Train loss: 0.024081883653998375, Test loss: 0.05625700578093529\n", + "Starting epoch 1918\n", + "Train loss: 0.0239561577886343, Test loss: 0.05217632278800011\n", + "Starting epoch 1919\n", + "Train loss: 0.023783498108386994, Test loss: 0.05590946599841118\n", + "Starting epoch 1920\n", + "Train loss: 0.02378921549767256, Test loss: 0.05902876332402229\n", + "Starting epoch 1921\n", + "Train loss: 0.023888502083718777, Test loss: 0.05339876562356949\n", + "Starting epoch 1922\n", + "Train loss: 0.02375606767833233, Test loss: 0.054695673286914825\n", + "Starting epoch 1923\n", + "Train loss: 0.023707159571349622, Test loss: 0.05801649019122124\n", + "Starting epoch 1924\n", + "Train loss: 0.023662021942436696, Test loss: 0.05269021540880203\n", + "Starting epoch 1925\n", + "Train loss: 0.023520503155887128, Test loss: 0.05843066796660423\n", + "Starting epoch 1926\n", + "Train loss: 0.02345291819423437, Test loss: 0.055702224373817444\n", + "Starting epoch 1927\n", + "Train loss: 0.02326694142073393, Test loss: 0.05839908868074417\n", + "Starting epoch 1928\n", + "Train loss: 0.023274349607527257, Test loss: 0.053844090551137924\n", + "Starting epoch 1929\n", + "Train loss: 0.02302941545844078, Test loss: 0.05796913430094719\n", + "Starting epoch 1930\n", + "Train loss: 0.023226686865091325, Test loss: 0.058226484805345535\n", + "Starting epoch 1931\n", + "Train loss: 0.023269794397056103, Test loss: 0.05193103849887848\n", + "Starting epoch 1932\n", + "Train loss: 0.023004418574273586, Test loss: 0.049324579536914825\n", + "Starting epoch 1933\n", + "Train loss: 0.0232508110627532, Test loss: 0.05692097917199135\n", + "Starting epoch 1934\n", + "Train loss: 0.023245652467012407, Test loss: 0.056595541536808014\n", + "Starting epoch 1935\n", + "Train loss: 0.023240169398486613, Test loss: 0.05293763801455498\n", + "Starting epoch 1936\n", + "Train loss: 0.022963469661772252, Test loss: 0.05763852223753929\n", + "Starting epoch 1937\n", + "Train loss: 0.023053142093122005, Test loss: 0.05687861517071724\n", + "Starting epoch 1938\n", + "Train loss: 0.022901302874088286, Test loss: 0.056037478148937225\n", + "Starting epoch 1939\n", + "Train loss: 0.02288473319262266, Test loss: 0.05485739931464195\n", + "Starting epoch 1940\n", + "Train loss: 0.022985660769045353, Test loss: 0.052705563604831696\n", + "Starting epoch 1941\n", + "Train loss: 0.02286418206989765, Test loss: 0.05332188308238983\n", + "Starting epoch 1942\n", + "Train loss: 0.022787893004715443, Test loss: 0.05379366874694824\n", + "Starting epoch 1943\n", + "Train loss: 0.02270457733422518, Test loss: 0.058900147676467896\n", + "Starting epoch 1944\n", + "Train loss: 0.02259230464696884, Test loss: 0.05239087715744972\n", + "Starting epoch 1945\n", + "Train loss: 0.02260917466133833, Test loss: 0.05198966711759567\n", + "Starting epoch 1946\n", + "Train loss: 0.022579322308301925, Test loss: 0.05688270926475525\n", + "Starting epoch 1947\n", + "Train loss: 0.02252401623874903, Test loss: 0.05226026475429535\n", + "Starting epoch 1948\n", + "Train loss: 0.022543756440281868, Test loss: 0.058123745024204254\n", + "Starting epoch 1949\n", + "Train loss: 0.022434647120535375, Test loss: 0.0483047254383564\n", + "Starting epoch 1950\n", + "Train loss: 0.02242034699767828, Test loss: 0.05502728000283241\n", + "Starting epoch 1951\n", + "Train loss: 0.02234582643955946, Test loss: 0.053170979022979736\n", + "Starting epoch 1952\n", + "Train loss: 0.022292362153530122, Test loss: 0.05502784252166748\n", + "Starting epoch 1953\n", + "Train loss: 0.022263308726251124, Test loss: 0.056571800261735916\n", + "Starting epoch 1954\n", + "Train loss: 0.022238240726292133, Test loss: 0.05524539202451706\n", + "Starting epoch 1955\n", + "Train loss: 0.02229790735989809, Test loss: 0.05462072789669037\n", + "Starting epoch 1956\n", + "Train loss: 0.022283168397843837, Test loss: 0.05252789333462715\n", + "Starting epoch 1957\n", + "Train loss: 0.022202948033809664, Test loss: 0.05255329608917236\n", + "Starting epoch 1958\n", + "Train loss: 0.02213085114955902, Test loss: 0.05078111216425896\n", + "Starting epoch 1959\n", + "Train loss: 0.0220580992102623, Test loss: 0.05309412628412247\n", + "Starting epoch 1960\n", + "Train loss: 0.022057415395975114, Test loss: 0.05042993649840355\n", + "Starting epoch 1961\n", + "Train loss: 0.02200995396822691, Test loss: 0.052040454000234604\n", + "Starting epoch 1962\n", + "Train loss: 0.022043879628181457, Test loss: 0.05280786380171776\n", + "Starting epoch 1963\n", + "Train loss: 0.021967017389833927, Test loss: 0.05527549982070923\n", + "Starting epoch 1964\n", + "Train loss: 0.02200009822845459, Test loss: 0.05179332196712494\n", + "Starting epoch 1965\n", + "Train loss: 0.021965313628315927, Test loss: 0.04916156455874443\n", + "Starting epoch 1966\n", + "Train loss: 0.02189308378845453, Test loss: 0.053076352924108505\n", + "Starting epoch 1967\n", + "Train loss: 0.02191346600651741, Test loss: 0.05129900947213173\n", + "Starting epoch 1968\n", + "Train loss: 0.021830542348325252, Test loss: 0.04509676620364189\n", + "Starting epoch 1969\n", + "Train loss: 0.021192361041903494, Test loss: 0.04221038520336151\n", + "Starting epoch 1970\n", + "Train loss: 0.021816130205988885, Test loss: 0.050008587539196014\n", + "Starting epoch 1971\n", + "Train loss: 0.021960563622415066, Test loss: 0.052027225494384766\n", + "Starting epoch 1972\n", + "Train loss: 0.02194459058344364, Test loss: 0.04808204621076584\n", + "Starting epoch 1973\n", + "Train loss: 0.021906075701117515, Test loss: 0.04999813809990883\n", + "Starting epoch 1974\n", + "Train loss: 0.021759676672518255, Test loss: 0.04961545765399933\n", + "Starting epoch 1975\n", + "Train loss: 0.021718162111938, Test loss: 0.0507676936686039\n", + "Starting epoch 1976\n", + "Train loss: 0.021645201928913595, Test loss: 0.05268407240509987\n", + "Starting epoch 1977\n", + "Train loss: 0.021634825728833674, Test loss: 0.04727649316191673\n", + "Starting epoch 1978\n", + "Train loss: 0.021541524417698382, Test loss: 0.05302266776561737\n", + "Starting epoch 1979\n", + "Train loss: 0.021409081146121025, Test loss: 0.05162244290113449\n", + "Starting epoch 1980\n", + "Train loss: 0.021440673694014548, Test loss: 0.04523897543549538\n", + "Starting epoch 1981\n", + "Train loss: 0.02152136567980051, Test loss: 0.04960382357239723\n", + "Starting epoch 1982\n", + "Train loss: 0.021501718908548354, Test loss: 0.0480075441300869\n", + "Starting epoch 1983\n", + "Train loss: 0.021602153442800046, Test loss: 0.04893152788281441\n", + "Starting epoch 1984\n", + "Train loss: 0.02147187814116478, Test loss: 0.047491490840911865\n", + "Starting epoch 1985\n", + "Train loss: 0.02144068233668804, Test loss: 0.05261486768722534\n", + "Starting epoch 1986\n", + "Train loss: 0.021344390138983725, Test loss: 0.05026901140809059\n", + "Starting epoch 1987\n", + "Train loss: 0.02149045843631029, Test loss: 0.04564516618847847\n", + "Starting epoch 1988\n", + "Train loss: 0.021400151960551738, Test loss: 0.04468045383691788\n", + "Starting epoch 1989\n", + "Train loss: 0.021276156194508077, Test loss: 0.05051671341061592\n", + "Starting epoch 1990\n", + "Train loss: 0.02132299106568098, Test loss: 0.05369742214679718\n", + "Starting epoch 1991\n", + "Train loss: 0.021296781934797764, Test loss: 0.04858476668596268\n", + "Starting epoch 1992\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.021198114007711412, Test loss: 0.052307453006505966\n", + "Starting epoch 1993\n", + "Train loss: 0.02118571426719427, Test loss: 0.04677620530128479\n", + "Starting epoch 1994\n", + "Train loss: 0.021192117929458618, Test loss: 0.05300043150782585\n", + "Starting epoch 1995\n", + "Train loss: 0.021078874357044697, Test loss: 0.045020341873168945\n", + "Starting epoch 1996\n", + "Train loss: 0.021119009293615817, Test loss: 0.04641454666852951\n", + "Starting epoch 1997\n", + "Train loss: 0.021011554040014743, Test loss: 0.047330956906080246\n", + "Starting epoch 1998\n", + "Train loss: 0.021077455207705497, Test loss: 0.04511994123458862\n", + "Starting epoch 1999\n", + "Train loss: 0.021146302223205567, Test loss: 0.04367154464125633\n", + "Starting epoch 2000\n", + "Train loss: 0.02102144729346037, Test loss: 0.04819456860423088\n", + "Starting epoch 2001\n", + "Train loss: 0.021100222207605837, Test loss: 0.04685637727379799\n", + "Starting epoch 2002\n", + "Train loss: 0.020968985967338084, Test loss: 0.04372468590736389\n", + "Starting epoch 2003\n", + "Train loss: 0.021036774404346942, Test loss: 0.044876981526613235\n", + "Starting epoch 2004\n", + "Train loss: 0.02098987367004156, Test loss: 0.048090312629938126\n", + "Starting epoch 2005\n", + "Train loss: 0.021017327383160592, Test loss: 0.045213282108306885\n", + "Starting epoch 2006\n", + "Train loss: 0.02087822277098894, Test loss: 0.04803938791155815\n", + "Starting epoch 2007\n", + "Train loss: 0.020651721246540546, Test loss: 0.04830215126276016\n", + "Starting epoch 2008\n", + "Train loss: 0.020945245698094368, Test loss: 0.05185816064476967\n", + "Starting epoch 2009\n", + "Train loss: 0.02085471224039793, Test loss: 0.045494213700294495\n", + "Starting epoch 2010\n", + "Train loss: 0.020892838127911092, Test loss: 0.04671400785446167\n", + "Starting epoch 2011\n", + "Train loss: 0.020695973336696625, Test loss: 0.046920131891965866\n", + "Starting epoch 2012\n", + "Train loss: 0.020687882974743842, Test loss: 0.04759984090924263\n", + "Starting epoch 2013\n", + "Train loss: 0.020730231255292893, Test loss: 0.04639260470867157\n", + "Starting epoch 2014\n", + "Train loss: 0.02067291360348463, Test loss: 0.047661978751420975\n", + "Starting epoch 2015\n", + "Train loss: 0.020707371868193148, Test loss: 0.043859079480171204\n", + "Starting epoch 2016\n", + "Train loss: 0.02065607186406851, Test loss: 0.04711954668164253\n", + "Starting epoch 2017\n", + "Train loss: 0.02059365380555391, Test loss: 0.045943114906549454\n", + "Starting epoch 2018\n", + "Train loss: 0.020568616949021817, Test loss: 0.04355999454855919\n", + "Starting epoch 2019\n", + "Train loss: 0.020657685585319996, Test loss: 0.044644422829151154\n", + "Starting epoch 2020\n", + "Train loss: 0.02058703761547804, Test loss: 0.048102568835020065\n", + "Starting epoch 2021\n", + "Train loss: 0.020561900995671747, Test loss: 0.051487281918525696\n", + "Starting epoch 2022\n", + "Train loss: 0.020567063130438327, Test loss: 0.04561367258429527\n", + "Starting epoch 2023\n", + "Train loss: 0.020473762936890124, Test loss: 0.048825860023498535\n", + "Starting epoch 2024\n", + "Train loss: 0.020571776814758778, Test loss: 0.050479285418987274\n", + "Starting epoch 2025\n", + "Train loss: 0.020530365817248823, Test loss: 0.046223096549510956\n", + "Starting epoch 2026\n", + "Train loss: 0.020435388945043088, Test loss: 0.04407761991024017\n", + "Starting epoch 2027\n", + "Train loss: 0.020462001115083693, Test loss: 0.045283976942300797\n", + "Starting epoch 2028\n", + "Train loss: 0.0204389975219965, Test loss: 0.0471787191927433\n", + "Starting epoch 2029\n", + "Train loss: 0.02041975811123848, Test loss: 0.04581240937113762\n", + "Starting epoch 2030\n", + "Train loss: 0.02040696479380131, Test loss: 0.044052623212337494\n", + "Starting epoch 2031\n", + "Train loss: 0.020371033847332, Test loss: 0.044602829962968826\n", + "Starting epoch 2032\n", + "Train loss: 0.0202997962012887, Test loss: 0.049877848476171494\n", + "Starting epoch 2033\n", + "Train loss: 0.020346884317696094, Test loss: 0.04724615067243576\n", + "Starting epoch 2034\n", + "Train loss: 0.02030192032456398, Test loss: 0.04707741364836693\n", + "Starting epoch 2035\n", + "Train loss: 0.020278553552925585, Test loss: 0.04269106686115265\n", + "Starting epoch 2036\n", + "Train loss: 0.020294102542102337, Test loss: 0.04489927738904953\n", + "Starting epoch 2037\n", + "Train loss: 0.020283789709210395, Test loss: 0.04478560760617256\n", + "Starting epoch 2038\n", + "Train loss: 0.02023708313703537, Test loss: 0.04604508727788925\n", + "Starting epoch 2039\n", + "Train loss: 0.020102147944271564, Test loss: 0.043189406394958496\n", + "Starting epoch 2040\n", + "Train loss: 0.020265669971704484, Test loss: 0.041925106197595596\n", + "Starting epoch 2041\n", + "Train loss: 0.020303089506924152, Test loss: 0.04284597560763359\n", + "Starting epoch 2042\n", + "Train loss: 0.02019626207649708, Test loss: 0.04532612860202789\n", + "Starting epoch 2043\n", + "Train loss: 0.020320132151246072, Test loss: 0.043572477996349335\n", + "Starting epoch 2044\n", + "Train loss: 0.02008311040699482, Test loss: 0.04865844175219536\n", + "Starting epoch 2045\n", + "Train loss: 0.020208997912704946, Test loss: 0.03978574275970459\n", + "Starting epoch 2046\n", + "Train loss: 0.020207718685269355, Test loss: 0.04154612496495247\n", + "Starting epoch 2047\n", + "Train loss: 0.020181526094675065, Test loss: 0.04317038133740425\n", + "Starting epoch 2048\n", + "Train loss: 0.02018817339092493, Test loss: 0.04198060557246208\n", + "Starting epoch 2049\n", + "Train loss: 0.020129083320498467, Test loss: 0.0442316047847271\n", + "Starting epoch 2050\n", + "Train loss: 0.020077900998294353, Test loss: 0.04332442581653595\n", + "Starting epoch 2051\n", + "Train loss: 0.02000573240220547, Test loss: 0.04650530964136124\n", + "Starting epoch 2052\n", + "Train loss: 0.01995220221579075, Test loss: 0.04345207288861275\n", + "Starting epoch 2053\n", + "Train loss: 0.019976371712982655, Test loss: 0.039605725556612015\n", + "Starting epoch 2054\n", + "Train loss: 0.01996787339448929, Test loss: 0.04792588949203491\n", + "Starting epoch 2055\n", + "Train loss: 0.01989976465702057, Test loss: 0.04185771197080612\n", + "Starting epoch 2056\n", + "Train loss: 0.01986939113587141, Test loss: 0.03935110196471214\n", + "Starting epoch 2057\n", + "Train loss: 0.019925890192389488, Test loss: 0.040700968354940414\n", + "Starting epoch 2058\n", + "Train loss: 0.019883868396282197, Test loss: 0.04139452055096626\n", + "Starting epoch 2059\n", + "Train loss: 0.0198115099593997, Test loss: 0.04115395247936249\n", + "Starting epoch 2060\n", + "Train loss: 0.019829495288431646, Test loss: 0.04174274578690529\n", + "Starting epoch 2061\n", + "Train loss: 0.019757461100816727, Test loss: 0.04607722535729408\n", + "Starting epoch 2062\n", + "Train loss: 0.019621331579983234, Test loss: 0.03253643587231636\n", + "Starting epoch 2063\n", + "Train loss: 0.019821068458259105, Test loss: 0.04196614772081375\n", + "Starting epoch 2064\n", + "Train loss: 0.019884468466043474, Test loss: 0.04166286811232567\n", + "Starting epoch 2065\n", + "Train loss: 0.01980845108628273, Test loss: 0.03872547671198845\n", + "Starting epoch 2066\n", + "Train loss: 0.01981883741915226, Test loss: 0.04207102209329605\n", + "Starting epoch 2067\n", + "Train loss: 0.019815947078168392, Test loss: 0.040584590286016464\n", + "Starting epoch 2068\n", + "Train loss: 0.0197950928658247, Test loss: 0.042765479534864426\n", + "Starting epoch 2069\n", + "Train loss: 0.019791033342480658, Test loss: 0.039142731577157974\n", + "Starting epoch 2070\n", + "Train loss: 0.01985072370618582, Test loss: 0.027023879811167717\n", + "Starting epoch 2071\n", + "Train loss: 0.02026649296283722, Test loss: 0.02335016243159771\n", + "Starting epoch 2072\n", + "Train loss: 0.019827956072986125, Test loss: 0.02381184697151184\n", + "Starting epoch 2073\n", + "Train loss: 0.019840459078550338, Test loss: 0.02181403525173664\n", + "Starting epoch 2074\n", + "Train loss: 0.01988290760666132, Test loss: 0.02453801967203617\n", + "Starting epoch 2075\n", + "Train loss: 0.019760296046733856, Test loss: 0.025301024317741394\n", + "Starting epoch 2076\n", + "Train loss: 0.0196921918168664, Test loss: 0.023298470303416252\n", + "Starting epoch 2077\n", + "Train loss: 0.01965638976544142, Test loss: 0.022463813424110413\n", + "Starting epoch 2078\n", + "Train loss: 0.019585579968988895, Test loss: 0.022522129118442535\n", + "Starting epoch 2079\n", + "Train loss: 0.019602192379534245, Test loss: 0.023250825703144073\n", + "Starting epoch 2080\n", + "Train loss: 0.019529640451073648, Test loss: 0.022989269345998764\n", + "Starting epoch 2081\n", + "Train loss: 0.01951862346380949, Test loss: 0.024582574144005775\n", + "Starting epoch 2082\n", + "Train loss: 0.019520900622010232, Test loss: 0.023142749443650246\n", + "Starting epoch 2083\n", + "Train loss: 0.01937850508838892, Test loss: 0.023701611906290054\n", + "Starting epoch 2084\n", + "Train loss: 0.019425585232675076, Test loss: 0.021565841510891914\n", + "Starting epoch 2085\n", + "Train loss: 0.019436419419944288, Test loss: 0.02478446625173092\n", + "Starting epoch 2086\n", + "Train loss: 0.01937973767518997, Test loss: 0.021328838542103767\n", + "Starting epoch 2087\n", + "Train loss: 0.019366344586014748, Test loss: 0.02301379293203354\n", + "Starting epoch 2088\n", + "Train loss: 0.019405263625085354, Test loss: 0.023486141115427017\n", + "Starting epoch 2089\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.019410686567425728, Test loss: 0.02221180871129036\n", + "Starting epoch 2090\n", + "Train loss: 0.019351279996335506, Test loss: 0.023961124941706657\n", + "Starting epoch 2091\n", + "Train loss: 0.019298107624053956, Test loss: 0.02123734913766384\n", + "Starting epoch 2092\n", + "Train loss: 0.019382210075855257, Test loss: 0.0233148206025362\n", + "Starting epoch 2093\n", + "Train loss: 0.01923378333449364, Test loss: 0.021967854350805283\n", + "Starting epoch 2094\n", + "Train loss: 0.019246016405522825, Test loss: 0.026360103860497475\n", + "Starting epoch 2095\n", + "Train loss: 0.019165568854659795, Test loss: 0.04579531028866768\n", + "Starting epoch 2096\n", + "Train loss: 0.01956490468233824, Test loss: 0.0401647575199604\n", + "Starting epoch 2097\n", + "Train loss: 0.019388032890856267, Test loss: 0.03817995265126228\n", + "Starting epoch 2098\n", + "Train loss: 0.019458220042288303, Test loss: 0.03872714936733246\n", + "Starting epoch 2099\n", + "Train loss: 0.01937035482376814, Test loss: 0.04231109470129013\n", + "Starting epoch 2100\n", + "Train loss: 0.019329756759107114, Test loss: 0.04219914227724075\n", + "Starting epoch 2101\n", + "Train loss: 0.01934938434511423, Test loss: 0.04201962426304817\n", + "Starting epoch 2102\n", + "Train loss: 0.019300903975963592, Test loss: 0.044141605496406555\n", + "Starting epoch 2103\n", + "Train loss: 0.019197035878896714, Test loss: 0.039482902735471725\n", + "Starting epoch 2104\n", + "Train loss: 0.019188731908798218, Test loss: 0.0394609235227108\n", + "Starting epoch 2105\n", + "Train loss: 0.019128848761320115, Test loss: 0.03805123269557953\n", + "Starting epoch 2106\n", + "Train loss: 0.019142805971205233, Test loss: 0.038374681025743484\n", + "Starting epoch 2107\n", + "Train loss: 0.01913586921989918, Test loss: 0.041423290967941284\n", + "Starting epoch 2108\n", + "Train loss: 0.01907219473272562, Test loss: 0.03956782445311546\n", + "Starting epoch 2109\n", + "Train loss: 0.01904585774987936, Test loss: 0.03769832104444504\n", + "Starting epoch 2110\n", + "Train loss: 0.019080316945910454, Test loss: 0.038065966218709946\n", + "Starting epoch 2111\n", + "Train loss: 0.019163507968187332, Test loss: 0.03671453893184662\n", + "Starting epoch 2112\n", + "Train loss: 0.019071906208992004, Test loss: 0.035057954490184784\n", + "Starting epoch 2113\n", + "Train loss: 0.019144511409103872, Test loss: 0.04036194086074829\n", + "Starting epoch 2114\n", + "Train loss: 0.01904125455766916, Test loss: 0.04374151676893234\n", + "Starting epoch 2115\n", + "Train loss: 0.019016896747052668, Test loss: 0.03767852112650871\n", + "Starting epoch 2116\n", + "Train loss: 0.019016517363488673, Test loss: 0.03721022978425026\n", + "Starting epoch 2117\n", + "Train loss: 0.019032991454005242, Test loss: 0.03591478615999222\n", + "Starting epoch 2118\n", + "Train loss: 0.01895733643323183, Test loss: 0.03474511578679085\n", + "Starting epoch 2119\n", + "Train loss: 0.018892764188349245, Test loss: 0.039314791560173035\n", + "Starting epoch 2120\n", + "Train loss: 0.018850072398781777, Test loss: 0.042078837752342224\n", + "Starting epoch 2121\n", + "Train loss: 0.01901771452277899, Test loss: 0.03671357408165932\n", + "Starting epoch 2122\n", + "Train loss: 0.018916176818311214, Test loss: 0.03536398336291313\n", + "Starting epoch 2123\n", + "Train loss: 0.01902858290821314, Test loss: 0.03965675085783005\n", + "Starting epoch 2124\n", + "Train loss: 0.01896260779350996, Test loss: 0.03705881908535957\n", + "Starting epoch 2125\n", + "Train loss: 0.018927214816212654, Test loss: 0.037766583263874054\n", + "Starting epoch 2126\n", + "Train loss: 0.0189072522893548, Test loss: 0.039202045649290085\n", + "Starting epoch 2127\n", + "Train loss: 0.01890500735491514, Test loss: 0.04072660207748413\n", + "Starting epoch 2128\n", + "Train loss: 0.018908384293317794, Test loss: 0.039446886628866196\n", + "Starting epoch 2129\n", + "Train loss: 0.018811973966658114, Test loss: 0.034850798547267914\n", + "Starting epoch 2130\n", + "Train loss: 0.018700219113379716, Test loss: 0.04570235684514046\n", + "Starting epoch 2131\n", + "Train loss: 0.018846163526177406, Test loss: 0.04082578793168068\n", + "Starting epoch 2132\n", + "Train loss: 0.018797321543097496, Test loss: 0.042124468833208084\n", + "Starting epoch 2133\n", + "Train loss: 0.01875049412250519, Test loss: 0.04064967483282089\n", + "Starting epoch 2134\n", + "Train loss: 0.018859678730368614, Test loss: 0.03564188629388809\n", + "Starting epoch 2135\n", + "Train loss: 0.01881157036870718, Test loss: 0.036214374005794525\n", + "Starting epoch 2136\n", + "Train loss: 0.018848903626203537, Test loss: 0.03768104314804077\n", + "Starting epoch 2137\n", + "Train loss: 0.018828990682959557, Test loss: 0.0359419621527195\n", + "Starting epoch 2138\n", + "Train loss: 0.01874207064509392, Test loss: 0.038259927183389664\n", + "Starting epoch 2139\n", + "Train loss: 0.018772257417440416, Test loss: 0.037789732217788696\n", + "Starting epoch 2140\n", + "Train loss: 0.01879082277417183, Test loss: 0.03750942274928093\n", + "Starting epoch 2141\n", + "Train loss: 0.018552082851529123, Test loss: 0.044584061950445175\n", + "Starting epoch 2142\n", + "Train loss: 0.018689649701118468, Test loss: 0.03845428675413132\n", + "Starting epoch 2143\n", + "Train loss: 0.01883617039769888, Test loss: 0.03410361707210541\n", + "Starting epoch 2144\n", + "Train loss: 0.01881988435983658, Test loss: 0.0387447364628315\n", + "Starting epoch 2145\n", + "Train loss: 0.01881197836250067, Test loss: 0.03833102062344551\n", + "Starting epoch 2146\n", + "Train loss: 0.018757206872105598, Test loss: 0.034420907497406006\n", + "Starting epoch 2147\n", + "Train loss: 0.018622641414403916, Test loss: 0.03732014819979668\n", + "Starting epoch 2148\n", + "Train loss: 0.018662739917635916, Test loss: 0.03934994339942932\n", + "Starting epoch 2149\n", + "Train loss: 0.01873536728322506, Test loss: 0.03530032932758331\n", + "Starting epoch 2150\n", + "Train loss: 0.018655346781015394, Test loss: 0.03566153347492218\n", + "Starting epoch 2151\n", + "Train loss: 0.018622734807431696, Test loss: 0.03867049887776375\n", + "Starting epoch 2152\n", + "Train loss: 0.018525468166917563, Test loss: 0.02067350409924984\n", + "Starting epoch 2153\n", + "Train loss: 0.018657220453023912, Test loss: 0.018528973683714867\n", + "Starting epoch 2154\n", + "Train loss: 0.018752123787999155, Test loss: 0.020694240927696228\n", + "Starting epoch 2155\n", + "Train loss: 0.018618450313806535, Test loss: 0.019954858347773552\n", + "Starting epoch 2156\n", + "Train loss: 0.018577707447111607, Test loss: 0.020431049168109894\n", + "Starting epoch 2157\n", + "Train loss: 0.01862508950755, Test loss: 0.021306635811924934\n", + "Starting epoch 2158\n", + "Train loss: 0.018720145039260387, Test loss: 0.021292980760335922\n", + "Starting epoch 2159\n", + "Train loss: 0.018623144626617433, Test loss: 0.02054951898753643\n", + "Starting epoch 2160\n", + "Train loss: 0.018592525757849216, Test loss: 0.0209454745054245\n", + "Starting epoch 2161\n", + "Train loss: 0.01859654627740383, Test loss: 0.02119326964020729\n", + "Starting epoch 2162\n", + "Train loss: 0.018555751256644725, Test loss: 0.019113579764962196\n", + "Starting epoch 2163\n", + "Train loss: 0.01849451243877411, Test loss: 0.019127000123262405\n", + "Starting epoch 2164\n", + "Train loss: 0.01848384417593479, Test loss: 0.019362490624189377\n", + "Starting epoch 2165\n", + "Train loss: 0.01852773778140545, Test loss: 0.019488442689180374\n", + "Starting epoch 2166\n", + "Train loss: 0.018480473421514033, Test loss: 0.021057946607470512\n", + "Starting epoch 2167\n", + "Train loss: 0.018465960454195737, Test loss: 0.020053979009389877\n", + "Starting epoch 2168\n", + "Train loss: 0.018397269267588855, Test loss: 0.022995837032794952\n", + "Starting epoch 2169\n", + "Train loss: 0.018495513796806334, Test loss: 0.02023467980325222\n", + "Starting epoch 2170\n", + "Train loss: 0.018344370499253273, Test loss: 0.020597970113158226\n", + "Starting epoch 2171\n", + "Train loss: 0.018471831306815147, Test loss: 0.019620735198259354\n", + "Starting epoch 2172\n", + "Train loss: 0.018431235440075397, Test loss: 0.0201284047216177\n", + "Starting epoch 2173\n", + "Train loss: 0.018360949084162713, Test loss: 0.020334811881184578\n", + "Starting epoch 2174\n", + "Train loss: 0.018421908132731913, Test loss: 0.02065177634358406\n", + "Starting epoch 2175\n", + "Train loss: 0.018402155227959155, Test loss: 0.018046990036964417\n", + "Starting epoch 2176\n", + "Train loss: 0.018352855518460272, Test loss: 0.022292176261544228\n", + "Starting epoch 2177\n", + "Train loss: 0.01842761505395174, Test loss: 0.02097323350608349\n", + "Starting epoch 2178\n", + "Train loss: 0.0183395853638649, Test loss: 0.021731384098529816\n", + "Starting epoch 2179\n", + "Train loss: 0.018258525878190993, Test loss: 0.01914004050195217\n", + "Starting epoch 2180\n", + "Train loss: 0.0183895018696785, Test loss: 0.019849903881549835\n", + "Starting epoch 2181\n", + "Train loss: 0.01828854702413082, Test loss: 0.020165299996733665\n", + "Starting epoch 2182\n", + "Train loss: 0.018285201117396354, Test loss: 0.019947819411754608\n", + "Starting epoch 2183\n", + "Train loss: 0.018280431665480136, Test loss: 0.02016778104007244\n", + "Starting epoch 2184\n", + "Train loss: 0.018324316702783106, Test loss: 0.01944529451429844\n", + "Starting epoch 2185\n", + "Train loss: 0.018172055184841156, Test loss: 0.017996488139033318\n", + "Starting epoch 2186\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.018271752037107944, Test loss: 0.01930377632379532\n", + "Starting epoch 2187\n", + "Train loss: 0.018235794007778167, Test loss: 0.020357858389616013\n", + "Starting epoch 2188\n", + "Train loss: 0.01831539046019316, Test loss: 0.019389882683753967\n", + "Starting epoch 2189\n", + "Train loss: 0.018231831043958664, Test loss: 0.019372450187802315\n", + "Starting epoch 2190\n", + "Train loss: 0.0182117884978652, Test loss: 0.01988806016743183\n", + "Starting epoch 2191\n", + "Train loss: 0.018187677040696144, Test loss: 0.01933661662042141\n", + "Starting epoch 2192\n", + "Train loss: 0.018194106742739677, Test loss: 0.019041258841753006\n", + "Starting epoch 2193\n", + "Train loss: 0.018186942785978318, Test loss: 0.018551988527178764\n", + "Starting epoch 2194\n", + "Train loss: 0.01801816837862134, Test loss: 0.01948460377752781\n", + "Starting epoch 2195\n", + "Train loss: 0.018185098208487033, Test loss: 0.022601857781410217\n", + "Starting epoch 2196\n", + "Train loss: 0.018146482277661562, Test loss: 0.018999392166733742\n", + "Starting epoch 2197\n", + "Train loss: 0.018210685960948467, Test loss: 0.020325930789113045\n", + "Starting epoch 2198\n", + "Train loss: 0.018081774339079856, Test loss: 0.018526816740632057\n", + "Starting epoch 2199\n", + "Train loss: 0.017920967116951944, Test loss: 0.03520054742693901\n", + "Starting epoch 2200\n", + "Train loss: 0.01825382586568594, Test loss: 0.03298592194914818\n", + "Starting epoch 2201\n", + "Train loss: 0.01818548880517483, Test loss: 0.03564484789967537\n", + "Starting epoch 2202\n", + "Train loss: 0.018168878704309464, Test loss: 0.036750469356775284\n", + "Starting epoch 2203\n", + "Train loss: 0.01818069603294134, Test loss: 0.04054106026887894\n", + "Starting epoch 2204\n", + "Train loss: 0.0181681240350008, Test loss: 0.038772501051425934\n", + "Starting epoch 2205\n", + "Train loss: 0.017809434868395327, Test loss: 0.019591785967350006\n", + "Starting epoch 2206\n", + "Train loss: 0.018242568112909793, Test loss: 0.019002405926585197\n", + "Starting epoch 2207\n", + "Train loss: 0.018219020292162894, Test loss: 0.019757134839892387\n", + "Starting epoch 2208\n", + "Train loss: 0.01817495286464691, Test loss: 0.018354544416069984\n", + "Starting epoch 2209\n", + "Train loss: 0.01815566424280405, Test loss: 0.01860162988305092\n", + "Starting epoch 2210\n", + "Train loss: 0.018134172074496746, Test loss: 0.02103157714009285\n", + "Starting epoch 2211\n", + "Train loss: 0.017952193021774292, Test loss: 0.020603956654667854\n", + "Starting epoch 2212\n", + "Train loss: 0.01815620731562376, Test loss: 0.01825367845594883\n", + "Starting epoch 2213\n", + "Train loss: 0.018070730753242968, Test loss: 0.020089440047740936\n", + "Starting epoch 2214\n", + "Train loss: 0.018066148571670056, Test loss: 0.018695421516895294\n", + "Starting epoch 2215\n", + "Train loss: 0.01796230297535658, Test loss: 0.0207653921097517\n", + "Starting epoch 2216\n", + "Train loss: 0.017961106877774, Test loss: 0.02165962941944599\n", + "Starting epoch 2217\n", + "Train loss: 0.01804171912372112, Test loss: 0.020321231335401535\n", + "Starting epoch 2218\n", + "Train loss: 0.0179533264786005, Test loss: 0.021435139700770378\n", + "Starting epoch 2219\n", + "Train loss: 0.018098262157291174, Test loss: 0.019546542316675186\n", + "Starting epoch 2220\n", + "Train loss: 0.018007536605000495, Test loss: 0.01854395866394043\n", + "Starting epoch 2221\n", + "Train loss: 0.018622750509530307, Test loss: 0.020878667011857033\n", + "Starting epoch 2222\n", + "Train loss: 0.01811580153182149, Test loss: 0.0190883856266737\n", + "Starting epoch 2223\n", + "Train loss: 0.018190302178263663, Test loss: 0.01913895457983017\n", + "Starting epoch 2224\n", + "Train loss: 0.018118361718952657, Test loss: 0.01876891776919365\n", + "Starting epoch 2225\n", + "Train loss: 0.018036221042275428, Test loss: 0.01850653439760208\n", + "Starting epoch 2226\n", + "Train loss: 0.01805778346955776, Test loss: 0.018236657604575157\n", + "Starting epoch 2227\n", + "Train loss: 0.01797714076936245, Test loss: 0.02011980302631855\n", + "Starting epoch 2228\n", + "Train loss: 0.01796326257288456, Test loss: 0.019733380526304245\n", + "Starting epoch 2229\n", + "Train loss: 0.017973966114223004, Test loss: 0.020023107528686523\n", + "Starting epoch 2230\n", + "Train loss: 0.017950118593871592, Test loss: 0.019527820870280266\n", + "Starting epoch 2231\n", + "Train loss: 0.017887307815253734, Test loss: 0.018757876008749008\n", + "Starting epoch 2232\n", + "Train loss: 0.017846304662525653, Test loss: 0.018101833760738373\n", + "Starting epoch 2233\n", + "Train loss: 0.017901497818529608, Test loss: 0.020389696583151817\n", + "Starting epoch 2234\n", + "Train loss: 0.01782870780676603, Test loss: 0.018349992111325264\n", + "Starting epoch 2235\n", + "Train loss: 0.017774358037859202, Test loss: 0.018631208688020706\n", + "Starting epoch 2236\n", + "Train loss: 0.017855847217142583, Test loss: 0.01881830394268036\n", + "Starting epoch 2237\n", + "Train loss: 0.017848607823252677, Test loss: 0.0195716992020607\n", + "Starting epoch 2238\n", + "Train loss: 0.017745796274393797, Test loss: 0.020580843091011047\n", + "Starting epoch 2239\n", + "Train loss: 0.01790830012410879, Test loss: 0.019381137564778328\n", + "Starting epoch 2240\n", + "Train loss: 0.017771183736622332, Test loss: 0.018154211342334747\n", + "Starting epoch 2241\n", + "Train loss: 0.01784597657620907, Test loss: 0.01848175749182701\n", + "Starting epoch 2242\n", + "Train loss: 0.017775538265705108, Test loss: 0.01911516673862934\n", + "Starting epoch 2243\n", + "Train loss: 0.01776304066181183, Test loss: 0.01743950881063938\n", + "Starting epoch 2244\n", + "Train loss: 0.01779439691454172, Test loss: 0.018747584894299507\n", + "Starting epoch 2245\n", + "Train loss: 0.017766247019171715, Test loss: 0.019411904737353325\n", + "Starting epoch 2246\n", + "Train loss: 0.01779777094721794, Test loss: 0.018118571490049362\n", + "Starting epoch 2247\n", + "Train loss: 0.017770195119082927, Test loss: 0.018853111192584038\n", + "Starting epoch 2248\n", + "Train loss: 0.01768225997686386, Test loss: 0.019307512789964676\n", + "Starting epoch 2249\n", + "Train loss: 0.017735976316034793, Test loss: 0.01793432980775833\n", + "Starting epoch 2250\n", + "Train loss: 0.017772616408765315, Test loss: 0.020524105057120323\n", + "Starting epoch 2251\n", + "Train loss: 0.017634904999285935, Test loss: 0.017795760184526443\n", + "Starting epoch 2252\n", + "Train loss: 0.01765448644757271, Test loss: 0.018664808943867683\n", + "Starting epoch 2253\n", + "Train loss: 0.01774179220199585, Test loss: 0.01785155013203621\n", + "Starting epoch 2254\n", + "Train loss: 0.017748941518366337, Test loss: 0.01815228909254074\n", + "Starting epoch 2255\n", + "Train loss: 0.017741136886179447, Test loss: 0.018849318847060204\n", + "Starting epoch 2256\n", + "Train loss: 0.017577706407755612, Test loss: 0.017809079959988594\n", + "Starting epoch 2257\n", + "Train loss: 0.017668454572558404, Test loss: 0.017967261373996735\n", + "Starting epoch 2258\n", + "Train loss: 0.017651575598865748, Test loss: 0.018887128680944443\n", + "Starting epoch 2259\n", + "Train loss: 0.017707356102764606, Test loss: 0.017929047346115112\n", + "Starting epoch 2260\n", + "Train loss: 0.017689543180167674, Test loss: 0.018465327098965645\n", + "Starting epoch 2261\n", + "Train loss: 0.017730868235230447, Test loss: 0.017487112432718277\n", + "Starting epoch 2262\n", + "Train loss: 0.017678623162209986, Test loss: 0.01787840947508812\n", + "Starting epoch 2263\n", + "Train loss: 0.01758855476975441, Test loss: 0.019032828509807587\n", + "Starting epoch 2264\n", + "Train loss: 0.017638042010366915, Test loss: 0.018596714362502098\n", + "Starting epoch 2265\n", + "Train loss: 0.017604775577783584, Test loss: 0.020018385723233223\n", + "Starting epoch 2266\n", + "Train loss: 0.017673431895673275, Test loss: 0.018186312168836594\n", + "Starting epoch 2267\n", + "Train loss: 0.017542494907975197, Test loss: 0.018926434218883514\n", + "Starting epoch 2268\n", + "Train loss: 0.017621118333190678, Test loss: 0.016431983560323715\n", + "Starting epoch 2269\n", + "Train loss: 0.01751113874837756, Test loss: 0.018307318910956383\n", + "Starting epoch 2270\n", + "Train loss: 0.017575190905481576, Test loss: 0.017087411135435104\n", + "Starting epoch 2271\n", + "Train loss: 0.017606851104646923, Test loss: 0.020351450890302658\n", + "Starting epoch 2272\n", + "Train loss: 0.017641004249453543, Test loss: 0.017702531069517136\n", + "Starting epoch 2273\n", + "Train loss: 0.0176438607275486, Test loss: 0.017673177644610405\n", + "Starting epoch 2274\n", + "Train loss: 0.017582262456417082, Test loss: 0.01908881403505802\n", + "Starting epoch 2275\n", + "Train loss: 0.017563036903738977, Test loss: 0.019115960225462914\n", + "Starting epoch 2276\n", + "Train loss: 0.017491365000605583, Test loss: 0.0188354030251503\n", + "Starting epoch 2277\n", + "Train loss: 0.01755417577922344, Test loss: 0.016612647101283073\n", + "Starting epoch 2278\n", + "Train loss: 0.017502151727676392, Test loss: 0.017225833609700203\n", + "Starting epoch 2279\n", + "Train loss: 0.017606652677059173, Test loss: 0.017354147508740425\n", + "Starting epoch 2280\n", + "Train loss: 0.017522330693900585, Test loss: 0.0187898688018322\n", + "Starting epoch 2281\n", + "Train loss: 0.01749111920595169, Test loss: 0.01693800278007984\n", + "Starting epoch 2282\n", + "Train loss: 0.0175260728597641, Test loss: 0.017227953299880028\n", + "Starting epoch 2283\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.017517695128917696, Test loss: 0.019410111010074615\n", + "Starting epoch 2284\n", + "Train loss: 0.017409885693341495, Test loss: 0.017726365476846695\n", + "Starting epoch 2285\n", + "Train loss: 0.01761689081788063, Test loss: 0.018709998577833176\n", + "Starting epoch 2286\n", + "Train loss: 0.017533371541649103, Test loss: 0.01779100112617016\n", + "Starting epoch 2287\n", + "Train loss: 0.017495023533701897, Test loss: 0.020465711131691933\n", + "Starting epoch 2288\n", + "Train loss: 0.01746927186846733, Test loss: 0.017400870099663734\n", + "Starting epoch 2289\n", + "Train loss: 0.01745240431278944, Test loss: 0.020079519599676132\n", + "Starting epoch 2290\n", + "Train loss: 0.017362806778401137, Test loss: 0.02072438783943653\n", + "Starting epoch 2291\n", + "Train loss: 0.017430900856852533, Test loss: 0.021561842411756516\n", + "Starting epoch 2292\n", + "Train loss: 0.017323187813162805, Test loss: 0.01771678403019905\n", + "Starting epoch 2293\n", + "Train loss: 0.017215128671377897, Test loss: 0.022870292887091637\n", + "Starting epoch 2294\n", + "Train loss: 0.01736036267131567, Test loss: 0.027820510789752007\n", + "Starting epoch 2295\n", + "Train loss: 0.01758145207539201, Test loss: 0.03293022885918617\n", + "Starting epoch 2296\n", + "Train loss: 0.017613409943878652, Test loss: 0.03389449417591095\n", + "Starting epoch 2297\n", + "Train loss: 0.017567801997065546, Test loss: 0.038122907280921936\n", + "Starting epoch 2298\n", + "Train loss: 0.017450912874192, Test loss: 0.03563996031880379\n", + "Starting epoch 2299\n", + "Train loss: 0.01741249196231365, Test loss: 0.03145327791571617\n", + "Starting epoch 2300\n", + "Train loss: 0.017380823623389007, Test loss: 0.037848301231861115\n", + "Starting epoch 2301\n", + "Train loss: 0.01741334715858102, Test loss: 0.03330383449792862\n", + "Starting epoch 2302\n", + "Train loss: 0.01743981594219804, Test loss: 0.034298546612262726\n", + "Starting epoch 2303\n", + "Train loss: 0.017391244508326055, Test loss: 0.032098811119794846\n", + "Starting epoch 2304\n", + "Train loss: 0.017267494965344667, Test loss: 0.036526232957839966\n", + "Starting epoch 2305\n", + "Train loss: 0.01742931913584471, Test loss: 0.037274688482284546\n", + "Starting epoch 2306\n", + "Train loss: 0.01734086321666837, Test loss: 0.03253263607621193\n", + "Starting epoch 2307\n", + "Train loss: 0.017347137220203878, Test loss: 0.03241272643208504\n", + "Starting epoch 2308\n", + "Train loss: 0.017360515091568232, Test loss: 0.03711288049817085\n", + "Starting epoch 2309\n", + "Train loss: 0.017358822152018547, Test loss: 0.036395374685525894\n", + "Starting epoch 2310\n", + "Train loss: 0.01731393186375499, Test loss: 0.034708257764577866\n", + "Starting epoch 2311\n", + "Train loss: 0.017279839031398295, Test loss: 0.03332114592194557\n", + "Starting epoch 2312\n", + "Train loss: 0.017291351184248925, Test loss: 0.03618715703487396\n", + "Starting epoch 2313\n", + "Train loss: 0.017327619045972825, Test loss: 0.03481272608041763\n", + "Starting epoch 2314\n", + "Train loss: 0.017251845076680185, Test loss: 0.03350963070988655\n", + "Starting epoch 2315\n", + "Train loss: 0.01727556508034468, Test loss: 0.03212117776274681\n", + "Starting epoch 2316\n", + "Train loss: 0.01723538152873516, Test loss: 0.03286505490541458\n", + "Starting epoch 2317\n", + "Train loss: 0.01727357951924205, Test loss: 0.035146016627550125\n", + "Starting epoch 2318\n", + "Train loss: 0.017216596454381942, Test loss: 0.03803963586688042\n", + "Starting epoch 2319\n", + "Train loss: 0.01728496592491865, Test loss: 0.03557911887764931\n", + "Starting epoch 2320\n", + "Train loss: 0.017241216115653515, Test loss: 0.03752187639474869\n", + "Starting epoch 2321\n", + "Train loss: 0.017257231883704663, Test loss: 0.03475561738014221\n", + "Starting epoch 2322\n", + "Train loss: 0.017230992894619702, Test loss: 0.0378955714404583\n", + "Starting epoch 2323\n", + "Train loss: 0.01719549834728241, Test loss: 0.030713709071278572\n", + "Starting epoch 2324\n", + "Train loss: 0.017310834024101496, Test loss: 0.03860870748758316\n", + "Starting epoch 2325\n", + "Train loss: 0.01728316105902195, Test loss: 0.035303473472595215\n", + "Starting epoch 2326\n", + "Train loss: 0.01728643387556076, Test loss: 0.03449603170156479\n", + "Starting epoch 2327\n", + "Train loss: 0.017199465129524468, Test loss: 0.03505679592490196\n", + "Starting epoch 2328\n", + "Train loss: 0.017164431642740965, Test loss: 0.033917538821697235\n", + "Starting epoch 2329\n", + "Train loss: 0.0170913115888834, Test loss: 0.033080872148275375\n", + "Starting epoch 2330\n", + "Train loss: 0.01726319445297122, Test loss: 0.028415188193321228\n", + "Starting epoch 2331\n", + "Train loss: 0.017187695633620024, Test loss: 0.0301777645945549\n", + "Starting epoch 2332\n", + "Train loss: 0.01727515786886215, Test loss: 0.03572477400302887\n", + "Starting epoch 2333\n", + "Train loss: 0.0172011131234467, Test loss: 0.032514944672584534\n", + "Starting epoch 2334\n", + "Train loss: 0.017274427097290754, Test loss: 0.031726300716400146\n", + "Starting epoch 2335\n", + "Train loss: 0.017349246069788934, Test loss: 0.03703504428267479\n", + "Starting epoch 2336\n", + "Train loss: 0.017210708670318126, Test loss: 0.03431037813425064\n", + "Starting epoch 2337\n", + "Train loss: 0.017157025840133427, Test loss: 0.03653569146990776\n", + "Starting epoch 2338\n", + "Train loss: 0.01715904664248228, Test loss: 0.03644638508558273\n", + "Starting epoch 2339\n", + "Train loss: 0.01722095761448145, Test loss: 0.032076455652713776\n", + "Starting epoch 2340\n", + "Train loss: 0.01715108899399638, Test loss: 0.03316769003868103\n", + "Starting epoch 2341\n", + "Train loss: 0.017208270728588104, Test loss: 0.029713448137044907\n", + "Starting epoch 2342\n", + "Train loss: 0.017104294169694185, Test loss: 0.03446049615740776\n", + "Starting epoch 2343\n", + "Train loss: 0.01722651991993189, Test loss: 0.03613985702395439\n", + "Starting epoch 2344\n", + "Train loss: 0.01714855244383216, Test loss: 0.03370460495352745\n", + "Starting epoch 2345\n", + "Train loss: 0.01697824664413929, Test loss: 0.017734941095113754\n", + "Starting epoch 2346\n", + "Train loss: 0.017158043794333933, Test loss: 0.016077259555459023\n", + "Starting epoch 2347\n", + "Train loss: 0.017118173446506263, Test loss: 0.017696326598525047\n", + "Starting epoch 2348\n", + "Train loss: 0.017166616432368755, Test loss: 0.01668119803071022\n", + "Starting epoch 2349\n", + "Train loss: 0.017041706331074238, Test loss: 0.0185548048466444\n", + "Starting epoch 2350\n", + "Train loss: 0.017120683584362268, Test loss: 0.02156923897564411\n", + "Starting epoch 2351\n", + "Train loss: 0.016971869599074124, Test loss: 0.01887425221502781\n", + "Starting epoch 2352\n", + "Train loss: 0.01707080293446779, Test loss: 0.018284011632204056\n", + "Starting epoch 2353\n", + "Train loss: 0.017062962017953397, Test loss: 0.01860344409942627\n", + "Starting epoch 2354\n", + "Train loss: 0.017059290669858455, Test loss: 0.01663571037352085\n", + "Starting epoch 2355\n", + "Train loss: 0.01697699857875705, Test loss: 0.019925234839320183\n", + "Starting epoch 2356\n", + "Train loss: 0.017037728037685156, Test loss: 0.016679981723427773\n", + "Starting epoch 2357\n", + "Train loss: 0.016960874609649183, Test loss: 0.016682475805282593\n", + "Starting epoch 2358\n", + "Train loss: 0.01702421933412552, Test loss: 0.019548244774341583\n", + "Starting epoch 2359\n", + "Train loss: 0.017040083352476358, Test loss: 0.01749625988304615\n", + "Starting epoch 2360\n", + "Train loss: 0.016909348238259556, Test loss: 0.017920901998877525\n", + "Starting epoch 2361\n", + "Train loss: 0.017082419116050004, Test loss: 0.016427133232355118\n", + "Starting epoch 2362\n", + "Train loss: 0.01703111259266734, Test loss: 0.015749353915452957\n", + "Starting epoch 2363\n", + "Train loss: 0.01705647265538573, Test loss: 0.01653175801038742\n", + "Starting epoch 2364\n", + "Train loss: 0.017043164670467376, Test loss: 0.019115814939141273\n", + "Starting epoch 2365\n", + "Train loss: 0.016945014782249926, Test loss: 0.01968330144882202\n", + "Starting epoch 2366\n", + "Train loss: 0.01699199717491865, Test loss: 0.017414579167962074\n", + "Starting epoch 2367\n", + "Train loss: 0.01678783731535077, Test loss: 0.021507157012820244\n", + "Starting epoch 2368\n", + "Train loss: 0.01692648882046342, Test loss: 0.017251422628760338\n", + "Starting epoch 2369\n", + "Train loss: 0.017049386631697416, Test loss: 0.016260115429759026\n", + "Starting epoch 2370\n", + "Train loss: 0.016946627292782068, Test loss: 0.018896017223596573\n", + "Starting epoch 2371\n", + "Train loss: 0.01698960781097412, Test loss: 0.017498744651675224\n", + "Starting epoch 2372\n", + "Train loss: 0.016974624320864676, Test loss: 0.017858106642961502\n", + "Starting epoch 2373\n", + "Train loss: 0.01695160448551178, Test loss: 0.017504315823316574\n", + "Starting epoch 2374\n", + "Train loss: 0.016794541478157045, Test loss: 0.01815037988126278\n", + "Starting epoch 2375\n", + "Train loss: 0.01694217784330249, Test loss: 0.017272384837269783\n", + "Starting epoch 2376\n", + "Train loss: 0.016823538690805436, Test loss: 0.021172184497117996\n", + "Starting epoch 2377\n", + "Train loss: 0.01700552327558398, Test loss: 0.017089150846004486\n", + "Starting epoch 2378\n", + "Train loss: 0.017035526446998118, Test loss: 0.019306086003780365\n", + "Starting epoch 2379\n", + "Train loss: 0.01702820759266615, Test loss: 0.0159442201256752\n", + "Starting epoch 2380\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.016933523565530777, Test loss: 0.019471876323223114\n", + "Starting epoch 2381\n", + "Train loss: 0.0169099785014987, Test loss: 0.01608731597661972\n", + "Starting epoch 2382\n", + "Train loss: 0.016938608959317206, Test loss: 0.018824582919478416\n", + "Starting epoch 2383\n", + "Train loss: 0.016910126078873874, Test loss: 0.01874520070850849\n", + "Starting epoch 2384\n", + "Train loss: 0.016927834749221802, Test loss: 0.017280060797929764\n", + "Starting epoch 2385\n", + "Train loss: 0.016834607496857643, Test loss: 0.0179618950933218\n", + "Starting epoch 2386\n", + "Train loss: 0.016797603163868188, Test loss: 0.015867944806814194\n", + "Starting epoch 2387\n", + "Train loss: 0.016931932140141725, Test loss: 0.021056467667222023\n", + "Starting epoch 2388\n", + "Train loss: 0.016697321068495512, Test loss: 0.01771165430545807\n", + "Starting epoch 2389\n", + "Train loss: 0.016904480326920747, Test loss: 0.018936118111014366\n", + "Starting epoch 2390\n", + "Train loss: 0.016877289470285176, Test loss: 0.017235277220606804\n", + "Starting epoch 2391\n", + "Train loss: 0.016919767558574675, Test loss: 0.016443172469735146\n", + "Starting epoch 2392\n", + "Train loss: 0.016731517240405083, Test loss: 0.016554340720176697\n", + "Starting epoch 2393\n", + "Train loss: 0.016890621203929188, Test loss: 0.0175497867166996\n", + "Starting epoch 2394\n", + "Train loss: 0.01683309506624937, Test loss: 0.01732325740158558\n", + "Starting epoch 2395\n", + "Train loss: 0.016886883489787577, Test loss: 0.01994621939957142\n", + "Starting epoch 2396\n", + "Train loss: 0.01679583393037319, Test loss: 0.02027842029929161\n", + "Starting epoch 2397\n", + "Train loss: 0.01682364420965314, Test loss: 0.01824815385043621\n", + "Starting epoch 2398\n", + "Train loss: 0.016808930914849043, Test loss: 0.017375357449054718\n", + "Starting epoch 2399\n", + "Train loss: 0.01680874714627862, Test loss: 0.018731655552983284\n", + "Starting epoch 2400\n", + "Train loss: 0.016871173325926067, Test loss: 0.016682088375091553\n", + "Starting epoch 2401\n", + "Train loss: 0.016791666485369207, Test loss: 0.01730426773428917\n", + "Starting epoch 2402\n", + "Train loss: 0.016781705096364022, Test loss: 0.016085175797343254\n", + "Starting epoch 2403\n", + "Train loss: 0.01684186274185777, Test loss: 0.01863808184862137\n", + "Starting epoch 2404\n", + "Train loss: 0.01680937921628356, Test loss: 0.017990265041589737\n", + "Starting epoch 2405\n", + "Train loss: 0.01672470623627305, Test loss: 0.017815833911299706\n", + "Starting epoch 2406\n", + "Train loss: 0.01678182855248451, Test loss: 0.019435575231909752\n", + "Starting epoch 2407\n", + "Train loss: 0.016773562338203193, Test loss: 0.0170846339315176\n", + "Starting epoch 2408\n", + "Train loss: 0.016692768558859827, Test loss: 0.01870599389076233\n", + "Starting epoch 2409\n", + "Train loss: 0.016865627318620683, Test loss: 0.018343063071370125\n", + "Starting epoch 2410\n", + "Train loss: 0.016677318196743725, Test loss: 0.01703544519841671\n", + "Starting epoch 2411\n", + "Train loss: 0.016729315612465143, Test loss: 0.01640336588025093\n", + "Starting epoch 2412\n", + "Train loss: 0.01681311661377549, Test loss: 0.017151743173599243\n", + "Starting epoch 2413\n", + "Train loss: 0.016718471050262453, Test loss: 0.016791265457868576\n", + "Starting epoch 2414\n", + "Train loss: 0.01691916661337018, Test loss: 0.03700356185436249\n", + "Starting epoch 2415\n", + "Train loss: 0.016856127474457024, Test loss: 0.03165622055530548\n", + "Starting epoch 2416\n", + "Train loss: 0.01675062209367752, Test loss: 0.03402603045105934\n", + "Starting epoch 2417\n", + "Train loss: 0.01659922506660223, Test loss: 0.017386965453624725\n", + "Starting epoch 2418\n", + "Train loss: 0.016955741345882416, Test loss: 0.01832759566605091\n", + "Starting epoch 2419\n", + "Train loss: 0.016861429288983345, Test loss: 0.017089754343032837\n", + "Starting epoch 2420\n", + "Train loss: 0.01684281077235937, Test loss: 0.03458128497004509\n", + "Starting epoch 2421\n", + "Train loss: 0.01698864618316293, Test loss: 0.030886095017194748\n", + "Starting epoch 2422\n", + "Train loss: 0.0168280591070652, Test loss: 0.0370035283267498\n", + "Starting epoch 2423\n", + "Train loss: 0.016768888402730227, Test loss: 0.03138938173651695\n", + "Starting epoch 2424\n", + "Train loss: 0.016725603211671113, Test loss: 0.022272834554314613\n", + "Starting epoch 2425\n", + "Train loss: 0.016912296693772078, Test loss: 0.01874774694442749\n", + "Starting epoch 2426\n", + "Train loss: 0.01683709939941764, Test loss: 0.01656348817050457\n", + "Starting epoch 2427\n", + "Train loss: 0.016780951898545025, Test loss: 0.01692606322467327\n", + "Starting epoch 2428\n", + "Train loss: 0.01671815725043416, Test loss: 0.01682959869503975\n", + "Starting epoch 2429\n", + "Train loss: 0.016755004450678825, Test loss: 0.01588089019060135\n", + "Starting epoch 2430\n", + "Train loss: 0.01671066677197814, Test loss: 0.019621465355157852\n", + "Starting epoch 2431\n", + "Train loss: 0.01667891673743725, Test loss: 0.018759755417704582\n", + "Starting epoch 2432\n", + "Train loss: 0.01657909233123064, Test loss: 0.016883425414562225\n", + "Starting epoch 2433\n", + "Train loss: 0.016703260149806737, Test loss: 0.015481814742088318\n", + "Starting epoch 2434\n", + "Train loss: 0.01667736105620861, Test loss: 0.016609149053692818\n", + "Starting epoch 2435\n", + "Train loss: 0.01654303288087249, Test loss: 0.01731492578983307\n", + "Starting epoch 2436\n", + "Train loss: 0.01660675024613738, Test loss: 0.016170376911759377\n", + "Starting epoch 2437\n", + "Train loss: 0.016646220255643128, Test loss: 0.01719518192112446\n", + "Starting epoch 2438\n", + "Train loss: 0.016658070888370277, Test loss: 0.01668483018875122\n", + "Starting epoch 2439\n", + "Train loss: 0.01699783166870475, Test loss: 0.03176269307732582\n", + "Starting epoch 2440\n", + "Train loss: 0.016772423833608628, Test loss: 0.017463775351643562\n", + "Starting epoch 2441\n", + "Train loss: 0.01671503756195307, Test loss: 0.017295515164732933\n", + "Starting epoch 2442\n", + "Train loss: 0.016518174540251493, Test loss: 0.017340747639536858\n", + "Starting epoch 2443\n", + "Train loss: 0.016640520375221967, Test loss: 0.016449758782982826\n", + "Starting epoch 2444\n", + "Train loss: 0.016642945744097234, Test loss: 0.0186749417334795\n", + "Starting epoch 2445\n", + "Train loss: 0.01663224009796977, Test loss: 0.016754550859332085\n", + "Starting epoch 2446\n", + "Train loss: 0.01657114828005433, Test loss: 0.017483320087194443\n", + "Starting epoch 2447\n", + "Train loss: 0.01662233566865325, Test loss: 0.01748652569949627\n", + "Starting epoch 2448\n", + "Train loss: 0.016606419011950493, Test loss: 0.016456570476293564\n", + "Starting epoch 2449\n", + "Train loss: 0.01657965689897537, Test loss: 0.015626462176442146\n", + "Starting epoch 2450\n", + "Train loss: 0.016548244263976813, Test loss: 0.018162023276090622\n", + "Starting epoch 2451\n", + "Train loss: 0.01654094522818923, Test loss: 0.017386222258210182\n", + "Starting epoch 2452\n", + "Train loss: 0.01651269281283021, Test loss: 0.017558002844452858\n", + "Starting epoch 2453\n", + "Train loss: 0.01652479900047183, Test loss: 0.016223128885030746\n", + "Starting epoch 2454\n", + "Train loss: 0.0164486500993371, Test loss: 0.016931632533669472\n", + "Starting epoch 2455\n", + "Train loss: 0.016493380311876537, Test loss: 0.01543706376105547\n", + "Starting epoch 2456\n", + "Train loss: 0.016444821525365114, Test loss: 0.016035832464694977\n", + "Starting epoch 2457\n", + "Train loss: 0.0164613207988441, Test loss: 0.018962809816002846\n", + "Starting epoch 2458\n", + "Train loss: 0.016492690704762936, Test loss: 0.015185327269136906\n", + "Starting epoch 2459\n", + "Train loss: 0.016456549800932408, Test loss: 0.018683187663555145\n", + "Starting epoch 2460\n", + "Train loss: 0.016371292974799872, Test loss: 0.01785985194146633\n", + "Starting epoch 2461\n", + "Train loss: 0.016476905569434167, Test loss: 0.016770586371421814\n", + "Starting epoch 2462\n", + "Train loss: 0.016484266109764577, Test loss: 0.01540171355009079\n", + "Starting epoch 2463\n", + "Train loss: 0.01642627676948905, Test loss: 0.01690499112010002\n", + "Starting epoch 2464\n", + "Train loss: 0.016500820983201266, Test loss: 0.016903523355722427\n", + "Starting epoch 2465\n", + "Train loss: 0.01646294619888067, Test loss: 0.016666419804096222\n", + "Starting epoch 2466\n", + "Train loss: 0.01647839767858386, Test loss: 0.01637352630496025\n", + "Starting epoch 2467\n", + "Train loss: 0.01648515269160271, Test loss: 0.017593612894415855\n", + "Starting epoch 2468\n", + "Train loss: 0.016504910923540592, Test loss: 0.017615385353565216\n", + "Starting epoch 2469\n", + "Train loss: 0.01642874700948596, Test loss: 0.017124567180871964\n", + "Starting epoch 2470\n", + "Train loss: 0.01641391510143876, Test loss: 0.017192894592881203\n", + "Starting epoch 2471\n", + "Train loss: 0.016426687873899937, Test loss: 0.01662687212228775\n", + "Starting epoch 2472\n", + "Train loss: 0.016445749439299107, Test loss: 0.016025738790631294\n", + "Starting epoch 2473\n", + "Train loss: 0.016488200630992652, Test loss: 0.017939116805791855\n", + "Starting epoch 2474\n", + "Train loss: 0.0164240469224751, Test loss: 0.016313565894961357\n", + "Starting epoch 2475\n", + "Train loss: 0.016346108336001634, Test loss: 0.019598929211497307\n", + "Starting epoch 2476\n", + "Train loss: 0.016424241345375776, Test loss: 0.01644456945359707\n", + "Starting epoch 2477\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.016394154094159605, Test loss: 0.017011471092700958\n", + "Starting epoch 2478\n", + "Train loss: 0.016413086205720902, Test loss: 0.015551410615444183\n", + "Starting epoch 2479\n", + "Train loss: 0.016352020613849164, Test loss: 0.016229869797825813\n", + "Starting epoch 2480\n", + "Train loss: 0.01637901984155178, Test loss: 0.015080480836331844\n", + "Starting epoch 2481\n", + "Train loss: 0.01632297854870558, Test loss: 0.019334988668560982\n", + "Starting epoch 2482\n", + "Train loss: 0.016273592710494996, Test loss: 0.016599727794528008\n", + "Starting epoch 2483\n", + "Train loss: 0.01644375881180167, Test loss: 0.01682957075536251\n", + "Starting epoch 2484\n", + "Train loss: 0.016314322892576456, Test loss: 0.018862472847104073\n", + "Starting epoch 2485\n", + "Train loss: 0.016332814525812863, Test loss: 0.01779235526919365\n", + "Starting epoch 2486\n", + "Train loss: 0.016388618927448986, Test loss: 0.014991870149970055\n", + "Starting epoch 2487\n", + "Train loss: 0.016352825872600077, Test loss: 0.016049599274992943\n", + "Starting epoch 2488\n", + "Train loss: 0.01637860734015703, Test loss: 0.017295923084020615\n", + "Starting epoch 2489\n", + "Train loss: 0.016362490020692348, Test loss: 0.022630922496318817\n", + "Starting epoch 2490\n", + "Train loss: 0.01637588821351528, Test loss: 0.01600218564271927\n", + "Starting epoch 2491\n", + "Train loss: 0.016413011718541384, Test loss: 0.01768537238240242\n", + "Starting epoch 2492\n", + "Train loss: 0.016408360674977304, Test loss: 0.015605740249156952\n", + "Starting epoch 2493\n", + "Train loss: 0.016414422448724507, Test loss: 0.019046958535909653\n", + "Starting epoch 2494\n", + "Train loss: 0.016315783504396678, Test loss: 0.016675755381584167\n", + "Starting epoch 2495\n", + "Train loss: 0.01625262074172497, Test loss: 0.01661447063088417\n", + "Starting epoch 2496\n", + "Train loss: 0.016389903258532287, Test loss: 0.016330666840076447\n", + "Starting epoch 2497\n", + "Train loss: 0.01633164083585143, Test loss: 0.01909962110221386\n", + "Starting epoch 2498\n", + "Train loss: 0.016303976103663444, Test loss: 0.01836652308702469\n", + "Starting epoch 2499\n", + "Train loss: 0.01613803954795003, Test loss: 0.032397426664829254\n", + "Starting epoch 2500\n", + "Train loss: 0.016444292459636925, Test loss: 0.033019017428159714\n", + "Starting epoch 2501\n", + "Train loss: 0.016301109418272973, Test loss: 0.03436394780874252\n", + "Starting epoch 2502\n", + "Train loss: 0.016308269407600166, Test loss: 0.03476030007004738\n", + "Starting epoch 2503\n", + "Train loss: 0.016376600731164218, Test loss: 0.03051372431218624\n", + "Starting epoch 2504\n", + "Train loss: 0.01630165932700038, Test loss: 0.03277261555194855\n", + "Starting epoch 2505\n", + "Train loss: 0.016306099388748406, Test loss: 0.03439066559076309\n", + "Starting epoch 2506\n", + "Train loss: 0.016239195503294467, Test loss: 0.03501419723033905\n", + "Starting epoch 2507\n", + "Train loss: 0.016340789683163166, Test loss: 0.031132647767663002\n", + "Starting epoch 2508\n", + "Train loss: 0.01624752301722765, Test loss: 0.02831972949206829\n", + "Starting epoch 2509\n", + "Train loss: 0.016261409148573876, Test loss: 0.02751079946756363\n", + "Starting epoch 2510\n", + "Train loss: 0.016399816051125525, Test loss: 0.02897927723824978\n", + "Starting epoch 2511\n", + "Train loss: 0.016323834024369717, Test loss: 0.02922101691365242\n", + "Starting epoch 2512\n", + "Train loss: 0.016270738374441863, Test loss: 0.03067125380039215\n", + "Starting epoch 2513\n", + "Train loss: 0.016224622577428818, Test loss: 0.030943425372242928\n", + "Starting epoch 2514\n", + "Train loss: 0.016396601051092148, Test loss: 0.032102521508932114\n", + "Starting epoch 2515\n", + "Train loss: 0.016252265851944683, Test loss: 0.03416823595762253\n", + "Starting epoch 2516\n", + "Train loss: 0.016418549325317146, Test loss: 0.029305361211299896\n", + "Starting epoch 2517\n", + "Train loss: 0.01634999582543969, Test loss: 0.032953690737485886\n", + "Starting epoch 2518\n", + "Train loss: 0.01635210309177637, Test loss: 0.02688089944422245\n", + "Starting epoch 2519\n", + "Train loss: 0.016288701500743626, Test loss: 0.03274567425251007\n", + "Starting epoch 2520\n", + "Train loss: 0.01626358671113849, Test loss: 0.03277609124779701\n", + "Starting epoch 2521\n", + "Train loss: 0.01619711544364691, Test loss: 0.03142109885811806\n", + "Starting epoch 2522\n", + "Train loss: 0.016185547690838575, Test loss: 0.03269346430897713\n", + "Starting epoch 2523\n", + "Train loss: 0.016153167225420473, Test loss: 0.01602805033326149\n", + "Starting epoch 2524\n", + "Train loss: 0.01625209491699934, Test loss: 0.017083872109651566\n", + "Starting epoch 2525\n", + "Train loss: 0.016249069292098284, Test loss: 0.01745094358921051\n", + "Starting epoch 2526\n", + "Train loss: 0.016218976713716984, Test loss: 0.01534237153828144\n", + "Starting epoch 2527\n", + "Train loss: 0.016169673427939416, Test loss: 0.016881199553608894\n", + "Starting epoch 2528\n", + "Train loss: 0.016086535658687354, Test loss: 0.016501301899552345\n", + "Starting epoch 2529\n", + "Train loss: 0.016199423652142286, Test loss: 0.016918407753109932\n", + "Starting epoch 2530\n", + "Train loss: 0.016223709136247634, Test loss: 0.015447413548827171\n", + "Starting epoch 2531\n", + "Train loss: 0.016224326118826866, Test loss: 0.01968308538198471\n", + "Starting epoch 2532\n", + "Train loss: 0.016189732123166323, Test loss: 0.016617942601442337\n", + "Starting epoch 2533\n", + "Train loss: 0.01622051604092121, Test loss: 0.018824882805347443\n", + "Starting epoch 2534\n", + "Train loss: 0.016189020425081253, Test loss: 0.016861071810126305\n", + "Starting epoch 2535\n", + "Train loss: 0.0161066453717649, Test loss: 0.01855815015733242\n", + "Starting epoch 2536\n", + "Train loss: 0.01616079345345497, Test loss: 0.01675340346992016\n", + "Starting epoch 2537\n", + "Train loss: 0.016146733444184064, Test loss: 0.016634810715913773\n", + "Starting epoch 2538\n", + "Train loss: 0.01608660804107785, Test loss: 0.015239041298627853\n", + "Starting epoch 2539\n", + "Train loss: 0.016047168131917713, Test loss: 0.03282970190048218\n", + "Starting epoch 2540\n", + "Train loss: 0.01617617465555668, Test loss: 0.032074689865112305\n", + "Starting epoch 2541\n", + "Train loss: 0.016061413399875164, Test loss: 0.03345310315489769\n", + "Starting epoch 2542\n", + "Train loss: 0.01613468324765563, Test loss: 0.03154533728957176\n", + "Starting epoch 2543\n", + "Train loss: 0.016138222962617875, Test loss: 0.03359474986791611\n", + "Starting epoch 2544\n", + "Train loss: 0.016153009813278912, Test loss: 0.03310242295265198\n", + "Starting epoch 2545\n", + "Train loss: 0.016130767688155176, Test loss: 0.028145750984549522\n", + "Starting epoch 2546\n", + "Train loss: 0.01610901465639472, Test loss: 0.030950995162129402\n", + "Starting epoch 2547\n", + "Train loss: 0.016082446202635766, Test loss: 0.029068440198898315\n", + "Starting epoch 2548\n", + "Train loss: 0.016143099591135978, Test loss: 0.032722052186727524\n", + "Starting epoch 2549\n", + "Train loss: 0.01615977980196476, Test loss: 0.034151751548051834\n", + "Starting epoch 2550\n", + "Train loss: 0.016157314889132978, Test loss: 0.035731974989175797\n", + "Starting epoch 2551\n", + "Train loss: 0.016101841628551484, Test loss: 0.03163818269968033\n", + "Starting epoch 2552\n", + "Train loss: 0.016080492064356803, Test loss: 0.030047887936234474\n", + "Starting epoch 2553\n", + "Train loss: 0.016097035184502603, Test loss: 0.03065258264541626\n", + "Starting epoch 2554\n", + "Train loss: 0.016603271160274744, Test loss: 0.0236259363591671\n", + "Starting epoch 2555\n", + "Train loss: 0.019176311176270246, Test loss: 0.02195332944393158\n", + "Starting epoch 2556\n", + "Train loss: 0.017479845341295003, Test loss: 0.018721934407949448\n", + "Starting epoch 2557\n", + "Train loss: 0.016649973504245283, Test loss: 0.016041215509176254\n", + "Starting epoch 2558\n", + "Train loss: 0.016599668264389036, Test loss: 0.01797194965183735\n", + "Starting epoch 2559\n", + "Train loss: 0.016451167091727257, Test loss: 0.01832265406847\n", + "Starting epoch 2560\n", + "Train loss: 0.016332280114293097, Test loss: 0.016017476096749306\n", + "Starting epoch 2561\n", + "Train loss: 0.016307599991559982, Test loss: 0.014049559831619263\n", + "Starting epoch 2562\n", + "Train loss: 0.01626377960667014, Test loss: 0.01725035160779953\n", + "Starting epoch 2563\n", + "Train loss: 0.01615007871761918, Test loss: 0.015947330743074417\n", + "Starting epoch 2564\n", + "Train loss: 0.016199102457612753, Test loss: 0.016011299565434456\n", + "Starting epoch 2565\n", + "Train loss: 0.016209970396012068, Test loss: 0.017916277050971985\n", + "Starting epoch 2566\n", + "Train loss: 0.01618200646713376, Test loss: 0.018059518188238144\n", + "Starting epoch 2567\n", + "Train loss: 0.015966748967766762, Test loss: 0.0163283534348011\n", + "Starting epoch 2568\n", + "Train loss: 0.01619021600112319, Test loss: 0.018657619133591652\n", + "Starting epoch 2569\n", + "Train loss: 0.01621246976777911, Test loss: 0.016166482120752335\n", + "Starting epoch 2570\n", + "Train loss: 0.01612902918830514, Test loss: 0.016479697078466415\n", + "Starting epoch 2571\n", + "Train loss: 0.01606420150026679, Test loss: 0.016255926340818405\n", + "Starting epoch 2572\n", + "Train loss: 0.016119525264948607, Test loss: 0.01611863635480404\n", + "Starting epoch 2573\n", + "Train loss: 0.016011935472488404, Test loss: 0.014518782496452332\n", + "Starting epoch 2574\n", + "Train loss: 0.016055925376713277, Test loss: 0.017974574118852615\n", + "Starting epoch 2575\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.015932666473090648, Test loss: 0.01744188368320465\n", + "Starting epoch 2576\n", + "Train loss: 0.016062212381511926, Test loss: 0.017313368618488312\n", + "Starting epoch 2577\n", + "Train loss: 0.016045814640820028, Test loss: 0.014983715489506721\n", + "Starting epoch 2578\n", + "Train loss: 0.016035888828337192, Test loss: 0.014929172582924366\n", + "Starting epoch 2579\n", + "Train loss: 0.016020060535520315, Test loss: 0.01713506504893303\n", + "Starting epoch 2580\n", + "Train loss: 0.016011788882315157, Test loss: 0.014993658289313316\n", + "Starting epoch 2581\n", + "Train loss: 0.015975116603076458, Test loss: 0.018034186214208603\n", + "Starting epoch 2582\n", + "Train loss: 0.015960425529628994, Test loss: 0.0157959945499897\n", + "Starting epoch 2583\n", + "Train loss: 0.015889314375817774, Test loss: 0.015249952673912048\n", + "Starting epoch 2584\n", + "Train loss: 0.01588168229907751, Test loss: 0.015114249661564827\n", + "Starting epoch 2585\n", + "Train loss: 0.015961481798440216, Test loss: 0.016124101355671883\n", + "Starting epoch 2586\n", + "Train loss: 0.016028988398611544, Test loss: 0.015612330287694931\n", + "Starting epoch 2587\n", + "Train loss: 0.015945276711136103, Test loss: 0.016631554812192917\n", + "Starting epoch 2588\n", + "Train loss: 0.015927796680480243, Test loss: 0.014660746790468693\n", + "Starting epoch 2589\n", + "Train loss: 0.015772193036973477, Test loss: 0.015564418397843838\n", + "Starting epoch 2590\n", + "Train loss: 0.01599788598716259, Test loss: 0.0175197534263134\n", + "Starting epoch 2591\n", + "Train loss: 0.015949643217027187, Test loss: 0.015757273882627487\n", + "Starting epoch 2592\n", + "Train loss: 0.01576920526102185, Test loss: 0.02650774084031582\n", + "Starting epoch 2593\n", + "Train loss: 0.016042789109051228, Test loss: 0.031667310744524\n", + "Starting epoch 2594\n", + "Train loss: 0.016091924719512464, Test loss: 0.0328872986137867\n", + "Starting epoch 2595\n", + "Train loss: 0.015974122136831283, Test loss: 0.02957664243876934\n", + "Starting epoch 2596\n", + "Train loss: 0.015968435741961003, Test loss: 0.03300663083791733\n", + "Starting epoch 2597\n", + "Train loss: 0.015956590585410595, Test loss: 0.03216041997075081\n", + "Starting epoch 2598\n", + "Train loss: 0.015808444246649744, Test loss: 0.03322982415556908\n", + "Starting epoch 2599\n", + "Train loss: 0.015961514990776776, Test loss: 0.03296790271997452\n", + "Starting epoch 2600\n", + "Train loss: 0.015903421603143217, Test loss: 0.03388983756303787\n", + "Starting epoch 2601\n", + "Train loss: 0.015951088424772025, Test loss: 0.02903268113732338\n", + "Starting epoch 2602\n", + "Train loss: 0.015891458690166473, Test loss: 0.03006685897707939\n", + "Starting epoch 2603\n", + "Train loss: 0.015838387198746206, Test loss: 0.03022727556526661\n", + "Starting epoch 2604\n", + "Train loss: 0.015831810720264912, Test loss: 0.03376968950033188\n", + "Starting epoch 2605\n", + "Train loss: 0.0159291847422719, Test loss: 0.03090698830783367\n", + "Starting epoch 2606\n", + "Train loss: 0.015883493684232235, Test loss: 0.029837965965270996\n", + "Starting epoch 2607\n", + "Train loss: 0.015832792762666942, Test loss: 0.031584106385707855\n", + "Starting epoch 2608\n", + "Train loss: 0.015917796809226274, Test loss: 0.03046995960175991\n", + "Starting epoch 2609\n", + "Train loss: 0.015837250985205175, Test loss: 0.030676664784550667\n", + "Starting epoch 2610\n", + "Train loss: 0.01590714268386364, Test loss: 0.029057443141937256\n", + "Starting epoch 2611\n", + "Train loss: 0.015766969639807938, Test loss: 0.029482411220669746\n", + "Starting epoch 2612\n", + "Train loss: 0.015897455401718618, Test loss: 0.03159213438630104\n", + "Starting epoch 2613\n", + "Train loss: 0.015930991061031817, Test loss: 0.02912641502916813\n", + "Starting epoch 2614\n", + "Train loss: 0.01585747390985489, Test loss: 0.03218688443303108\n", + "Starting epoch 2615\n", + "Train loss: 0.015853541009128093, Test loss: 0.029694698750972748\n", + "Starting epoch 2616\n", + "Train loss: 0.015785462222993374, Test loss: 0.03151143342256546\n", + "Starting epoch 2617\n", + "Train loss: 0.015840141270309686, Test loss: 0.029192356392741203\n", + "Starting epoch 2618\n", + "Train loss: 0.015855370853096247, Test loss: 0.030280781909823418\n", + "Starting epoch 2619\n", + "Train loss: 0.015814754236489533, Test loss: 0.028462551534175873\n", + "Starting epoch 2620\n", + "Train loss: 0.01578030327335, Test loss: 0.030673468485474586\n", + "Starting epoch 2621\n", + "Train loss: 0.015739406514912842, Test loss: 0.027926044538617134\n", + "Starting epoch 2622\n", + "Train loss: 0.01585524410009384, Test loss: 0.029443690553307533\n", + "Starting epoch 2623\n", + "Train loss: 0.01582760153338313, Test loss: 0.033676765859127045\n", + "Starting epoch 2624\n", + "Train loss: 0.01583425676450133, Test loss: 0.024275805801153183\n", + "Starting epoch 2625\n", + "Train loss: 0.015862814635038375, Test loss: 0.0317421592772007\n", + "Starting epoch 2626\n", + "Train loss: 0.015866207741200924, Test loss: 0.031270790845155716\n", + "Starting epoch 2627\n", + "Train loss: 0.0158817932382226, Test loss: 0.02866990491747856\n", + "Starting epoch 2628\n", + "Train loss: 0.015787521116435527, Test loss: 0.032904017716646194\n", + "Starting epoch 2629\n", + "Train loss: 0.015816903244704007, Test loss: 0.03066999837756157\n", + "Starting epoch 2630\n", + "Train loss: 0.015792176835238935, Test loss: 0.0319385826587677\n", + "Starting epoch 2631\n", + "Train loss: 0.015766905415803195, Test loss: 0.027205919846892357\n", + "Starting epoch 2632\n", + "Train loss: 0.01578889401629567, Test loss: 0.030805224552750587\n", + "Starting epoch 2633\n", + "Train loss: 0.015758034884929658, Test loss: 0.026537198573350906\n", + "Starting epoch 2634\n", + "Train loss: 0.01575109429657459, Test loss: 0.0230755303055048\n", + "Starting epoch 2635\n", + "Train loss: 0.01560862198472023, Test loss: 0.030559157952666283\n", + "Starting epoch 2636\n", + "Train loss: 0.015814398396760226, Test loss: 0.03131302818655968\n", + "Starting epoch 2637\n", + "Train loss: 0.015701211541891098, Test loss: 0.02459096908569336\n", + "Starting epoch 2638\n", + "Train loss: 0.015694928765296937, Test loss: 0.029037736356258392\n", + "Starting epoch 2639\n", + "Train loss: 0.015712576266378165, Test loss: 0.02951461263000965\n", + "Starting epoch 2640\n", + "Train loss: 0.015800440795719625, Test loss: 0.02846258506178856\n", + "Starting epoch 2641\n", + "Train loss: 0.01576839653775096, Test loss: 0.0294707790017128\n", + "Starting epoch 2642\n", + "Train loss: 0.015767598561942577, Test loss: 0.029549524188041687\n", + "Starting epoch 2643\n", + "Train loss: 0.015719205718487502, Test loss: 0.029049867764115334\n", + "Starting epoch 2644\n", + "Train loss: 0.01577937938272953, Test loss: 0.03420671448111534\n", + "Starting epoch 2645\n", + "Train loss: 0.015787851959466935, Test loss: 0.029397545382380486\n", + "Starting epoch 2646\n", + "Train loss: 0.015683136731386185, Test loss: 0.030836086720228195\n", + "Starting epoch 2647\n", + "Train loss: 0.01572637878358364, Test loss: 0.02865782380104065\n", + "Starting epoch 2648\n", + "Train loss: 0.01570526769384742, Test loss: 0.03127902373671532\n", + "Starting epoch 2649\n", + "Train loss: 0.015688881259411574, Test loss: 0.016451610252261162\n", + "Starting epoch 2650\n", + "Train loss: 0.015771647337824105, Test loss: 0.014857359230518341\n", + "Starting epoch 2651\n", + "Train loss: 0.01573018614202738, Test loss: 0.01358591765165329\n", + "Starting epoch 2652\n", + "Train loss: 0.015745774880051612, Test loss: 0.01826644502580166\n", + "Starting epoch 2653\n", + "Train loss: 0.015594057906419038, Test loss: 0.025478839874267578\n", + "Starting epoch 2654\n", + "Train loss: 0.015752571132034064, Test loss: 0.02996203862130642\n", + "Starting epoch 2655\n", + "Train loss: 0.015805323589593172, Test loss: 0.023926453664898872\n", + "Starting epoch 2656\n", + "Train loss: 0.01570737099274993, Test loss: 0.027183815836906433\n", + "Starting epoch 2657\n", + "Train loss: 0.015797584298998118, Test loss: 0.030184898525476456\n", + "Starting epoch 2658\n", + "Train loss: 0.015730973836034536, Test loss: 0.03014479950070381\n", + "Starting epoch 2659\n", + "Train loss: 0.015576686691492796, Test loss: 0.02971581369638443\n", + "Starting epoch 2660\n", + "Train loss: 0.01564266553148627, Test loss: 0.02951866388320923\n", + "Starting epoch 2661\n", + "Train loss: 0.01566409572958946, Test loss: 0.03242087736725807\n", + "Starting epoch 2662\n", + "Train loss: 0.015686768982559444, Test loss: 0.0287898238748312\n", + "Starting epoch 2663\n", + "Train loss: 0.015690609384328126, Test loss: 0.028791211545467377\n", + "Starting epoch 2664\n", + "Train loss: 0.01563092289492488, Test loss: 0.02781740017235279\n", + "Starting epoch 2665\n", + "Train loss: 0.01566413728520274, Test loss: 0.026976723223924637\n", + "Starting epoch 2666\n", + "Train loss: 0.015631394181400537, Test loss: 0.030749700963497162\n", + "Starting epoch 2667\n", + "Train loss: 0.015577143151313066, Test loss: 0.02304920367896557\n", + "Starting epoch 2668\n", + "Train loss: 0.015619492009282112, Test loss: 0.02725670300424099\n", + "Starting epoch 2669\n", + "Train loss: 0.01565668446943164, Test loss: 0.03014192543923855\n", + "Starting epoch 2670\n", + "Train loss: 0.01563628040254116, Test loss: 0.03079843334853649\n", + "Starting epoch 2671\n", + "Train loss: 0.015667420327663422, Test loss: 0.024998251348733902\n", + "Starting epoch 2672\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.01557001968845725, Test loss: 0.03204454109072685\n", + "Starting epoch 2673\n", + "Train loss: 0.015523693207651377, Test loss: 0.02874395065009594\n", + "Starting epoch 2674\n", + "Train loss: 0.015602315869182348, Test loss: 0.03167818859219551\n", + "Starting epoch 2675\n", + "Train loss: 0.015637176148593425, Test loss: 0.03020796738564968\n", + "Starting epoch 2676\n", + "Train loss: 0.015570290237665176, Test loss: 0.024590734392404556\n", + "Starting epoch 2677\n", + "Train loss: 0.015586614068597555, Test loss: 0.01590234972536564\n", + "Starting epoch 2678\n", + "Train loss: 0.015739464815706015, Test loss: 0.015088872984051704\n", + "Starting epoch 2679\n", + "Train loss: 0.01563855869695544, Test loss: 0.014148076996207237\n", + "Starting epoch 2680\n", + "Train loss: 0.015577963646501302, Test loss: 0.01677052117884159\n", + "Starting epoch 2681\n", + "Train loss: 0.01560783164575696, Test loss: 0.014796608127653599\n", + "Starting epoch 2682\n", + "Train loss: 0.015668280478566884, Test loss: 0.016385754570364952\n", + "Starting epoch 2683\n", + "Train loss: 0.01561249703168869, Test loss: 0.014652550220489502\n", + "Starting epoch 2684\n", + "Train loss: 0.015522899013012648, Test loss: 0.01699058711528778\n", + "Starting epoch 2685\n", + "Train loss: 0.015565293077379465, Test loss: 0.029337797313928604\n", + "Starting epoch 2686\n", + "Train loss: 0.01560328332707286, Test loss: 0.028743484988808632\n", + "Starting epoch 2687\n", + "Train loss: 0.015524711534380913, Test loss: 0.029705505818128586\n", + "Starting epoch 2688\n", + "Train loss: 0.015537865832448006, Test loss: 0.03142194822430611\n", + "Starting epoch 2689\n", + "Train loss: 0.015574512351304293, Test loss: 0.028111284598708153\n", + "Starting epoch 2690\n", + "Train loss: 0.015551718808710575, Test loss: 0.031487397849559784\n", + "Starting epoch 2691\n", + "Train loss: 0.015530877150595188, Test loss: 0.033926717936992645\n", + "Starting epoch 2692\n", + "Train loss: 0.015609861686825752, Test loss: 0.030750401318073273\n", + "Starting epoch 2693\n", + "Train loss: 0.015607210993766784, Test loss: 0.02972208522260189\n", + "Starting epoch 2694\n", + "Train loss: 0.01556948009878397, Test loss: 0.029232148081064224\n", + "Starting epoch 2695\n", + "Train loss: 0.015514122620224953, Test loss: 0.03228231519460678\n", + "Starting epoch 2696\n", + "Train loss: 0.015552820302546025, Test loss: 0.02903815358877182\n", + "Starting epoch 2697\n", + "Train loss: 0.015583359654992818, Test loss: 0.02760697528719902\n", + "Starting epoch 2698\n", + "Train loss: 0.015421249624341726, Test loss: 0.032422300428152084\n", + "Starting epoch 2699\n", + "Train loss: 0.0154961128346622, Test loss: 0.028779765591025352\n", + "Starting epoch 2700\n", + "Train loss: 0.015480156764388084, Test loss: 0.026365458965301514\n", + "Starting epoch 2701\n", + "Train loss: 0.015499952752143144, Test loss: 0.030006008222699165\n", + "Starting epoch 2702\n", + "Train loss: 0.015543639957904816, Test loss: 0.029085852205753326\n", + "Starting epoch 2703\n", + "Train loss: 0.015550091732293367, Test loss: 0.02920464798808098\n", + "Starting epoch 2704\n", + "Train loss: 0.015487753748893739, Test loss: 0.03195355832576752\n", + "Starting epoch 2705\n", + "Train loss: 0.01557991597801447, Test loss: 0.029728930443525314\n", + "Starting epoch 2706\n", + "Train loss: 0.015505522843450309, Test loss: 0.024272672832012177\n", + "Starting epoch 2707\n", + "Train loss: 0.01557969095185399, Test loss: 0.03387816622853279\n", + "Starting epoch 2708\n", + "Train loss: 0.015471478067338467, Test loss: 0.032494738698005676\n", + "Starting epoch 2709\n", + "Train loss: 0.01550116591155529, Test loss: 0.028729042038321495\n", + "Starting epoch 2710\n", + "Train loss: 0.015486089866608382, Test loss: 0.030387308448553085\n", + "Starting epoch 2711\n", + "Train loss: 0.01548523360863328, Test loss: 0.03266342729330063\n", + "Starting epoch 2712\n", + "Train loss: 0.015502283703535796, Test loss: 0.026060672476887703\n", + "Starting epoch 2713\n", + "Train loss: 0.015520938113331795, Test loss: 0.03180834650993347\n", + "Starting epoch 2714\n", + "Train loss: 0.015499945525079965, Test loss: 0.027434619143605232\n", + "Starting epoch 2715\n", + "Train loss: 0.01543978538364172, Test loss: 0.030310548841953278\n", + "Starting epoch 2716\n", + "Train loss: 0.015489904060959816, Test loss: 0.026330837979912758\n", + "Starting epoch 2717\n", + "Train loss: 0.015410989932715892, Test loss: 0.029647914692759514\n", + "Starting epoch 2718\n", + "Train loss: 0.01549721771851182, Test loss: 0.028964538127183914\n", + "Starting epoch 2719\n", + "Train loss: 0.015573462657630443, Test loss: 0.030753981322050095\n", + "Starting epoch 2720\n", + "Train loss: 0.01543845696374774, Test loss: 0.029624875634908676\n", + "Starting epoch 2721\n", + "Train loss: 0.01545493582263589, Test loss: 0.02588457055389881\n", + "Starting epoch 2722\n", + "Train loss: 0.015442943852394819, Test loss: 0.029119864106178284\n", + "Starting epoch 2723\n", + "Train loss: 0.015455925781279802, Test loss: 0.02837682142853737\n", + "Starting epoch 2724\n", + "Train loss: 0.015406104251742363, Test loss: 0.03082358092069626\n", + "Starting epoch 2725\n", + "Train loss: 0.01545292679220438, Test loss: 0.034019190818071365\n", + "Starting epoch 2726\n", + "Train loss: 0.015502656158059836, Test loss: 0.033196382224559784\n", + "Starting epoch 2727\n", + "Train loss: 0.015512662306427956, Test loss: 0.02994934469461441\n", + "Starting epoch 2728\n", + "Train loss: 0.01541504006832838, Test loss: 0.030420297756791115\n", + "Starting epoch 2729\n", + "Train loss: 0.015380235631018877, Test loss: 0.02653295174241066\n", + "Starting epoch 2730\n", + "Train loss: 0.01546654812991619, Test loss: 0.02866983786225319\n", + "Starting epoch 2731\n", + "Train loss: 0.015317074246704579, Test loss: 0.027016866952180862\n", + "Starting epoch 2732\n", + "Train loss: 0.015462908241897822, Test loss: 0.025934729725122452\n", + "Starting epoch 2733\n", + "Train loss: 0.015467570964246988, Test loss: 0.026595132425427437\n", + "Starting epoch 2734\n", + "Train loss: 0.01542078960686922, Test loss: 0.0311738234013319\n", + "Starting epoch 2735\n", + "Train loss: 0.01543573385104537, Test loss: 0.029267987236380577\n", + "Starting epoch 2736\n", + "Train loss: 0.015370594169944525, Test loss: 0.036026015877723694\n", + "Starting epoch 2737\n", + "Train loss: 0.015459644198417664, Test loss: 0.031231220811605453\n", + "Starting epoch 2738\n", + "Train loss: 0.015320947039872407, Test loss: 0.03182373568415642\n", + "Starting epoch 2739\n", + "Train loss: 0.015430511478334664, Test loss: 0.03119020350277424\n", + "Starting epoch 2740\n", + "Train loss: 0.01537416098639369, Test loss: 0.014959143474698067\n", + "Starting epoch 2741\n", + "Train loss: 0.015434324406087399, Test loss: 0.016254328191280365\n", + "Starting epoch 2742\n", + "Train loss: 0.015452232658863068, Test loss: 0.014015162363648415\n", + "Starting epoch 2743\n", + "Train loss: 0.015362756811082363, Test loss: 0.015445750206708908\n", + "Starting epoch 2744\n", + "Train loss: 0.015332921911031007, Test loss: 0.015434164553880692\n", + "Starting epoch 2745\n", + "Train loss: 0.015394024085253477, Test loss: 0.015719473361968994\n", + "Starting epoch 2746\n", + "Train loss: 0.01524765795096755, Test loss: 0.014361020177602768\n", + "Starting epoch 2747\n", + "Train loss: 0.01705503188073635, Test loss: 0.021992597728967667\n", + "Starting epoch 2748\n", + "Train loss: 0.01590259123593569, Test loss: 0.032348040491342545\n", + "Starting epoch 2749\n", + "Train loss: 0.015607379637658595, Test loss: 0.030604105442762375\n", + "Starting epoch 2750\n", + "Train loss: 0.015550071578472853, Test loss: 0.029345912858843803\n", + "Starting epoch 2751\n", + "Train loss: 0.01548500582575798, Test loss: 0.031191209331154823\n", + "Starting epoch 2752\n", + "Train loss: 0.015445654150098562, Test loss: 0.02809712663292885\n", + "Starting epoch 2753\n", + "Train loss: 0.01546187249943614, Test loss: 0.027102511376142502\n", + "Starting epoch 2754\n", + "Train loss: 0.01539629627019167, Test loss: 0.02995363622903824\n", + "Starting epoch 2755\n", + "Train loss: 0.015355643220245838, Test loss: 0.03436955809593201\n", + "Starting epoch 2756\n", + "Train loss: 0.015423960089683532, Test loss: 0.02872013859450817\n", + "Starting epoch 2757\n", + "Train loss: 0.01526557508856058, Test loss: 0.03182633966207504\n", + "Starting epoch 2758\n", + "Train loss: 0.015271903052926063, Test loss: 0.032174885272979736\n", + "Starting epoch 2759\n", + "Train loss: 0.015260446518659591, Test loss: 0.02772880718111992\n", + "Starting epoch 2760\n", + "Train loss: 0.015311933550983668, Test loss: 0.01561993733048439\n", + "Starting epoch 2761\n", + "Train loss: 0.015357978064566851, Test loss: 0.01416845340281725\n", + "Starting epoch 2762\n", + "Train loss: 0.015247906371951104, Test loss: 0.016022497788071632\n", + "Starting epoch 2763\n", + "Train loss: 0.01534591929987073, Test loss: 0.015406650491058826\n", + "Starting epoch 2764\n", + "Train loss: 0.015316607411950827, Test loss: 0.01580275595188141\n", + "Starting epoch 2765\n", + "Train loss: 0.015210156980901956, Test loss: 0.015023545362055302\n", + "Starting epoch 2766\n", + "Train loss: 0.015306853335350751, Test loss: 0.015058564022183418\n", + "Starting epoch 2767\n", + "Train loss: 0.01523469902575016, Test loss: 0.018670951947569847\n", + "Starting epoch 2768\n", + "Train loss: 0.015277042742818594, Test loss: 0.0166181568056345\n", + "Starting epoch 2769\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.015266487691551447, Test loss: 0.015313318930566311\n", + "Starting epoch 2770\n", + "Train loss: 0.015303449593484402, Test loss: 0.015177513472735882\n", + "Starting epoch 2771\n", + "Train loss: 0.015303879026323557, Test loss: 0.014185450039803982\n", + "Starting epoch 2772\n", + "Train loss: 0.015245398618280887, Test loss: 0.016025133430957794\n", + "Starting epoch 2773\n", + "Train loss: 0.015216921642422675, Test loss: 0.014688904397189617\n", + "Starting epoch 2774\n", + "Train loss: 0.015165989845991134, Test loss: 0.01635805517435074\n", + "Starting epoch 2775\n", + "Train loss: 0.015166001524776221, Test loss: 0.01488802395761013\n", + "Starting epoch 2776\n", + "Train loss: 0.015288314763456583, Test loss: 0.017248008400201797\n", + "Starting epoch 2777\n", + "Train loss: 0.015262442789971829, Test loss: 0.01642431877553463\n", + "Starting epoch 2778\n", + "Train loss: 0.015236325562000275, Test loss: 0.013271735049784184\n", + "Starting epoch 2779\n", + "Train loss: 0.015190394185483456, Test loss: 0.013921303674578667\n", + "Starting epoch 2780\n", + "Train loss: 0.015264712050557136, Test loss: 0.014555622823536396\n", + "Starting epoch 2781\n", + "Train loss: 0.015249181725084781, Test loss: 0.018019821494817734\n", + "Starting epoch 2782\n", + "Train loss: 0.015168820526450872, Test loss: 0.020711377263069153\n", + "Starting epoch 2783\n", + "Train loss: 0.015207838211208581, Test loss: 0.019689470529556274\n", + "Starting epoch 2784\n", + "Train loss: 0.015128768626600503, Test loss: 0.02661828324198723\n", + "Starting epoch 2785\n", + "Train loss: 0.015332019347697496, Test loss: 0.028703829273581505\n", + "Starting epoch 2786\n", + "Train loss: 0.015263278130441904, Test loss: 0.029903249815106392\n", + "Starting epoch 2787\n", + "Train loss: 0.015172359105199575, Test loss: 0.023500366136431694\n", + "Starting epoch 2788\n", + "Train loss: 0.015089754033833742, Test loss: 0.03273063153028488\n", + "Starting epoch 2789\n", + "Train loss: 0.015121220648288726, Test loss: 0.028869520872831345\n", + "Starting epoch 2790\n", + "Train loss: 0.015302953254431486, Test loss: 0.015611061826348305\n", + "Starting epoch 2791\n", + "Train loss: 0.015217471197247505, Test loss: 0.013383511453866959\n", + "Starting epoch 2792\n", + "Train loss: 0.015258375518023969, Test loss: 0.016718529164791107\n", + "Starting epoch 2793\n", + "Train loss: 0.01526099158450961, Test loss: 0.01735180802643299\n", + "Starting epoch 2794\n", + "Train loss: 0.015283456426113845, Test loss: 0.019906679168343544\n", + "Starting epoch 2795\n", + "Train loss: 0.015182958859950304, Test loss: 0.015550403855741024\n", + "Starting epoch 2796\n", + "Train loss: 0.015287741273641586, Test loss: 0.014966794289648533\n", + "Starting epoch 2797\n", + "Train loss: 0.015160729326307774, Test loss: 0.018268266692757607\n", + "Starting epoch 2798\n", + "Train loss: 0.015200879499316216, Test loss: 0.015924843028187752\n", + "Starting epoch 2799\n", + "Train loss: 0.015270318929105997, Test loss: 0.016199970617890358\n", + "Starting epoch 2800\n", + "Train loss: 0.01508291371166706, Test loss: 0.014791660942137241\n", + "Starting epoch 2801\n", + "Train loss: 0.015223417021334172, Test loss: 0.015033924952149391\n", + "Starting epoch 2802\n", + "Train loss: 0.015153693109750748, Test loss: 0.016252426430583\n", + "Starting epoch 2803\n", + "Train loss: 0.01520112318918109, Test loss: 0.01569460891187191\n", + "Starting epoch 2804\n", + "Train loss: 0.015189525000751018, Test loss: 0.013930692337453365\n", + "Starting epoch 2805\n", + "Train loss: 0.01517780164256692, Test loss: 0.015229727141559124\n", + "Starting epoch 2806\n", + "Train loss: 0.016553527247160672, Test loss: 0.022848451510071754\n", + "Starting epoch 2807\n", + "Train loss: 0.019879570845514536, Test loss: 0.022330768406391144\n", + "Starting epoch 2808\n", + "Train loss: 0.015562643371522427, Test loss: 0.014782417565584183\n", + "Starting epoch 2809\n", + "Train loss: 0.01551545675843954, Test loss: 0.01372476201504469\n", + "Starting epoch 2810\n", + "Train loss: 0.015624057129025459, Test loss: 0.015537923201918602\n", + "Starting epoch 2811\n", + "Train loss: 0.015497419219464063, Test loss: 0.01410609483718872\n", + "Starting epoch 2812\n", + "Train loss: 0.015387654043734074, Test loss: 0.01483621634542942\n", + "Starting epoch 2813\n", + "Train loss: 0.01526026276871562, Test loss: 0.01409743819385767\n", + "Starting epoch 2814\n", + "Train loss: 0.015281537845730782, Test loss: 0.016131313517689705\n", + "Starting epoch 2815\n", + "Train loss: 0.015207131039351224, Test loss: 0.015047240070998669\n", + "Starting epoch 2816\n", + "Train loss: 0.015143916793167591, Test loss: 0.014974530786275864\n", + "Starting epoch 2817\n", + "Train loss: 0.015163634456694126, Test loss: 0.015926742926239967\n", + "Starting epoch 2818\n", + "Train loss: 0.015148378647863866, Test loss: 0.020418915897607803\n", + "Starting epoch 2819\n", + "Train loss: 0.015099497362971306, Test loss: 0.028548985719680786\n", + "Starting epoch 2820\n", + "Train loss: 0.015046982616186142, Test loss: 0.029485050588846207\n", + "Starting epoch 2821\n", + "Train loss: 0.015113854650408029, Test loss: 0.0161493718624115\n", + "Starting epoch 2822\n", + "Train loss: 0.015148465018719436, Test loss: 0.014119971543550491\n", + "Starting epoch 2823\n", + "Train loss: 0.015040066242218018, Test loss: 0.0331595316529274\n", + "Starting epoch 2824\n", + "Train loss: 0.015145476125180721, Test loss: 0.02512289397418499\n", + "Starting epoch 2825\n", + "Train loss: 0.015095947869122028, Test loss: 0.024948155507445335\n", + "Starting epoch 2826\n", + "Train loss: 0.014949930906295776, Test loss: 0.029170118272304535\n", + "Starting epoch 2827\n", + "Train loss: 0.015150444358587265, Test loss: 0.03155924007296562\n", + "Starting epoch 2828\n", + "Train loss: 0.015046336762607098, Test loss: 0.02330285683274269\n", + "Starting epoch 2829\n", + "Train loss: 0.015070206522941589, Test loss: 0.029892580583691597\n", + "Starting epoch 2830\n", + "Train loss: 0.015029165092855692, Test loss: 0.030133303254842758\n", + "Starting epoch 2831\n", + "Train loss: 0.015034537874162196, Test loss: 0.029470711946487427\n", + "Starting epoch 2832\n", + "Train loss: 0.015052613709121943, Test loss: 0.022447561845183372\n", + "Starting epoch 2833\n", + "Train loss: 0.01491339085623622, Test loss: 0.024394012987613678\n", + "Starting epoch 2834\n", + "Train loss: 0.015126688536256552, Test loss: 0.027436697855591774\n", + "Starting epoch 2835\n", + "Train loss: 0.015133697036653758, Test loss: 0.031014516949653625\n", + "Starting epoch 2836\n", + "Train loss: 0.015063189957290888, Test loss: 0.03005124069750309\n", + "Starting epoch 2837\n", + "Train loss: 0.015080959964543582, Test loss: 0.020546164363622665\n", + "Starting epoch 2838\n", + "Train loss: 0.015116390474140644, Test loss: 0.030281009152531624\n", + "Starting epoch 2839\n", + "Train loss: 0.014994713105261325, Test loss: 0.019209086894989014\n", + "Starting epoch 2840\n", + "Train loss: 0.015076409094035625, Test loss: 0.013046995736658573\n", + "Starting epoch 2841\n", + "Train loss: 0.014889087490737439, Test loss: 0.013074385933578014\n", + "Starting epoch 2842\n", + "Train loss: 0.01506100369617343, Test loss: 0.014741573482751846\n", + "Starting epoch 2843\n", + "Train loss: 0.015023107882589102, Test loss: 0.014524488709867\n", + "Starting epoch 2844\n", + "Train loss: 0.015058436375111341, Test loss: 0.014974531717598438\n", + "Starting epoch 2845\n", + "Train loss: 0.01501509740948677, Test loss: 0.014307512901723385\n", + "Starting epoch 2846\n", + "Train loss: 0.014855567049235106, Test loss: 0.015760943293571472\n", + "Starting epoch 2847\n", + "Train loss: 0.01501277718693018, Test loss: 0.01465813908725977\n", + "Starting epoch 2848\n", + "Train loss: 0.014787993859499693, Test loss: 0.01505599357187748\n", + "Starting epoch 2849\n", + "Train loss: 0.01499851293861866, Test loss: 0.014244630932807922\n", + "Starting epoch 2850\n", + "Train loss: 0.014982087016105651, Test loss: 0.01664081960916519\n", + "Starting epoch 2851\n", + "Train loss: 0.014948948230594396, Test loss: 0.01816719025373459\n", + "Starting epoch 2852\n", + "Train loss: 0.014999424386769534, Test loss: 0.015171530656516552\n", + "Starting epoch 2853\n", + "Train loss: 0.015004136878997088, Test loss: 0.015372750349342823\n", + "Starting epoch 2854\n", + "Train loss: 0.015093174371868372, Test loss: 0.014101045206189156\n", + "Starting epoch 2855\n", + "Train loss: 0.015015698112547398, Test loss: 0.015043887309730053\n", + "Starting epoch 2856\n", + "Train loss: 0.014935793858021497, Test loss: 0.015655923634767532\n", + "Starting epoch 2857\n", + "Train loss: 0.014863375071436168, Test loss: 0.015507983043789864\n", + "Starting epoch 2858\n", + "Train loss: 0.015033759865909815, Test loss: 0.013635768555104733\n", + "Starting epoch 2859\n", + "Train loss: 0.014963833130896092, Test loss: 0.014307226985692978\n", + "Starting epoch 2860\n", + "Train loss: 0.014916985407471656, Test loss: 0.015389377251267433\n", + "Starting epoch 2861\n", + "Train loss: 0.014879768695682288, Test loss: 0.01413489319384098\n", + "Starting epoch 2862\n", + "Train loss: 0.014959851466119289, Test loss: 0.014078006148338318\n", + "Starting epoch 2863\n", + "Train loss: 0.014957004319876432, Test loss: 0.015555819496512413\n", + "Starting epoch 2864\n", + "Train loss: 0.014921694602817297, Test loss: 0.014552666805684566\n", + "Starting epoch 2865\n", + "Train loss: 0.015042711850255727, Test loss: 0.01371670700609684\n", + "Starting epoch 2866\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.014996828995645046, Test loss: 0.016041545197367668\n", + "Starting epoch 2867\n", + "Train loss: 0.014905556160956622, Test loss: 0.01380607858300209\n", + "Starting epoch 2868\n", + "Train loss: 0.01503941010683775, Test loss: 0.018145008012652397\n", + "Starting epoch 2869\n", + "Train loss: 0.014917822647839784, Test loss: 0.013986795209348202\n", + "Starting epoch 2870\n", + "Train loss: 0.014908953253179789, Test loss: 0.014368578791618347\n", + "Starting epoch 2871\n", + "Train loss: 0.014963247943669558, Test loss: 0.014410000294446945\n", + "Starting epoch 2872\n", + "Train loss: 0.01492367547005415, Test loss: 0.014274852350354195\n", + "Starting epoch 2873\n", + "Train loss: 0.014897151105105876, Test loss: 0.01793072558939457\n", + "Starting epoch 2874\n", + "Train loss: 0.01493529723957181, Test loss: 0.01711324416100979\n", + "Starting epoch 2875\n", + "Train loss: 0.01488083427771926, Test loss: 0.01458062045276165\n", + "Starting epoch 2876\n", + "Train loss: 0.014860849343240261, Test loss: 0.016256187111139297\n", + "Starting epoch 2877\n", + "Train loss: 0.014874163344502448, Test loss: 0.016327369958162308\n", + "Starting epoch 2878\n", + "Train loss: 0.01492365898564458, Test loss: 0.013781869783997536\n", + "Starting epoch 2879\n", + "Train loss: 0.01491054005920887, Test loss: 0.012362857349216938\n", + "Starting epoch 2880\n", + "Train loss: 0.01487388253211975, Test loss: 0.014306260272860527\n", + "Starting epoch 2881\n", + "Train loss: 0.014775609634816646, Test loss: 0.01307334192097187\n", + "Starting epoch 2882\n", + "Train loss: 0.014951393399387598, Test loss: 0.01323850080370903\n", + "Starting epoch 2883\n", + "Train loss: 0.014908408802002669, Test loss: 0.013019540347158909\n", + "Starting epoch 2884\n", + "Train loss: 0.01486523350700736, Test loss: 0.014465597458183765\n", + "Starting epoch 2885\n", + "Train loss: 0.014762312434613704, Test loss: 0.015849675983190536\n", + "Starting epoch 2886\n", + "Train loss: 0.01483978034928441, Test loss: 0.014335745014250278\n", + "Starting epoch 2887\n", + "Train loss: 0.014854609761387109, Test loss: 0.02830066904425621\n", + "Starting epoch 2888\n", + "Train loss: 0.014883663672953844, Test loss: 0.02593962661921978\n", + "Starting epoch 2889\n", + "Train loss: 0.014949680827558041, Test loss: 0.026820108294487\n", + "Starting epoch 2890\n", + "Train loss: 0.014888646565377712, Test loss: 0.027384549379348755\n", + "Starting epoch 2891\n", + "Train loss: 0.014852968119084835, Test loss: 0.029022295027971268\n", + "Starting epoch 2892\n", + "Train loss: 0.014849596247076989, Test loss: 0.025893250480294228\n", + "Starting epoch 2893\n", + "Train loss: 0.014793531876057387, Test loss: 0.03262721002101898\n", + "Starting epoch 2894\n", + "Train loss: 0.014874533340334893, Test loss: 0.03313048183917999\n", + "Starting epoch 2895\n", + "Train loss: 0.014896286912262439, Test loss: 0.031643763184547424\n", + "Starting epoch 2896\n", + "Train loss: 0.014833942670375109, Test loss: 0.031347181648015976\n", + "Starting epoch 2897\n", + "Train loss: 0.014767552334815264, Test loss: 0.013033048249781132\n", + "Starting epoch 2898\n", + "Train loss: 0.01496657621115446, Test loss: 0.016350867226719856\n", + "Starting epoch 2899\n", + "Train loss: 0.014909319747239351, Test loss: 0.015739666298031807\n", + "Starting epoch 2900\n", + "Train loss: 0.01491167513653636, Test loss: 0.013728092424571514\n", + "Starting epoch 2901\n", + "Train loss: 0.01483504032716155, Test loss: 0.016397183761000633\n", + "Starting epoch 2902\n", + "Train loss: 0.014844880234450101, Test loss: 0.014488634653389454\n", + "Starting epoch 2903\n", + "Train loss: 0.014813776258379222, Test loss: 0.015715358778834343\n", + "Starting epoch 2904\n", + "Train loss: 0.014860446620732546, Test loss: 0.014806486666202545\n", + "Starting epoch 2905\n", + "Train loss: 0.01478854225948453, Test loss: 0.01336619071662426\n", + "Starting epoch 2906\n", + "Train loss: 0.014870366454124451, Test loss: 0.01446318719536066\n", + "Starting epoch 2907\n", + "Train loss: 0.014833022262901068, Test loss: 0.014983654953539371\n", + "Starting epoch 2908\n", + "Train loss: 0.014730278886854648, Test loss: 0.013126351870596409\n", + "Starting epoch 2909\n", + "Train loss: 0.01487717917189002, Test loss: 0.013518060557544231\n", + "Starting epoch 2910\n", + "Train loss: 0.014912682045251131, Test loss: 0.014420131221413612\n", + "Starting epoch 2911\n", + "Train loss: 0.014877607524394989, Test loss: 0.014694939367473125\n", + "Starting epoch 2912\n", + "Train loss: 0.014815269578248262, Test loss: 0.01642794907093048\n", + "Starting epoch 2913\n", + "Train loss: 0.014609676990658045, Test loss: 0.012356270104646683\n", + "Starting epoch 2914\n", + "Train loss: 0.01488989107310772, Test loss: 0.013528715819120407\n", + "Starting epoch 2915\n", + "Train loss: 0.014945539329200983, Test loss: 0.014456606470048428\n", + "Starting epoch 2916\n", + "Train loss: 0.014833978414535522, Test loss: 0.014833508990705013\n", + "Starting epoch 2917\n", + "Train loss: 0.014875548426061869, Test loss: 0.01940299943089485\n", + "Starting epoch 2918\n", + "Train loss: 0.014809992797672748, Test loss: 0.017826953902840614\n", + "Starting epoch 2919\n", + "Train loss: 0.01489218333736062, Test loss: 0.013526290655136108\n", + "Starting epoch 2920\n", + "Train loss: 0.014845853094011546, Test loss: 0.014680953696370125\n", + "Starting epoch 2921\n", + "Train loss: 0.014886046331375836, Test loss: 0.012920808047056198\n", + "Starting epoch 2922\n", + "Train loss: 0.014804461803287268, Test loss: 0.0157773494720459\n", + "Starting epoch 2923\n", + "Train loss: 0.014793483186513185, Test loss: 0.01379252877086401\n", + "Starting epoch 2924\n", + "Train loss: 0.014834174234420062, Test loss: 0.013558641076087952\n", + "Starting epoch 2925\n", + "Train loss: 0.01460583820939064, Test loss: 0.026367181912064552\n", + "Starting epoch 2926\n", + "Train loss: 0.014810584355145692, Test loss: 0.02901276759803295\n", + "Starting epoch 2927\n", + "Train loss: 0.014851567316800356, Test loss: 0.028818877413868904\n", + "Starting epoch 2928\n", + "Train loss: 0.014708289112895727, Test loss: 0.02843385376036167\n", + "Starting epoch 2929\n", + "Train loss: 0.014836620967835189, Test loss: 0.02702106349170208\n", + "Starting epoch 2930\n", + "Train loss: 0.014737154431641102, Test loss: 0.024426093325018883\n", + "Starting epoch 2931\n", + "Train loss: 0.014815589152276517, Test loss: 0.02758055366575718\n", + "Starting epoch 2932\n", + "Train loss: 0.01468813732266426, Test loss: 0.027623319998383522\n", + "Starting epoch 2933\n", + "Train loss: 0.014844866208732129, Test loss: 0.024691792204976082\n", + "Starting epoch 2934\n", + "Train loss: 0.014734803698956966, Test loss: 0.02863669954240322\n", + "Starting epoch 2935\n", + "Train loss: 0.014744747653603554, Test loss: 0.029387205839157104\n", + "Starting epoch 2936\n", + "Train loss: 0.014800121672451497, Test loss: 0.029242495074868202\n", + "Starting epoch 2937\n", + "Train loss: 0.014653530679643154, Test loss: 0.025835148990154266\n", + "Starting epoch 2938\n", + "Train loss: 0.014777030013501644, Test loss: 0.026525074616074562\n", + "Starting epoch 2939\n", + "Train loss: 0.014784085359424352, Test loss: 0.029862061142921448\n", + "Starting epoch 2940\n", + "Train loss: 0.014683138839900493, Test loss: 0.029597029089927673\n", + "Starting epoch 2941\n", + "Train loss: 0.014676430504769087, Test loss: 0.026247963309288025\n", + "Starting epoch 2942\n", + "Train loss: 0.014658209271728992, Test loss: 0.026656555011868477\n", + "Starting epoch 2943\n", + "Train loss: 0.014661463145166635, Test loss: 0.02634064108133316\n", + "Starting epoch 2944\n", + "Train loss: 0.014765007384121419, Test loss: 0.025493988767266273\n", + "Starting epoch 2945\n", + "Train loss: 0.01469658687710762, Test loss: 0.032317619770765305\n", + "Starting epoch 2946\n", + "Train loss: 0.014710325580090285, Test loss: 0.030481047928333282\n", + "Starting epoch 2947\n", + "Train loss: 0.0148025887273252, Test loss: 0.024277111515402794\n", + "Starting epoch 2948\n", + "Train loss: 0.014664171673357486, Test loss: 0.014826294034719467\n", + "Starting epoch 2949\n", + "Train loss: 0.014773443583399057, Test loss: 0.014208993874490261\n", + "Starting epoch 2950\n", + "Train loss: 0.01474290644749999, Test loss: 0.01632341369986534\n", + "Starting epoch 2951\n", + "Train loss: 0.014667523913085461, Test loss: 0.01617169938981533\n", + "Starting epoch 2952\n", + "Train loss: 0.01464485701173544, Test loss: 0.014896850101649761\n", + "Starting epoch 2953\n", + "Train loss: 0.01466091088950634, Test loss: 0.015319217927753925\n", + "Starting epoch 2954\n", + "Train loss: 0.015058073680847883, Test loss: 0.02677992731332779\n", + "Starting epoch 2955\n", + "Train loss: 0.01477958431467414, Test loss: 0.02640371583402157\n", + "Starting epoch 2956\n", + "Train loss: 0.01472501253709197, Test loss: 0.029035916551947594\n", + "Starting epoch 2957\n", + "Train loss: 0.014712190944701433, Test loss: 0.02195051498711109\n", + "Starting epoch 2958\n", + "Train loss: 0.014614984840154648, Test loss: 0.029264409095048904\n", + "Starting epoch 2959\n", + "Train loss: 0.014748459961265326, Test loss: 0.02882285602390766\n", + "Starting epoch 2960\n", + "Train loss: 0.014712852109223604, Test loss: 0.028023406863212585\n", + "Starting epoch 2961\n", + "Train loss: 0.014692907407879829, Test loss: 0.028165249153971672\n", + "Starting epoch 2962\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.014606118369847536, Test loss: 0.02871687151491642\n", + "Starting epoch 2963\n", + "Train loss: 0.014599507320672273, Test loss: 0.02349700778722763\n", + "Starting epoch 2964\n", + "Train loss: 0.014602104183286429, Test loss: 0.02698564901947975\n", + "Starting epoch 2965\n", + "Train loss: 0.014580845460295677, Test loss: 0.02546817809343338\n", + "Starting epoch 2966\n", + "Train loss: 0.014724387507885694, Test loss: 0.026396315544843674\n", + "Starting epoch 2967\n", + "Train loss: 0.014612097889184952, Test loss: 0.024358095601201057\n", + "Starting epoch 2968\n", + "Train loss: 0.014570850133895873, Test loss: 0.0319075770676136\n", + "Starting epoch 2969\n", + "Train loss: 0.014681064020842315, Test loss: 0.024787260219454765\n", + "Starting epoch 2970\n", + "Train loss: 0.014565219059586526, Test loss: 0.027802415192127228\n", + "Starting epoch 2971\n", + "Train loss: 0.014461825974285603, Test loss: 0.022547537460923195\n", + "Starting epoch 2972\n", + "Train loss: 0.014478531554341317, Test loss: 0.03115655854344368\n", + "Starting epoch 2973\n", + "Train loss: 0.014737046472728253, Test loss: 0.02601947821676731\n", + "Starting epoch 2974\n", + "Train loss: 0.014640442170202732, Test loss: 0.028014613315463066\n", + "Starting epoch 2975\n", + "Train loss: 0.014630578979849816, Test loss: 0.023598164319992065\n", + "Starting epoch 2976\n", + "Train loss: 0.014644252806901932, Test loss: 0.02857487089931965\n", + "Starting epoch 2977\n", + "Train loss: 0.014655782841145992, Test loss: 0.02833777666091919\n", + "Starting epoch 2978\n", + "Train loss: 0.014660765323787928, Test loss: 0.025312194600701332\n", + "Starting epoch 2979\n", + "Train loss: 0.014668399915099144, Test loss: 0.02757851965725422\n", + "Starting epoch 2980\n", + "Train loss: 0.01464099993929267, Test loss: 0.021983031183481216\n", + "Starting epoch 2981\n", + "Train loss: 0.014596951752901077, Test loss: 0.021145718172192574\n", + "Starting epoch 2982\n", + "Train loss: 0.014521996732801199, Test loss: 0.02861613966524601\n", + "Starting epoch 2983\n", + "Train loss: 0.01461118269711733, Test loss: 0.024940138682723045\n", + "Starting epoch 2984\n", + "Train loss: 0.014573043473064899, Test loss: 0.027737751603126526\n", + "Starting epoch 2985\n", + "Train loss: 0.014535785857588053, Test loss: 0.026317983865737915\n", + "Starting epoch 2986\n", + "Train loss: 0.014566133264452219, Test loss: 0.027563979849219322\n", + "Starting epoch 2987\n", + "Train loss: 0.01453916499391198, Test loss: 0.030426504090428352\n", + "Starting epoch 2988\n", + "Train loss: 0.014570909570902586, Test loss: 0.015096360817551613\n", + "Starting epoch 2989\n", + "Train loss: 0.014610631894320249, Test loss: 0.014074274338781834\n", + "Starting epoch 2990\n", + "Train loss: 0.014649116192013025, Test loss: 0.013807822950184345\n", + "Starting epoch 2991\n", + "Train loss: 0.014541861470788717, Test loss: 0.02382161095738411\n", + "Starting epoch 2992\n", + "Train loss: 0.014624746255576611, Test loss: 0.02880083955824375\n", + "Starting epoch 2993\n", + "Train loss: 0.014600089862942695, Test loss: 0.026364604011178017\n", + "Starting epoch 2994\n", + "Train loss: 0.01457083886489272, Test loss: 0.029138054698705673\n", + "Starting epoch 2995\n", + "Train loss: 0.014481494762003421, Test loss: 0.025183649733662605\n", + "Starting epoch 2996\n", + "Train loss: 0.01455381965264678, Test loss: 0.014261335134506226\n", + "Starting epoch 2997\n", + "Train loss: 0.014596231207251549, Test loss: 0.013802808709442616\n", + "Starting epoch 2998\n", + "Train loss: 0.014615332391113043, Test loss: 0.014458549208939075\n", + "Starting epoch 2999\n", + "Train loss: 0.014558436460793018, Test loss: 0.0150105906650424\n", + "Starting epoch 3000\n", + "Train loss: 0.014530067574232816, Test loss: 0.015008045360445976\n", + "Starting epoch 3001\n", + "Train loss: 0.014591627251356841, Test loss: 0.01370159536600113\n", + "Starting epoch 3002\n", + "Train loss: 0.014516795463860034, Test loss: 0.020185988396406174\n", + "Starting epoch 3003\n", + "Train loss: 0.014451842326670884, Test loss: 0.030071966350078583\n", + "Starting epoch 3004\n", + "Train loss: 0.01461637057363987, Test loss: 0.029370730742812157\n", + "Starting epoch 3005\n", + "Train loss: 0.014609482251107692, Test loss: 0.030869899317622185\n", + "Starting epoch 3006\n", + "Train loss: 0.014573505651205777, Test loss: 0.024389861151576042\n", + "Starting epoch 3007\n", + "Train loss: 0.014530595485121011, Test loss: 0.02622910961508751\n", + "Starting epoch 3008\n", + "Train loss: 0.01441576711833477, Test loss: 0.027950435876846313\n", + "Starting epoch 3009\n", + "Train loss: 0.014439597018063069, Test loss: 0.027209434658288956\n", + "Starting epoch 3010\n", + "Train loss: 0.01440979603677988, Test loss: 0.028086943551898003\n", + "Starting epoch 3011\n", + "Train loss: 0.014590189903974532, Test loss: 0.012285739183425903\n", + "Starting epoch 3012\n", + "Train loss: 0.014578135535120964, Test loss: 0.012973426841199398\n", + "Starting epoch 3013\n", + "Train loss: 0.014553614910691976, Test loss: 0.01298707164824009\n", + "Starting epoch 3014\n", + "Train loss: 0.014568014480173587, Test loss: 0.016518568620085716\n", + "Starting epoch 3015\n", + "Train loss: 0.01451914394274354, Test loss: 0.013540741056203842\n", + "Starting epoch 3016\n", + "Train loss: 0.0145459857955575, Test loss: 0.013337058946490288\n", + "Starting epoch 3017\n", + "Train loss: 0.014593631830066442, Test loss: 0.017559507861733437\n", + "Starting epoch 3018\n", + "Train loss: 0.014507570695132017, Test loss: 0.014612407423555851\n", + "Starting epoch 3019\n", + "Train loss: 0.014454368390142918, Test loss: 0.01503801066428423\n", + "Starting epoch 3020\n", + "Train loss: 0.014424293898046017, Test loss: 0.01532213855534792\n", + "Starting epoch 3021\n", + "Train loss: 0.014507219512015581, Test loss: 0.014993366785347462\n", + "Starting epoch 3022\n", + "Train loss: 0.01451191257685423, Test loss: 0.015173538587987423\n", + "Starting epoch 3023\n", + "Train loss: 0.014475921615958214, Test loss: 0.0175315011292696\n", + "Starting epoch 3024\n", + "Train loss: 0.014472791273146868, Test loss: 0.011894418857991695\n", + "Starting epoch 3025\n", + "Train loss: 0.014449079595506191, Test loss: 0.013187609612941742\n", + "Starting epoch 3026\n", + "Train loss: 0.01444110132753849, Test loss: 0.01643507368862629\n", + "Starting epoch 3027\n", + "Train loss: 0.014476468544453383, Test loss: 0.012247877195477486\n", + "Starting epoch 3028\n", + "Train loss: 0.014430246260017156, Test loss: 0.015472015365958214\n", + "Starting epoch 3029\n", + "Train loss: 0.014382831361144781, Test loss: 0.013266757130622864\n", + "Starting epoch 3030\n", + "Train loss: 0.014498477727174759, Test loss: 0.02972021885216236\n", + "Starting epoch 3031\n", + "Train loss: 0.014505880475044251, Test loss: 0.029903754591941833\n", + "Starting epoch 3032\n", + "Train loss: 0.014496300499886275, Test loss: 0.022705182433128357\n", + "Starting epoch 3033\n", + "Train loss: 0.015072397086769343, Test loss: 0.02807062305510044\n", + "Starting epoch 3034\n", + "Train loss: 0.014590909797698259, Test loss: 0.027593694627285004\n", + "Starting epoch 3035\n", + "Train loss: 0.014465633668005467, Test loss: 0.02486257441341877\n", + "Starting epoch 3036\n", + "Train loss: 0.014458449557423592, Test loss: 0.030467554926872253\n", + "Starting epoch 3037\n", + "Train loss: 0.014447867069393397, Test loss: 0.02853746898472309\n", + "Starting epoch 3038\n", + "Train loss: 0.0144545578956604, Test loss: 0.030027085915207863\n", + "Starting epoch 3039\n", + "Train loss: 0.014392269812524318, Test loss: 0.02222713641822338\n", + "Starting epoch 3040\n", + "Train loss: 0.014322000704705715, Test loss: 0.028697317466139793\n", + "Starting epoch 3041\n", + "Train loss: 0.014450272172689438, Test loss: 0.02515641413629055\n", + "Starting epoch 3042\n", + "Train loss: 0.014451003670692443, Test loss: 0.0303302314132452\n", + "Starting epoch 3043\n", + "Train loss: 0.014443703908473254, Test loss: 0.03196043521165848\n", + "Starting epoch 3044\n", + "Train loss: 0.014408446904271842, Test loss: 0.026237016543745995\n", + "Starting epoch 3045\n", + "Train loss: 0.014392486624419689, Test loss: 0.023250488564372063\n", + "Starting epoch 3046\n", + "Train loss: 0.014390652608126402, Test loss: 0.025207990780472755\n", + "Starting epoch 3047\n", + "Train loss: 0.014525320455431938, Test loss: 0.025323567911982536\n", + "Starting epoch 3048\n", + "Train loss: 0.01449846975505352, Test loss: 0.02505224384367466\n", + "Starting epoch 3049\n", + "Train loss: 0.014448099657893182, Test loss: 0.028750423341989517\n", + "Starting epoch 3050\n", + "Train loss: 0.014451598729938269, Test loss: 0.027538876980543137\n", + "Starting epoch 3051\n", + "Train loss: 0.014423218443989754, Test loss: 0.024226335808634758\n", + "Starting epoch 3052\n", + "Train loss: 0.014297743588685989, Test loss: 0.01613941602408886\n", + "Starting epoch 3053\n", + "Train loss: 0.014486003033816815, Test loss: 0.030416881665587425\n", + "Starting epoch 3054\n", + "Train loss: 0.014478470496833325, Test loss: 0.025613468140363693\n", + "Starting epoch 3055\n", + "Train loss: 0.014409994091838598, Test loss: 0.026436492800712585\n", + "Starting epoch 3056\n", + "Train loss: 0.014384722858667374, Test loss: 0.02976321429014206\n", + "Starting epoch 3057\n", + "Train loss: 0.014375031478703022, Test loss: 0.025727691128849983\n", + "Starting epoch 3058\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.014419965464621782, Test loss: 0.026400569826364517\n", + "Starting epoch 3059\n", + "Train loss: 0.014354635998606681, Test loss: 0.013601495884358883\n", + "Starting epoch 3060\n", + "Train loss: 0.01439631164073944, Test loss: 0.012820511125028133\n", + "Starting epoch 3061\n", + "Train loss: 0.014332171361893416, Test loss: 0.02845992147922516\n", + "Starting epoch 3062\n", + "Train loss: 0.014404730889946222, Test loss: 0.0265429075807333\n", + "Starting epoch 3063\n", + "Train loss: 0.014376378655433654, Test loss: 0.02751556970179081\n", + "Starting epoch 3064\n", + "Train loss: 0.014373681284487247, Test loss: 0.02611720934510231\n", + "Starting epoch 3065\n", + "Train loss: 0.014383005369454622, Test loss: 0.022101327776908875\n", + "Starting epoch 3066\n", + "Train loss: 0.014340063631534576, Test loss: 0.013525236397981644\n", + "Starting epoch 3067\n", + "Train loss: 0.014325378779321909, Test loss: 0.013219980522990227\n", + "Starting epoch 3068\n", + "Train loss: 0.014323367010802031, Test loss: 0.014649268239736557\n", + "Starting epoch 3069\n", + "Train loss: 0.01437052546069026, Test loss: 0.014220922254025936\n", + "Starting epoch 3070\n", + "Train loss: 0.014328804593533277, Test loss: 0.012166034430265427\n", + "Starting epoch 3071\n", + "Train loss: 0.01430215997621417, Test loss: 0.02919306606054306\n", + "Starting epoch 3072\n", + "Train loss: 0.01438544012606144, Test loss: 0.027824178338050842\n", + "Starting epoch 3073\n", + "Train loss: 0.014342784658074379, Test loss: 0.030573152005672455\n", + "Starting epoch 3074\n", + "Train loss: 0.014405735041946173, Test loss: 0.02890438213944435\n", + "Starting epoch 3075\n", + "Train loss: 0.014376081675291062, Test loss: 0.02602526545524597\n", + "Starting epoch 3076\n", + "Train loss: 0.014262564312666655, Test loss: 0.024998802691698074\n", + "Starting epoch 3077\n", + "Train loss: 0.014353551845997571, Test loss: 0.02483132667839527\n", + "Starting epoch 3078\n", + "Train loss: 0.014280345812439919, Test loss: 0.027289820834994316\n", + "Starting epoch 3079\n", + "Train loss: 0.014229134395718575, Test loss: 0.024120423942804337\n", + "Starting epoch 3080\n", + "Train loss: 0.014348563812673092, Test loss: 0.032814864069223404\n", + "Starting epoch 3081\n", + "Train loss: 0.0144288956746459, Test loss: 0.02944287285208702\n", + "Starting epoch 3082\n", + "Train loss: 0.014412051066756248, Test loss: 0.027495259419083595\n", + "Starting epoch 3083\n", + "Train loss: 0.014332557041198015, Test loss: 0.028913071379065514\n", + "Starting epoch 3084\n", + "Train loss: 0.014342379793524741, Test loss: 0.026062136515975\n", + "Starting epoch 3085\n", + "Train loss: 0.014376489855349063, Test loss: 0.024661866948008537\n", + "Starting epoch 3086\n", + "Train loss: 0.014344999343156814, Test loss: 0.028307590633630753\n", + "Starting epoch 3087\n", + "Train loss: 0.014360174965113402, Test loss: 0.026680713519454002\n", + "Starting epoch 3088\n", + "Train loss: 0.014294057786464691, Test loss: 0.024530241265892982\n", + "Starting epoch 3089\n", + "Train loss: 0.014326786007732153, Test loss: 0.026709843426942825\n", + "Starting epoch 3090\n", + "Train loss: 0.014318118616938592, Test loss: 0.02540300413966179\n", + "Starting epoch 3091\n", + "Train loss: 0.0142541323043406, Test loss: 0.026403134688735008\n", + "Starting epoch 3092\n", + "Train loss: 0.014319202583283185, Test loss: 0.02730529010295868\n", + "Starting epoch 3093\n", + "Train loss: 0.014295227210968733, Test loss: 0.02431473135948181\n", + "Starting epoch 3094\n", + "Train loss: 0.014273237343877554, Test loss: 0.027901574969291687\n", + "Starting epoch 3095\n", + "Train loss: 0.014266655761748553, Test loss: 0.0274895578622818\n", + "Starting epoch 3096\n", + "Train loss: 0.014252655692398548, Test loss: 0.02718881517648697\n", + "Starting epoch 3097\n", + "Train loss: 0.01406571377068758, Test loss: 0.01402804534882307\n", + "Starting epoch 3098\n", + "Train loss: 0.014295556619763373, Test loss: 0.012351431883871555\n", + "Starting epoch 3099\n", + "Train loss: 0.014224862568080426, Test loss: 0.015995126217603683\n", + "Starting epoch 3100\n", + "Train loss: 0.014277513083070516, Test loss: 0.026270858943462372\n", + "Starting epoch 3101\n", + "Train loss: 0.014337433893233538, Test loss: 0.02588784135878086\n", + "Starting epoch 3102\n", + "Train loss: 0.014285673554986715, Test loss: 0.02907298132777214\n", + "Starting epoch 3103\n", + "Train loss: 0.014269714038819074, Test loss: 0.02283184416592121\n", + "Starting epoch 3104\n", + "Train loss: 0.014255117774009705, Test loss: 0.025233712047338486\n", + "Starting epoch 3105\n", + "Train loss: 0.014280559420585632, Test loss: 0.02654653787612915\n", + "Starting epoch 3106\n", + "Train loss: 0.01420996367931366, Test loss: 0.027836771681904793\n", + "Starting epoch 3107\n", + "Train loss: 0.014222456496208907, Test loss: 0.03135602921247482\n", + "Starting epoch 3108\n", + "Train loss: 0.014219819344580173, Test loss: 0.024181410670280457\n", + "Starting epoch 3109\n", + "Train loss: 0.014202458411455154, Test loss: 0.024775652214884758\n", + "Starting epoch 3110\n", + "Train loss: 0.014232428073883056, Test loss: 0.024342603981494904\n", + "Starting epoch 3111\n", + "Train loss: 0.014141585174947978, Test loss: 0.011566733941435814\n", + "Starting epoch 3112\n", + "Train loss: 0.014224080983549357, Test loss: 0.014396886341273785\n", + "Starting epoch 3113\n", + "Train loss: 0.014247171469032765, Test loss: 0.012657497078180313\n", + "Starting epoch 3114\n", + "Train loss: 0.014194913525134325, Test loss: 0.011498779058456421\n", + "Starting epoch 3115\n", + "Train loss: 0.014225731045007706, Test loss: 0.013802970759570599\n", + "Starting epoch 3116\n", + "Train loss: 0.01420769814401865, Test loss: 0.012896111235022545\n", + "Starting epoch 3117\n", + "Train loss: 0.014255697261542083, Test loss: 0.01345914788544178\n", + "Starting epoch 3118\n", + "Train loss: 0.014058492071926594, Test loss: 0.013362106867134571\n", + "Starting epoch 3119\n", + "Train loss: 0.014308656435459852, Test loss: 0.013069301843643188\n", + "Starting epoch 3120\n", + "Train loss: 0.014258506372570992, Test loss: 0.011897861026227474\n", + "Starting epoch 3121\n", + "Train loss: 0.014203858077526093, Test loss: 0.014586039818823338\n", + "Starting epoch 3122\n", + "Train loss: 0.014296228308230638, Test loss: 0.013994839042425156\n", + "Starting epoch 3123\n", + "Train loss: 0.014187384676188231, Test loss: 0.012869789265096188\n", + "Starting epoch 3124\n", + "Train loss: 0.01417613547295332, Test loss: 0.013353575021028519\n", + "Starting epoch 3125\n", + "Train loss: 0.014233298022300006, Test loss: 0.02366284653544426\n", + "Starting epoch 3126\n", + "Train loss: 0.014145339149981737, Test loss: 0.027106398716568947\n", + "Starting epoch 3127\n", + "Train loss: 0.014006770495325328, Test loss: 0.031297679990530014\n", + "Starting epoch 3128\n", + "Train loss: 0.014275576174259185, Test loss: 0.02571018412709236\n", + "Starting epoch 3129\n", + "Train loss: 0.01430002074688673, Test loss: 0.02825777418911457\n", + "Starting epoch 3130\n", + "Train loss: 0.014221742395311593, Test loss: 0.025990640744566917\n", + "Starting epoch 3131\n", + "Train loss: 0.014234491046518088, Test loss: 0.028463419526815414\n", + "Starting epoch 3132\n", + "Train loss: 0.014162161611020565, Test loss: 0.029746241867542267\n", + "Starting epoch 3133\n", + "Train loss: 0.014154559094458818, Test loss: 0.0311186071485281\n", + "Starting epoch 3134\n", + "Train loss: 0.014152101185172796, Test loss: 0.026526108384132385\n", + "Starting epoch 3135\n", + "Train loss: 0.014253621976822615, Test loss: 0.025684058666229248\n", + "Starting epoch 3136\n", + "Train loss: 0.014197144974023104, Test loss: 0.02470925636589527\n", + "Starting epoch 3137\n", + "Train loss: 0.014139192793518305, Test loss: 0.02191580832004547\n", + "Starting epoch 3138\n", + "Train loss: 0.01417267082259059, Test loss: 0.02739240601658821\n", + "Starting epoch 3139\n", + "Train loss: 0.01414647625759244, Test loss: 0.026552103459835052\n", + "Starting epoch 3140\n", + "Train loss: 0.014191235341131687, Test loss: 0.027797577902674675\n", + "Starting epoch 3141\n", + "Train loss: 0.0141047671623528, Test loss: 0.02530643902719021\n", + "Starting epoch 3142\n", + "Train loss: 0.01408187009394169, Test loss: 0.02523856610059738\n", + "Starting epoch 3143\n", + "Train loss: 0.014103641919791698, Test loss: 0.026895223185420036\n", + "Starting epoch 3144\n", + "Train loss: 0.01416557228192687, Test loss: 0.024417662993073463\n", + "Starting epoch 3145\n", + "Train loss: 0.014041063506156207, Test loss: 0.021026914939284325\n", + "Starting epoch 3146\n", + "Train loss: 0.014175795875489712, Test loss: 0.026850052177906036\n", + "Starting epoch 3147\n", + "Train loss: 0.014268123824149371, Test loss: 0.028005171567201614\n", + "Starting epoch 3148\n", + "Train loss: 0.014222396966069938, Test loss: 0.02224048413336277\n", + "Starting epoch 3149\n", + "Train loss: 0.014253145325928926, Test loss: 0.021711232140660286\n", + "Starting epoch 3150\n", + "Train loss: 0.014081800431013107, Test loss: 0.025720613077282906\n", + "Starting epoch 3151\n", + "Train loss: 0.014270231984555721, Test loss: 0.0255312267690897\n", + "Starting epoch 3152\n", + "Train loss: 0.014199095889925956, Test loss: 0.02486850880086422\n", + "Starting epoch 3153\n", + "Train loss: 0.014069787859916687, Test loss: 0.02645525150001049\n", + "Starting epoch 3154\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.01425040651112795, Test loss: 0.02248060144484043\n", + "Starting epoch 3155\n", + "Train loss: 0.014117298852652312, Test loss: 0.027803542092442513\n", + "Starting epoch 3156\n", + "Train loss: 0.01406531684100628, Test loss: 0.026721013709902763\n", + "Starting epoch 3157\n", + "Train loss: 0.014138710927218199, Test loss: 0.026922445744276047\n", + "Starting epoch 3158\n", + "Train loss: 0.014141483288258315, Test loss: 0.024452928453683853\n", + "Starting epoch 3159\n", + "Train loss: 0.014082150626927614, Test loss: 0.02998332679271698\n", + "Starting epoch 3160\n", + "Train loss: 0.014175743032246828, Test loss: 0.02274676412343979\n", + "Starting epoch 3161\n", + "Train loss: 0.01413166593760252, Test loss: 0.022415753453969955\n", + "Starting epoch 3162\n", + "Train loss: 0.014168860111385583, Test loss: 0.02674095146358013\n", + "Starting epoch 3163\n", + "Train loss: 0.014090993646532298, Test loss: 0.029894545674324036\n", + "Starting epoch 3164\n", + "Train loss: 0.014130187705159187, Test loss: 0.025101320818066597\n", + "Starting epoch 3165\n", + "Train loss: 0.014073329959064722, Test loss: 0.023100951686501503\n", + "Starting epoch 3166\n", + "Train loss: 0.014156624283641577, Test loss: 0.026477079838514328\n", + "Starting epoch 3167\n", + "Train loss: 0.014131358172744513, Test loss: 0.028581025078892708\n", + "Starting epoch 3168\n", + "Train loss: 0.014039149675518274, Test loss: 0.02891699969768524\n", + "Starting epoch 3169\n", + "Train loss: 0.01413692556321621, Test loss: 0.025468874722719193\n", + "Starting epoch 3170\n", + "Train loss: 0.013932494800537825, Test loss: 0.019468776881694794\n", + "Starting epoch 3171\n", + "Train loss: 0.014018518440425396, Test loss: 0.026450365781784058\n", + "Starting epoch 3172\n", + "Train loss: 0.014177743401378394, Test loss: 0.02409222349524498\n", + "Starting epoch 3173\n", + "Train loss: 0.014150445070117711, Test loss: 0.02389269694685936\n", + "Starting epoch 3174\n", + "Train loss: 0.01405910836532712, Test loss: 0.024489030241966248\n", + "Starting epoch 3175\n", + "Train loss: 0.014098687395453453, Test loss: 0.025126928463578224\n", + "Starting epoch 3176\n", + "Train loss: 0.014096098281443118, Test loss: 0.02791631408035755\n", + "Starting epoch 3177\n", + "Train loss: 0.013999696094542742, Test loss: 0.02258128672838211\n", + "Starting epoch 3178\n", + "Train loss: 0.01405549531802535, Test loss: 0.026632055640220642\n", + "Starting epoch 3179\n", + "Train loss: 0.01405235230922699, Test loss: 0.026526691392064095\n", + "Starting epoch 3180\n", + "Train loss: 0.014071683250367641, Test loss: 0.02762324921786785\n", + "Starting epoch 3181\n", + "Train loss: 0.014054825864732265, Test loss: 0.024159012362360954\n", + "Starting epoch 3182\n", + "Train loss: 0.013924752920866012, Test loss: 0.023933464661240578\n", + "Starting epoch 3183\n", + "Train loss: 0.01406056821346283, Test loss: 0.025757290422916412\n", + "Starting epoch 3184\n", + "Train loss: 0.014397711083292962, Test loss: 0.022845137864351273\n", + "Starting epoch 3185\n", + "Train loss: 0.014875273555517196, Test loss: 0.025245023891329765\n", + "Starting epoch 3186\n", + "Train loss: 0.014297453090548515, Test loss: 0.026889951899647713\n", + "Starting epoch 3187\n", + "Train loss: 0.014236536901444196, Test loss: 0.028807522729039192\n", + "Starting epoch 3188\n", + "Train loss: 0.014204602576792241, Test loss: 0.030644692480564117\n", + "Starting epoch 3189\n", + "Train loss: 0.014156229868531227, Test loss: 0.027748893946409225\n", + "Starting epoch 3190\n", + "Train loss: 0.014101539254188538, Test loss: 0.025807777419686317\n", + "Starting epoch 3191\n", + "Train loss: 0.014076208602637053, Test loss: 0.012911989353597164\n", + "Starting epoch 3192\n", + "Train loss: 0.014105587229132653, Test loss: 0.014468214474618435\n", + "Starting epoch 3193\n", + "Train loss: 0.014053852930665016, Test loss: 0.01345011591911316\n", + "Starting epoch 3194\n", + "Train loss: 0.01413111524656415, Test loss: 0.014242828823626041\n", + "Starting epoch 3195\n", + "Train loss: 0.01406328346580267, Test loss: 0.012330450117588043\n", + "Starting epoch 3196\n", + "Train loss: 0.014082264881581067, Test loss: 0.014258066192269325\n", + "Starting epoch 3197\n", + "Train loss: 0.01404208729043603, Test loss: 0.012541555799543858\n", + "Starting epoch 3198\n", + "Train loss: 0.013966249544173479, Test loss: 0.01185718271881342\n", + "Starting epoch 3199\n", + "Train loss: 0.013930614721029997, Test loss: 0.026347747072577477\n", + "Starting epoch 3200\n", + "Train loss: 0.013970749881118537, Test loss: 0.02565557137131691\n", + "Starting epoch 3201\n", + "Train loss: 0.01406375976279378, Test loss: 0.022808367386460304\n", + "Starting epoch 3202\n", + "Train loss: 0.014066833350807429, Test loss: 0.02909800410270691\n", + "Starting epoch 3203\n", + "Train loss: 0.015044111423194409, Test loss: 0.01968430168926716\n", + "Starting epoch 3204\n", + "Train loss: 0.015110324081033468, Test loss: 0.024486416950821877\n", + "Starting epoch 3205\n", + "Train loss: 0.014368836972862483, Test loss: 0.025418994948267937\n", + "Starting epoch 3206\n", + "Train loss: 0.014232442155480385, Test loss: 0.0249599851667881\n", + "Starting epoch 3207\n", + "Train loss: 0.014068646058440208, Test loss: 0.0273946113884449\n", + "Starting epoch 3208\n", + "Train loss: 0.014105481188744306, Test loss: 0.027193035930395126\n", + "Starting epoch 3209\n", + "Train loss: 0.014116442706435918, Test loss: 0.027416709810495377\n", + "Starting epoch 3210\n", + "Train loss: 0.014022191911935806, Test loss: 0.011745951138436794\n", + "Starting epoch 3211\n", + "Train loss: 0.014155776202678681, Test loss: 0.014114740304648876\n", + "Starting epoch 3212\n", + "Train loss: 0.014063518103212118, Test loss: 0.012829362414777279\n", + "Starting epoch 3213\n", + "Train loss: 0.013978530056774616, Test loss: 0.0151413194835186\n", + "Starting epoch 3214\n", + "Train loss: 0.013958951253443956, Test loss: 0.013042239472270012\n", + "Starting epoch 3215\n", + "Train loss: 0.013961456157267094, Test loss: 0.012980367057025433\n", + "Starting epoch 3216\n", + "Train loss: 0.01395656295120716, Test loss: 0.012350794859230518\n", + "Starting epoch 3217\n", + "Train loss: 0.013945225588977338, Test loss: 0.011861241422593594\n", + "Starting epoch 3218\n", + "Train loss: 0.013923465386033057, Test loss: 0.012304731644690037\n", + "Starting epoch 3219\n", + "Train loss: 0.013868046104907989, Test loss: 0.025489136576652527\n", + "Starting epoch 3220\n", + "Train loss: 0.013913005851209163, Test loss: 0.027724724262952805\n", + "Starting epoch 3221\n", + "Train loss: 0.014536691382527351, Test loss: 0.022659318521618843\n", + "Starting epoch 3222\n", + "Train loss: 0.014066640567034484, Test loss: 0.028716066852211952\n", + "Starting epoch 3223\n", + "Train loss: 0.014089407492429018, Test loss: 0.02812964841723442\n", + "Starting epoch 3224\n", + "Train loss: 0.014072404839098454, Test loss: 0.020654870197176933\n", + "Starting epoch 3225\n", + "Train loss: 0.013935167640447617, Test loss: 0.02484479732811451\n", + "Starting epoch 3226\n", + "Train loss: 0.014092456363141537, Test loss: 0.027104586362838745\n", + "Starting epoch 3227\n", + "Train loss: 0.014048886001110077, Test loss: 0.02652202919125557\n", + "Starting epoch 3228\n", + "Train loss: 0.014032574351876975, Test loss: 0.028480734676122665\n", + "Starting epoch 3229\n", + "Train loss: 0.01401299236342311, Test loss: 0.025537872686982155\n", + "Starting epoch 3230\n", + "Train loss: 0.013898817282170057, Test loss: 0.023236146196722984\n", + "Starting epoch 3231\n", + "Train loss: 0.013970697838813067, Test loss: 0.024461237713694572\n", + "Starting epoch 3232\n", + "Train loss: 0.014004769790917635, Test loss: 0.026479629799723625\n", + "Starting epoch 3233\n", + "Train loss: 0.01397653853520751, Test loss: 0.026936933398246765\n", + "Starting epoch 3234\n", + "Train loss: 0.013902857601642608, Test loss: 0.024338582530617714\n", + "Starting epoch 3235\n", + "Train loss: 0.013880821708589793, Test loss: 0.026403438299894333\n", + "Starting epoch 3236\n", + "Train loss: 0.013915385641157628, Test loss: 0.02531329356133938\n", + "Starting epoch 3237\n", + "Train loss: 0.013873981349170207, Test loss: 0.0180529598146677\n", + "Starting epoch 3238\n", + "Train loss: 0.013929245583713055, Test loss: 0.02562537044286728\n", + "Starting epoch 3239\n", + "Train loss: 0.013903848435729742, Test loss: 0.02639887109398842\n", + "Starting epoch 3240\n", + "Train loss: 0.01386777650564909, Test loss: 0.027866305783391\n", + "Starting epoch 3241\n", + "Train loss: 0.013772785253822804, Test loss: 0.025097480043768883\n", + "Starting epoch 3242\n", + "Train loss: 0.013893420305103064, Test loss: 0.027688147500157356\n", + "Starting epoch 3243\n", + "Train loss: 0.013995122946798801, Test loss: 0.02651342749595642\n", + "Starting epoch 3244\n", + "Train loss: 0.013864343110471964, Test loss: 0.02427096478641033\n", + "Starting epoch 3245\n", + "Train loss: 0.013918973430991173, Test loss: 0.026776224374771118\n", + "Starting epoch 3246\n", + "Train loss: 0.013945736065506935, Test loss: 0.02596772089600563\n", + "Starting epoch 3247\n", + "Train loss: 0.013908190950751305, Test loss: 0.013065493665635586\n", + "Starting epoch 3248\n", + "Train loss: 0.013942226860672235, Test loss: 0.012589084915816784\n", + "Starting epoch 3249\n", + "Train loss: 0.01384259771555662, Test loss: 0.014806818217039108\n", + "Starting epoch 3250\n", + "Train loss: 0.01394931562244892, Test loss: 0.013666623272001743\n", + "Starting epoch 3251\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.013889062348753214, Test loss: 0.016169700771570206\n", + "Starting epoch 3252\n", + "Train loss: 0.013830839600414037, Test loss: 0.01153399795293808\n", + "Starting epoch 3253\n", + "Train loss: 0.013894326351583004, Test loss: 0.012782399542629719\n", + "Starting epoch 3254\n", + "Train loss: 0.013789192754775286, Test loss: 0.020603526383638382\n", + "Starting epoch 3255\n", + "Train loss: 0.013899051900953054, Test loss: 0.0258079394698143\n", + "Starting epoch 3256\n", + "Train loss: 0.013872764464467764, Test loss: 0.026413248851895332\n", + "Starting epoch 3257\n", + "Train loss: 0.01390653794631362, Test loss: 0.026000235229730606\n", + "Starting epoch 3258\n", + "Train loss: 0.013952617030590773, Test loss: 0.02523161843419075\n", + "Starting epoch 3259\n", + "Train loss: 0.013895724881440401, Test loss: 0.02798188477754593\n", + "Starting epoch 3260\n", + "Train loss: 0.013951428681612015, Test loss: 0.02128710225224495\n", + "Starting epoch 3261\n", + "Train loss: 0.013886192608624696, Test loss: 0.024911940097808838\n", + "Starting epoch 3262\n", + "Train loss: 0.013880277145653963, Test loss: 0.02120342291891575\n", + "Starting epoch 3263\n", + "Train loss: 0.013862344697117805, Test loss: 0.012080875225365162\n", + "Starting epoch 3264\n", + "Train loss: 0.013918675221502782, Test loss: 0.012974243611097336\n", + "Starting epoch 3265\n", + "Train loss: 0.013874120991677046, Test loss: 0.012298764660954475\n", + "Starting epoch 3266\n", + "Train loss: 0.013782157767564059, Test loss: 0.01657574437558651\n", + "Starting epoch 3267\n", + "Train loss: 0.01382303211838007, Test loss: 0.013783064670860767\n", + "Starting epoch 3268\n", + "Train loss: 0.01389460038393736, Test loss: 0.014613236300647259\n", + "Starting epoch 3269\n", + "Train loss: 0.013822494987398386, Test loss: 0.015172231011092663\n", + "Starting epoch 3270\n", + "Train loss: 0.01394146041944623, Test loss: 0.012318228371441364\n", + "Starting epoch 3271\n", + "Train loss: 0.013773456458002329, Test loss: 0.015583913773298264\n", + "Starting epoch 3272\n", + "Train loss: 0.013835090212523937, Test loss: 0.027547065168619156\n", + "Starting epoch 3273\n", + "Train loss: 0.01394634772092104, Test loss: 0.02334464155137539\n", + "Starting epoch 3274\n", + "Train loss: 0.013866488765925169, Test loss: 0.011928862892091274\n", + "Starting epoch 3275\n", + "Train loss: 0.013938570581376552, Test loss: 0.015151593834161758\n", + "Starting epoch 3276\n", + "Train loss: 0.01383643563836813, Test loss: 0.011743457056581974\n", + "Starting epoch 3277\n", + "Train loss: 0.013889427855610847, Test loss: 0.012510290369391441\n", + "Starting epoch 3278\n", + "Train loss: 0.013876585625112056, Test loss: 0.013177408836781979\n", + "Starting epoch 3279\n", + "Train loss: 0.013781850188970567, Test loss: 0.011842080391943455\n", + "Starting epoch 3280\n", + "Train loss: 0.013804298862814902, Test loss: 0.013009159825742245\n", + "Starting epoch 3281\n", + "Train loss: 0.01384152639657259, Test loss: 0.01289288979023695\n", + "Starting epoch 3282\n", + "Train loss: 0.013800540026277303, Test loss: 0.012779896147549152\n", + "Starting epoch 3283\n", + "Train loss: 0.013888030853122473, Test loss: 0.012893930077552795\n", + "Starting epoch 3284\n", + "Train loss: 0.013750614877790213, Test loss: 0.014217127114534378\n", + "Starting epoch 3285\n", + "Train loss: 0.013824729342013597, Test loss: 0.014605678617954254\n", + "Starting epoch 3286\n", + "Train loss: 0.013839500658214092, Test loss: 0.012944639660418034\n", + "Starting epoch 3287\n", + "Train loss: 0.013897674828767777, Test loss: 0.012678735889494419\n", + "Starting epoch 3288\n", + "Train loss: 0.013838891237974167, Test loss: 0.013904708437621593\n", + "Starting epoch 3289\n", + "Train loss: 0.013794009555131197, Test loss: 0.011780889704823494\n", + "Starting epoch 3290\n", + "Train loss: 0.013817465472966433, Test loss: 0.013187067583203316\n", + "Starting epoch 3291\n", + "Train loss: 0.013876984957605601, Test loss: 0.013540462590754032\n", + "Starting epoch 3292\n", + "Train loss: 0.01384684856981039, Test loss: 0.012556585483253002\n", + "Starting epoch 3293\n", + "Train loss: 0.013890626039355994, Test loss: 0.018374282866716385\n", + "Starting epoch 3294\n", + "Train loss: 0.013807443659752608, Test loss: 0.011824398301541805\n", + "Starting epoch 3295\n", + "Train loss: 0.01373306980356574, Test loss: 0.014845320954918861\n", + "Starting epoch 3296\n", + "Train loss: 0.01376885075122118, Test loss: 0.011512779630720615\n", + "Starting epoch 3297\n", + "Train loss: 0.013859963640570641, Test loss: 0.013288904912769794\n", + "Starting epoch 3298\n", + "Train loss: 0.013843885250389576, Test loss: 0.012957344762980938\n", + "Starting epoch 3299\n", + "Train loss: 0.013803572449833155, Test loss: 0.012411107309162617\n", + "Starting epoch 3300\n", + "Train loss: 0.013815130647271872, Test loss: 0.0122825363650918\n", + "Starting epoch 3301\n", + "Train loss: 0.013752832617610693, Test loss: 0.01487045083194971\n", + "Starting epoch 3302\n", + "Train loss: 0.013677052836865187, Test loss: 0.012650088407099247\n", + "Starting epoch 3303\n", + "Train loss: 0.013840926140546798, Test loss: 0.012777455151081085\n", + "Starting epoch 3304\n", + "Train loss: 0.013791632447391748, Test loss: 0.013627899810671806\n", + "Starting epoch 3305\n", + "Train loss: 0.013784351777285337, Test loss: 0.012429364025592804\n", + "Starting epoch 3306\n", + "Train loss: 0.013832921143621207, Test loss: 0.013157308101654053\n", + "Starting epoch 3307\n", + "Train loss: 0.013678680565208197, Test loss: 0.012259774841368198\n", + "Starting epoch 3308\n", + "Train loss: 0.013732130341231823, Test loss: 0.012750057503581047\n", + "Starting epoch 3309\n", + "Train loss: 0.013760169874876738, Test loss: 0.014283910393714905\n", + "Starting epoch 3310\n", + "Train loss: 0.013811877518892289, Test loss: 0.01547855231910944\n", + "Starting epoch 3311\n", + "Train loss: 0.013840615656226874, Test loss: 0.013722539879381657\n", + "Starting epoch 3312\n", + "Train loss: 0.013753766752779483, Test loss: 0.012634233571588993\n", + "Starting epoch 3313\n", + "Train loss: 0.013696264773607254, Test loss: 0.01578310690820217\n", + "Starting epoch 3314\n", + "Train loss: 0.013785363361239433, Test loss: 0.012899321503937244\n", + "Starting epoch 3315\n", + "Train loss: 0.013792622871696948, Test loss: 0.011586347594857216\n", + "Starting epoch 3316\n", + "Train loss: 0.013726730775088071, Test loss: 0.017680862918496132\n", + "Starting epoch 3317\n", + "Train loss: 0.013823169898241758, Test loss: 0.013395260088145733\n", + "Starting epoch 3318\n", + "Train loss: 0.013770685698837042, Test loss: 0.012674913741648197\n", + "Starting epoch 3319\n", + "Train loss: 0.013735419232398271, Test loss: 0.013872053474187851\n", + "Starting epoch 3320\n", + "Train loss: 0.01370933661237359, Test loss: 0.01483822986483574\n", + "Starting epoch 3321\n", + "Train loss: 0.013770256303250789, Test loss: 0.01184113323688507\n", + "Starting epoch 3322\n", + "Train loss: 0.013819812145084143, Test loss: 0.014473705552518368\n", + "Starting epoch 3323\n", + "Train loss: 0.013798785377293826, Test loss: 0.01416736375540495\n", + "Starting epoch 3324\n", + "Train loss: 0.013805007580667734, Test loss: 0.014497802592813969\n", + "Starting epoch 3325\n", + "Train loss: 0.01367776507511735, Test loss: 0.01092718169093132\n", + "Starting epoch 3326\n", + "Train loss: 0.013629772905260325, Test loss: 0.019256114959716797\n", + "Starting epoch 3327\n", + "Train loss: 0.013786590434610843, Test loss: 0.027626927942037582\n", + "Starting epoch 3328\n", + "Train loss: 0.01382837638258934, Test loss: 0.027433207258582115\n", + "Starting epoch 3329\n", + "Train loss: 0.013776413276791572, Test loss: 0.023744046688079834\n", + "Starting epoch 3330\n", + "Train loss: 0.013745738435536623, Test loss: 0.026768147945404053\n", + "Starting epoch 3331\n", + "Train loss: 0.013660695776343346, Test loss: 0.02445961721241474\n", + "Starting epoch 3332\n", + "Train loss: 0.01372257299721241, Test loss: 0.025579417124390602\n", + "Starting epoch 3333\n", + "Train loss: 0.0137039228156209, Test loss: 0.02448003739118576\n", + "Starting epoch 3334\n", + "Train loss: 0.013755695577710867, Test loss: 0.021841494366526604\n", + "Starting epoch 3335\n", + "Train loss: 0.01377655727788806, Test loss: 0.026405787095427513\n", + "Starting epoch 3336\n", + "Train loss: 0.01369006011635065, Test loss: 0.02511320263147354\n", + "Starting epoch 3337\n", + "Train loss: 0.013755358662456274, Test loss: 0.026698123663663864\n", + "Starting epoch 3338\n", + "Train loss: 0.013757287841290235, Test loss: 0.023426542058587074\n", + "Starting epoch 3339\n", + "Train loss: 0.013722604140639305, Test loss: 0.013320707716047764\n", + "Starting epoch 3340\n", + "Train loss: 0.013738852422684432, Test loss: 0.015053884126245975\n", + "Starting epoch 3341\n", + "Train loss: 0.01369057010859251, Test loss: 0.01294812373816967\n", + "Starting epoch 3342\n", + "Train loss: 0.013690258841961622, Test loss: 0.013321603648364544\n", + "Starting epoch 3343\n", + "Train loss: 0.01367202278226614, Test loss: 0.012793983332812786\n", + "Starting epoch 3344\n", + "Train loss: 0.01365233276039362, Test loss: 0.011626199819147587\n", + "Starting epoch 3345\n", + "Train loss: 0.013768343981355428, Test loss: 0.01587524823844433\n", + "Starting epoch 3346\n", + "Train loss: 0.013700652159750462, Test loss: 0.014149738475680351\n", + "Starting epoch 3347\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.013608818985521793, Test loss: 0.014146802015602589\n", + "Starting epoch 3348\n", + "Train loss: 0.01373562004417181, Test loss: 0.011128979735076427\n", + "Starting epoch 3349\n", + "Train loss: 0.013685525115579366, Test loss: 0.013059697113931179\n", + "Starting epoch 3350\n", + "Train loss: 0.013599527422338725, Test loss: 0.02507845126092434\n", + "Starting epoch 3351\n", + "Train loss: 0.013742164429277181, Test loss: 0.02587031200528145\n", + "Starting epoch 3352\n", + "Train loss: 0.013614717591553926, Test loss: 0.02025241032242775\n", + "Starting epoch 3353\n", + "Train loss: 0.013585716784000397, Test loss: 0.0253860205411911\n", + "Starting epoch 3354\n", + "Train loss: 0.013767183758318424, Test loss: 0.02705344557762146\n", + "Starting epoch 3355\n", + "Train loss: 0.013738005813211203, Test loss: 0.026548996567726135\n", + "Starting epoch 3356\n", + "Train loss: 0.013639568760991096, Test loss: 0.022525856271386147\n", + "Starting epoch 3357\n", + "Train loss: 0.013730789367109537, Test loss: 0.021806681528687477\n", + "Starting epoch 3358\n", + "Train loss: 0.013678747843950987, Test loss: 0.014145230874419212\n", + "Starting epoch 3359\n", + "Train loss: 0.013721408471465111, Test loss: 0.014586380682885647\n", + "Starting epoch 3360\n", + "Train loss: 0.013674178086221218, Test loss: 0.012765379622578621\n", + "Starting epoch 3361\n", + "Train loss: 0.013674942553043365, Test loss: 0.012780960649251938\n", + "Starting epoch 3362\n", + "Train loss: 0.013683656826615334, Test loss: 0.012376143597066402\n", + "Starting epoch 3363\n", + "Train loss: 0.013607258908450603, Test loss: 0.01145872287452221\n", + "Starting epoch 3364\n", + "Train loss: 0.01368001477792859, Test loss: 0.012186295352876186\n", + "Starting epoch 3365\n", + "Train loss: 0.01368161702528596, Test loss: 0.012454887852072716\n", + "Starting epoch 3366\n", + "Train loss: 0.013751784786581994, Test loss: 0.015153602696955204\n", + "Starting epoch 3367\n", + "Train loss: 0.01373831184580922, Test loss: 0.012083737179636955\n", + "Starting epoch 3368\n", + "Train loss: 0.013684988673776387, Test loss: 0.014359129592776299\n", + "Starting epoch 3369\n", + "Train loss: 0.013637645747512578, Test loss: 0.014670253731310368\n", + "Starting epoch 3370\n", + "Train loss: 0.015228900108486414, Test loss: 0.02256542257964611\n", + "Starting epoch 3371\n", + "Train loss: 0.020110318679362537, Test loss: 0.01545186061412096\n", + "Starting epoch 3372\n", + "Train loss: 0.018502746671438218, Test loss: 0.017621492967009544\n", + "Starting epoch 3373\n", + "Train loss: 0.014808202981948853, Test loss: 0.014195248484611511\n", + "Starting epoch 3374\n", + "Train loss: 0.014147161338478327, Test loss: 0.011912940070033073\n", + "Starting epoch 3375\n", + "Train loss: 0.013918183296918869, Test loss: 0.01181529089808464\n", + "Starting epoch 3376\n", + "Train loss: 0.013918262533843517, Test loss: 0.014147271402180195\n", + "Starting epoch 3377\n", + "Train loss: 0.013829590268433094, Test loss: 0.013151920400559902\n", + "Starting epoch 3378\n", + "Train loss: 0.013736645691096783, Test loss: 0.011874956078827381\n", + "Starting epoch 3379\n", + "Train loss: 0.01371235204860568, Test loss: 0.012906722724437714\n", + "Starting epoch 3380\n", + "Train loss: 0.013699122183024883, Test loss: 0.01105064619332552\n", + "Starting epoch 3381\n", + "Train loss: 0.013599658142775297, Test loss: 0.024862343445420265\n", + "Starting epoch 3382\n", + "Train loss: 0.01366856649518013, Test loss: 0.02297777682542801\n", + "Starting epoch 3383\n", + "Train loss: 0.013643932715058326, Test loss: 0.024504369124770164\n", + "Starting epoch 3384\n", + "Train loss: 0.013635743614286185, Test loss: 0.02490060403943062\n", + "Starting epoch 3385\n", + "Train loss: 0.01361881086602807, Test loss: 0.02043982408940792\n", + "Starting epoch 3386\n", + "Train loss: 0.013562139440327883, Test loss: 0.02540011703968048\n", + "Starting epoch 3387\n", + "Train loss: 0.013505185637623072, Test loss: 0.023820115253329277\n", + "Starting epoch 3388\n", + "Train loss: 0.013549368754029274, Test loss: 0.02484147623181343\n", + "Starting epoch 3389\n", + "Train loss: 0.013522563315927983, Test loss: 0.025012152269482613\n", + "Starting epoch 3390\n", + "Train loss: 0.013523987643420696, Test loss: 0.02515069767832756\n", + "Starting epoch 3391\n", + "Train loss: 0.013484900798648596, Test loss: 0.027180997654795647\n", + "Starting epoch 3392\n", + "Train loss: 0.0135104120336473, Test loss: 0.0248943492770195\n", + "Starting epoch 3393\n", + "Train loss: 0.013443176876753568, Test loss: 0.012305675074458122\n", + "Starting epoch 3394\n", + "Train loss: 0.013514776453375817, Test loss: 0.02204812318086624\n", + "Starting epoch 3395\n", + "Train loss: 0.013623314686119556, Test loss: 0.024337906390428543\n", + "Starting epoch 3396\n", + "Train loss: 0.013586352989077568, Test loss: 0.02453169785439968\n", + "Starting epoch 3397\n", + "Train loss: 0.013547928761690855, Test loss: 0.01952246017754078\n", + "Starting epoch 3398\n", + "Train loss: 0.013528792764991521, Test loss: 0.01190401054918766\n", + "Starting epoch 3399\n", + "Train loss: 0.013536766897886991, Test loss: 0.02354786917567253\n", + "Starting epoch 3400\n", + "Train loss: 0.01362999051809311, Test loss: 0.024913286790251732\n", + "Starting epoch 3401\n", + "Train loss: 0.013625625763088465, Test loss: 0.02384076826274395\n", + "Starting epoch 3402\n", + "Train loss: 0.013579510506242514, Test loss: 0.021070441231131554\n", + "Starting epoch 3403\n", + "Train loss: 0.013600026611238719, Test loss: 0.02675836905837059\n", + "Starting epoch 3404\n", + "Train loss: 0.013577555511146784, Test loss: 0.02374468930065632\n", + "Starting epoch 3405\n", + "Train loss: 0.013551072161644697, Test loss: 0.027676833793520927\n", + "Starting epoch 3406\n", + "Train loss: 0.013555580526590347, Test loss: 0.026259690523147583\n", + "Starting epoch 3407\n", + "Train loss: 0.013560290541499853, Test loss: 0.02352231740951538\n", + "Starting epoch 3408\n", + "Train loss: 0.013601630721241235, Test loss: 0.02560519054532051\n", + "Starting epoch 3409\n", + "Train loss: 0.013552647028118371, Test loss: 0.01298593357205391\n", + "Starting epoch 3410\n", + "Train loss: 0.013636858686804771, Test loss: 0.012711863964796066\n", + "Starting epoch 3411\n", + "Train loss: 0.013574671279639006, Test loss: 0.013544042594730854\n", + "Starting epoch 3412\n", + "Train loss: 0.013537527602165937, Test loss: 0.027017807587981224\n", + "Starting epoch 3413\n", + "Train loss: 0.013595646470785141, Test loss: 0.025509245693683624\n", + "Starting epoch 3414\n", + "Train loss: 0.013545511625707149, Test loss: 0.024113250896334648\n", + "Starting epoch 3415\n", + "Train loss: 0.01353438152000308, Test loss: 0.027166469022631645\n", + "Starting epoch 3416\n", + "Train loss: 0.013543482981622218, Test loss: 0.013778266496956348\n", + "Starting epoch 3417\n", + "Train loss: 0.013560519702732563, Test loss: 0.012930973432958126\n", + "Starting epoch 3418\n", + "Train loss: 0.013528524823486805, Test loss: 0.012112735770642757\n", + "Starting epoch 3419\n", + "Train loss: 0.013555185366421939, Test loss: 0.011924130842089653\n", + "Starting epoch 3420\n", + "Train loss: 0.013517880085855723, Test loss: 0.012679177336394787\n", + "Starting epoch 3421\n", + "Train loss: 0.013490695804357529, Test loss: 0.013195239007472992\n", + "Starting epoch 3422\n", + "Train loss: 0.013474336192011833, Test loss: 0.012254630215466022\n", + "Starting epoch 3423\n", + "Train loss: 0.017305912226438524, Test loss: 0.023081105202436447\n", + "Starting epoch 3424\n", + "Train loss: 0.015095810387283564, Test loss: 0.022448547184467316\n", + "Starting epoch 3425\n", + "Train loss: 0.014058790914714337, Test loss: 0.024024944752454758\n", + "Starting epoch 3426\n", + "Train loss: 0.013923891931772231, Test loss: 0.024842889979481697\n", + "Starting epoch 3427\n", + "Train loss: 0.013830317091196776, Test loss: 0.025087809190154076\n", + "Starting epoch 3428\n", + "Train loss: 0.013626124504953622, Test loss: 0.023489005863666534\n", + "Starting epoch 3429\n", + "Train loss: 0.013756310399621725, Test loss: 0.02481609769165516\n", + "Starting epoch 3430\n", + "Train loss: 0.013677184861153364, Test loss: 0.024282686412334442\n", + "Starting epoch 3431\n", + "Train loss: 0.013639918565750121, Test loss: 0.020524481311440468\n", + "Starting epoch 3432\n", + "Train loss: 0.01343790367245674, Test loss: 0.02500339224934578\n", + "Starting epoch 3433\n", + "Train loss: 0.013753222525119782, Test loss: 0.0249240230768919\n", + "Starting epoch 3434\n", + "Train loss: 0.01370278563350439, Test loss: 0.024802222847938538\n", + "Starting epoch 3435\n", + "Train loss: 0.01361469205468893, Test loss: 0.0186535082757473\n", + "Starting epoch 3436\n", + "Train loss: 0.013631860706955195, Test loss: 0.027677418664097786\n", + "Starting epoch 3437\n", + "Train loss: 0.01363181922584772, Test loss: 0.026402117684483528\n", + "Starting epoch 3438\n", + "Train loss: 0.013554163910448551, Test loss: 0.025806782767176628\n", + "Starting epoch 3439\n", + "Train loss: 0.013575455714017153, Test loss: 0.02816266193985939\n", + "Starting epoch 3440\n", + "Train loss: 0.013555363658815623, Test loss: 0.01168806478381157\n", + "Starting epoch 3441\n", + "Train loss: 0.013554377239197493, Test loss: 0.01207832619547844\n", + "Starting epoch 3442\n", + "Train loss: 0.013604142386466265, Test loss: 0.025412647053599358\n", + "Starting epoch 3443\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.013546431306749582, Test loss: 0.01947598159313202\n", + "Starting epoch 3444\n", + "Train loss: 0.013544913548976183, Test loss: 0.024611392989754677\n", + "Starting epoch 3445\n", + "Train loss: 0.013498319536447525, Test loss: 0.024771330878138542\n", + "Starting epoch 3446\n", + "Train loss: 0.013529101368039847, Test loss: 0.028482073917984962\n", + "Starting epoch 3447\n", + "Train loss: 0.013500549159944057, Test loss: 0.02150140330195427\n", + "Starting epoch 3448\n", + "Train loss: 0.013494043461978435, Test loss: 0.025260183960199356\n", + "Starting epoch 3449\n", + "Train loss: 0.013565791025757789, Test loss: 0.02202565036714077\n", + "Starting epoch 3450\n", + "Train loss: 0.013406343646347523, Test loss: 0.02105327695608139\n", + "Starting epoch 3451\n", + "Train loss: 0.013542198091745377, Test loss: 0.021636174991726875\n", + "Starting epoch 3452\n", + "Train loss: 0.013481868468225001, Test loss: 0.020862549543380737\n", + "Starting epoch 3453\n", + "Train loss: 0.013488605469465256, Test loss: 0.022733010351657867\n", + "Starting epoch 3454\n", + "Train loss: 0.013500202000141143, Test loss: 0.02361728437244892\n", + "Starting epoch 3455\n", + "Train loss: 0.013508299011737107, Test loss: 0.02243552729487419\n", + "Starting epoch 3456\n", + "Train loss: 0.01337674655020237, Test loss: 0.026435548439621925\n", + "Starting epoch 3457\n", + "Train loss: 0.01349080566316843, Test loss: 0.0242278091609478\n", + "Starting epoch 3458\n", + "Train loss: 0.013482878878712655, Test loss: 0.02254415675997734\n", + "Starting epoch 3459\n", + "Train loss: 0.013431824371218681, Test loss: 0.0245091300457716\n", + "Starting epoch 3460\n", + "Train loss: 0.013541004881262779, Test loss: 0.023170869797468185\n", + "Starting epoch 3461\n", + "Train loss: 0.013502649404108525, Test loss: 0.02254827879369259\n", + "Starting epoch 3462\n", + "Train loss: 0.013453905899077655, Test loss: 0.02781243808567524\n", + "Starting epoch 3463\n", + "Train loss: 0.013416109960526229, Test loss: 0.01357474084943533\n", + "Starting epoch 3464\n", + "Train loss: 0.013495220206677913, Test loss: 0.015207458287477493\n", + "Starting epoch 3465\n", + "Train loss: 0.01348478626459837, Test loss: 0.011376838199794292\n", + "Starting epoch 3466\n", + "Train loss: 0.013430379405617714, Test loss: 0.011011212132871151\n", + "Starting epoch 3467\n", + "Train loss: 0.01335197253152728, Test loss: 0.023432472720742226\n", + "Starting epoch 3468\n", + "Train loss: 0.013406951129436493, Test loss: 0.022682074457406998\n", + "Starting epoch 3469\n", + "Train loss: 0.013465512786060571, Test loss: 0.020795771852135658\n", + "Starting epoch 3470\n", + "Train loss: 0.013479728903621434, Test loss: 0.024917269125580788\n", + "Starting epoch 3471\n", + "Train loss: 0.01342528274282813, Test loss: 0.02765907160937786\n", + "Starting epoch 3472\n", + "Train loss: 0.013421753924340009, Test loss: 0.02625095471739769\n", + "Starting epoch 3473\n", + "Train loss: 0.013366134501993657, Test loss: 0.011952796019613743\n", + "Starting epoch 3474\n", + "Train loss: 0.013510044720023871, Test loss: 0.013405931182205677\n", + "Starting epoch 3475\n", + "Train loss: 0.01356614476069808, Test loss: 0.01225280947983265\n", + "Starting epoch 3476\n", + "Train loss: 0.01354219052940607, Test loss: 0.013191690668463707\n", + "Starting epoch 3477\n", + "Train loss: 0.013456287793815136, Test loss: 0.027970021590590477\n", + "Starting epoch 3478\n", + "Train loss: 0.013487969692796468, Test loss: 0.01182960532605648\n", + "Starting epoch 3479\n", + "Train loss: 0.01345175039023161, Test loss: 0.014573216438293457\n", + "Starting epoch 3480\n", + "Train loss: 0.013350948803126811, Test loss: 0.011493757367134094\n", + "Starting epoch 3481\n", + "Train loss: 0.0134912090562284, Test loss: 0.010848451405763626\n", + "Starting epoch 3482\n", + "Train loss: 0.013433833234012127, Test loss: 0.017376556992530823\n", + "Starting epoch 3483\n", + "Train loss: 0.01344278583303094, Test loss: 0.021275702863931656\n", + "Starting epoch 3484\n", + "Train loss: 0.013467621393501758, Test loss: 0.02526465803384781\n", + "Starting epoch 3485\n", + "Train loss: 0.013358541037887335, Test loss: 0.022242749109864235\n", + "Starting epoch 3486\n", + "Train loss: 0.013462414517998695, Test loss: 0.023487424477934837\n", + "Starting epoch 3487\n", + "Train loss: 0.013408687077462672, Test loss: 0.025135191157460213\n", + "Starting epoch 3488\n", + "Train loss: 0.013411852214485407, Test loss: 0.02286955714225769\n", + "Starting epoch 3489\n", + "Train loss: 0.013395316693931817, Test loss: 0.021947629749774933\n", + "Starting epoch 3490\n", + "Train loss: 0.01348425155505538, Test loss: 0.025608409196138382\n", + "Starting epoch 3491\n", + "Train loss: 0.013488717917352915, Test loss: 0.019899781793355942\n", + "Starting epoch 3492\n", + "Train loss: 0.01344963300973177, Test loss: 0.026493368670344353\n", + "Starting epoch 3493\n", + "Train loss: 0.013391852881759404, Test loss: 0.017696518450975418\n", + "Starting epoch 3494\n", + "Train loss: 0.013286402113735676, Test loss: 0.012324108742177486\n", + "Starting epoch 3495\n", + "Train loss: 0.013450515046715737, Test loss: 0.012409116141498089\n", + "Starting epoch 3496\n", + "Train loss: 0.013437878414988518, Test loss: 0.014246390201151371\n", + "Starting epoch 3497\n", + "Train loss: 0.01336351877078414, Test loss: 0.01337502896785736\n", + "Starting epoch 3498\n", + "Train loss: 0.01352258762344718, Test loss: 0.027266599237918854\n", + "Starting epoch 3499\n", + "Train loss: 0.013397186808288098, Test loss: 0.024698885157704353\n", + "Starting epoch 3500\n", + "Train loss: 0.013550336807966232, Test loss: 0.022137798368930817\n", + "Starting epoch 3501\n", + "Train loss: 0.013437703177332878, Test loss: 0.025967514142394066\n", + "Starting epoch 3502\n", + "Train loss: 0.013384680841118097, Test loss: 0.024395298212766647\n", + "Starting epoch 3503\n", + "Train loss: 0.013406522888690233, Test loss: 0.02644500322639942\n", + "Starting epoch 3504\n", + "Train loss: 0.013384748008102179, Test loss: 0.02615884132683277\n", + "Starting epoch 3505\n", + "Train loss: 0.013436921946704388, Test loss: 0.024126382544636726\n", + "Starting epoch 3506\n", + "Train loss: 0.01328210648149252, Test loss: 0.02507169544696808\n", + "Starting epoch 3507\n", + "Train loss: 0.013409800864756106, Test loss: 0.027028797194361687\n", + "Starting epoch 3508\n", + "Train loss: 0.013445683382451534, Test loss: 0.02277233824133873\n", + "Starting epoch 3509\n", + "Train loss: 0.01342069337144494, Test loss: 0.023622438311576843\n", + "Starting epoch 3510\n", + "Train loss: 0.0134073156863451, Test loss: 0.02124360017478466\n", + "Starting epoch 3511\n", + "Train loss: 0.013406808581203222, Test loss: 0.020159872248768806\n", + "Starting epoch 3512\n", + "Train loss: 0.013338200952857733, Test loss: 0.01694665104150772\n", + "Starting epoch 3513\n", + "Train loss: 0.013327154032886028, Test loss: 0.011256435886025429\n", + "Starting epoch 3514\n", + "Train loss: 0.01340513300150633, Test loss: 0.01220946479588747\n", + "Starting epoch 3515\n", + "Train loss: 0.013410090040415525, Test loss: 0.014336840249598026\n", + "Starting epoch 3516\n", + "Train loss: 0.01337753675878048, Test loss: 0.012552089057862759\n", + "Starting epoch 3517\n", + "Train loss: 0.013354668766260147, Test loss: 0.014105243608355522\n", + "Starting epoch 3518\n", + "Train loss: 0.013309789095073939, Test loss: 0.012208347208797932\n", + "Starting epoch 3519\n", + "Train loss: 0.013391065523028373, Test loss: 0.013370836153626442\n", + "Starting epoch 3520\n", + "Train loss: 0.01336534008383751, Test loss: 0.014039983041584492\n", + "Starting epoch 3521\n", + "Train loss: 0.0133751037158072, Test loss: 0.011404416523873806\n", + "Starting epoch 3522\n", + "Train loss: 0.013386012353003025, Test loss: 0.012359684333205223\n", + "Starting epoch 3523\n", + "Train loss: 0.013383165802806615, Test loss: 0.010816285386681557\n", + "Starting epoch 3524\n", + "Train loss: 0.013341195285320281, Test loss: 0.012310607358813286\n", + "Starting epoch 3525\n", + "Train loss: 0.013372509721666575, Test loss: 0.012073970399796963\n", + "Starting epoch 3526\n", + "Train loss: 0.01332115801051259, Test loss: 0.013912907801568508\n", + "Starting epoch 3527\n", + "Train loss: 0.013377722688019275, Test loss: 0.011204738169908524\n", + "Starting epoch 3528\n", + "Train loss: 0.01328579692170024, Test loss: 0.02551059052348137\n", + "Starting epoch 3529\n", + "Train loss: 0.013374129608273506, Test loss: 0.021962616592645645\n", + "Starting epoch 3530\n", + "Train loss: 0.013388657923787833, Test loss: 0.025937557220458984\n", + "Starting epoch 3531\n", + "Train loss: 0.013366593457758426, Test loss: 0.02288658544421196\n", + "Starting epoch 3532\n", + "Train loss: 0.013423345182090998, Test loss: 0.023687992244958878\n", + "Starting epoch 3533\n", + "Train loss: 0.013416063375771045, Test loss: 0.02421300858259201\n", + "Starting epoch 3534\n", + "Train loss: 0.013342615570873023, Test loss: 0.020577803254127502\n", + "Starting epoch 3535\n", + "Train loss: 0.013372702579945326, Test loss: 0.021835235878825188\n", + "Starting epoch 3536\n", + "Train loss: 0.013355000913143157, Test loss: 0.02242928557097912\n", + "Starting epoch 3537\n", + "Train loss: 0.013327934667468072, Test loss: 0.02187422662973404\n", + "Starting epoch 3538\n", + "Train loss: 0.013289535287767648, Test loss: 0.02305847965180874\n", + "Starting epoch 3539\n", + "Train loss: 0.013364886734634637, Test loss: 0.024503318592905998\n", + "Starting epoch 3540\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.013362420555204153, Test loss: 0.015601391904056072\n", + "Starting epoch 3541\n", + "Train loss: 0.013328073751181365, Test loss: 0.011764193885028362\n", + "Starting epoch 3542\n", + "Train loss: 0.013345369789749383, Test loss: 0.014081720262765884\n", + "Starting epoch 3543\n", + "Train loss: 0.01333449048921466, Test loss: 0.01230755727738142\n", + "Starting epoch 3544\n", + "Train loss: 0.01331026017665863, Test loss: 0.01233359333127737\n", + "Starting epoch 3545\n", + "Train loss: 0.013351802323013544, Test loss: 0.014802254736423492\n", + "Starting epoch 3546\n", + "Train loss: 0.01326060228049755, Test loss: 0.01075370330363512\n", + "Starting epoch 3547\n", + "Train loss: 0.013350203391164541, Test loss: 0.0116750942543149\n", + "Starting epoch 3548\n", + "Train loss: 0.0131878755800426, Test loss: 0.011148634366691113\n", + "Starting epoch 3549\n", + "Train loss: 0.013313258234411478, Test loss: 0.012908310629427433\n", + "Starting epoch 3550\n", + "Train loss: 0.013271859940141439, Test loss: 0.01433142926543951\n", + "Starting epoch 3551\n", + "Train loss: 0.013257635347545147, Test loss: 0.010845904238522053\n", + "Starting epoch 3552\n", + "Train loss: 0.01332414673641324, Test loss: 0.014435079880058765\n", + "Starting epoch 3553\n", + "Train loss: 0.01325780488550663, Test loss: 0.012143295258283615\n", + "Starting epoch 3554\n", + "Train loss: 0.01327110879123211, Test loss: 0.0122596500441432\n", + "Starting epoch 3555\n", + "Train loss: 0.013145771734416485, Test loss: 0.013280746527016163\n", + "Starting epoch 3556\n", + "Train loss: 0.013307900205254554, Test loss: 0.012432614341378212\n", + "Starting epoch 3557\n", + "Train loss: 0.013316329885274172, Test loss: 0.011485022492706776\n", + "Starting epoch 3558\n", + "Train loss: 0.013197026997804642, Test loss: 0.012286732904613018\n", + "Starting epoch 3559\n", + "Train loss: 0.013338871169835329, Test loss: 0.01287867408245802\n", + "Starting epoch 3560\n", + "Train loss: 0.013300609681755305, Test loss: 0.011702093295753002\n", + "Starting epoch 3561\n", + "Train loss: 0.013291162792593242, Test loss: 0.01652628742158413\n", + "Starting epoch 3562\n", + "Train loss: 0.01329588808119297, Test loss: 0.013889821246266365\n", + "Starting epoch 3563\n", + "Train loss: 0.013299279846251011, Test loss: 0.022147560492157936\n", + "Starting epoch 3564\n", + "Train loss: 0.013301372230052948, Test loss: 0.02358068712055683\n", + "Starting epoch 3565\n", + "Train loss: 0.013319151885807514, Test loss: 0.023123880848288536\n", + "Starting epoch 3566\n", + "Train loss: 0.01322033142670989, Test loss: 0.024436360225081444\n", + "Starting epoch 3567\n", + "Train loss: 0.013262503053992987, Test loss: 0.020919756963849068\n", + "Starting epoch 3568\n", + "Train loss: 0.013261762373149395, Test loss: 0.024433007463812828\n", + "Starting epoch 3569\n", + "Train loss: 0.013225927967578172, Test loss: 0.02416042424738407\n", + "Starting epoch 3570\n", + "Train loss: 0.013353060483932495, Test loss: 0.022867226973176003\n", + "Starting epoch 3571\n", + "Train loss: 0.013326321709901095, Test loss: 0.026157202199101448\n", + "Starting epoch 3572\n", + "Train loss: 0.013332966286689042, Test loss: 0.02012723684310913\n", + "Starting epoch 3573\n", + "Train loss: 0.013286311775445939, Test loss: 0.020177336409687996\n", + "Starting epoch 3574\n", + "Train loss: 0.013250338584184647, Test loss: 0.012303929775953293\n", + "Starting epoch 3575\n", + "Train loss: 0.013275855556130409, Test loss: 0.021451789885759354\n", + "Starting epoch 3576\n", + "Train loss: 0.01331436900421977, Test loss: 0.02516581490635872\n", + "Starting epoch 3577\n", + "Train loss: 0.013294429313391446, Test loss: 0.019727695733308792\n", + "Starting epoch 3578\n", + "Train loss: 0.013189296144992113, Test loss: 0.01256806030869484\n", + "Starting epoch 3579\n", + "Train loss: 0.013281367626041174, Test loss: 0.012346269562840462\n", + "Starting epoch 3580\n", + "Train loss: 0.013309659212827682, Test loss: 0.012014782987535\n", + "Starting epoch 3581\n", + "Train loss: 0.013233197759836913, Test loss: 0.012667705304920673\n", + "Starting epoch 3582\n", + "Train loss: 0.013238874804228544, Test loss: 0.013188631273806095\n", + "Starting epoch 3583\n", + "Train loss: 0.013188012260943651, Test loss: 0.01557949185371399\n", + "Starting epoch 3584\n", + "Train loss: 0.01325343968346715, Test loss: 0.012369358912110329\n", + "Starting epoch 3585\n", + "Train loss: 0.013212852366268634, Test loss: 0.0152971176430583\n", + "Starting epoch 3586\n", + "Train loss: 0.013606268651783467, Test loss: 0.017999999225139618\n", + "Starting epoch 3587\n", + "Train loss: 0.013265460170805455, Test loss: 0.025317052379250526\n", + "Starting epoch 3588\n", + "Train loss: 0.013360684923827647, Test loss: 0.023221081122756004\n", + "Starting epoch 3589\n", + "Train loss: 0.013294013869017362, Test loss: 0.024476641789078712\n", + "Starting epoch 3590\n", + "Train loss: 0.013213484119623899, Test loss: 0.026144765317440033\n", + "Starting epoch 3591\n", + "Train loss: 0.01327802874147892, Test loss: 0.02118314616382122\n", + "Starting epoch 3592\n", + "Train loss: 0.013152612503618002, Test loss: 0.02692326530814171\n", + "Starting epoch 3593\n", + "Train loss: 0.01335844662040472, Test loss: 0.02296120673418045\n", + "Starting epoch 3594\n", + "Train loss: 0.013283744901418686, Test loss: 0.023005209863185883\n", + "Starting epoch 3595\n", + "Train loss: 0.013051281608641148, Test loss: 0.012593653984367847\n", + "Starting epoch 3596\n", + "Train loss: 0.013201035466045142, Test loss: 0.020247500389814377\n", + "Starting epoch 3597\n", + "Train loss: 0.013338160309940577, Test loss: 0.021917210891842842\n", + "Starting epoch 3598\n", + "Train loss: 0.013324497472494841, Test loss: 0.024480240419507027\n", + "Starting epoch 3599\n", + "Train loss: 0.013308639116585254, Test loss: 0.01867823116481304\n", + "Starting epoch 3600\n", + "Train loss: 0.013267541285604239, Test loss: 0.021116267889738083\n", + "Starting epoch 3601\n", + "Train loss: 0.013243664000183343, Test loss: 0.015343374572694302\n", + "Starting epoch 3602\n", + "Train loss: 0.013271381929516792, Test loss: 0.012269758619368076\n", + "Starting epoch 3603\n", + "Train loss: 0.013271864503622055, Test loss: 0.012285256758332253\n", + "Starting epoch 3604\n", + "Train loss: 0.013256303407251835, Test loss: 0.010474403388798237\n", + "Starting epoch 3605\n", + "Train loss: 0.013225727807730436, Test loss: 0.012379850260913372\n", + "Starting epoch 3606\n", + "Train loss: 0.013162191417068243, Test loss: 0.014043726958334446\n", + "Starting epoch 3607\n", + "Train loss: 0.013127949759364129, Test loss: 0.011589130386710167\n", + "Starting epoch 3608\n", + "Train loss: 0.013300370126962661, Test loss: 0.013725141063332558\n", + "Starting epoch 3609\n", + "Train loss: 0.013136610221117735, Test loss: 0.01323953177779913\n", + "Starting epoch 3610\n", + "Train loss: 0.013316760882735252, Test loss: 0.01220160722732544\n", + "Starting epoch 3611\n", + "Train loss: 0.013159729037433864, Test loss: 0.014854942448437214\n", + "Starting epoch 3612\n", + "Train loss: 0.01319576781243086, Test loss: 0.011989024467766285\n", + "Starting epoch 3613\n", + "Train loss: 0.013207624778151511, Test loss: 0.014873038977384567\n", + "Starting epoch 3614\n", + "Train loss: 0.01323362173512578, Test loss: 0.011027844622731209\n", + "Starting epoch 3615\n", + "Train loss: 0.013296848200261592, Test loss: 0.012960810214281082\n", + "Starting epoch 3616\n", + "Train loss: 0.013158081639558077, Test loss: 0.011255054734647274\n", + "Starting epoch 3617\n", + "Train loss: 0.013132957853376865, Test loss: 0.014423366636037827\n", + "Starting epoch 3618\n", + "Train loss: 0.013237122930586338, Test loss: 0.011898639611899853\n", + "Starting epoch 3619\n", + "Train loss: 0.013194218277931213, Test loss: 0.01312414463609457\n", + "Starting epoch 3620\n", + "Train loss: 0.013204587604850531, Test loss: 0.015342678874731064\n", + "Starting epoch 3621\n", + "Train loss: 0.01306989513337612, Test loss: 0.011980907991528511\n", + "Starting epoch 3622\n", + "Train loss: 0.01323176285251975, Test loss: 0.011844170279800892\n", + "Starting epoch 3623\n", + "Train loss: 0.01311499997973442, Test loss: 0.012074234895408154\n", + "Starting epoch 3624\n", + "Train loss: 0.013202304914593697, Test loss: 0.011823431588709354\n", + "Starting epoch 3625\n", + "Train loss: 0.013095147758722305, Test loss: 0.013264736160635948\n", + "Starting epoch 3626\n", + "Train loss: 0.013144830074161291, Test loss: 0.012997708283364773\n", + "Starting epoch 3627\n", + "Train loss: 0.01317428845912218, Test loss: 0.01261117309331894\n", + "Starting epoch 3628\n", + "Train loss: 0.013147207386791706, Test loss: 0.011490057222545147\n", + "Starting epoch 3629\n", + "Train loss: 0.013106388710439206, Test loss: 0.01306159421801567\n", + "Starting epoch 3630\n", + "Train loss: 0.013163313642144204, Test loss: 0.015892034396529198\n", + "Starting epoch 3631\n", + "Train loss: 0.013069164156913757, Test loss: 0.01141270063817501\n", + "Starting epoch 3632\n", + "Train loss: 0.013135354202240705, Test loss: 0.012840750627219677\n", + "Starting epoch 3633\n", + "Train loss: 0.01314731664955616, Test loss: 0.010046250186860561\n", + "Starting epoch 3634\n", + "Train loss: 0.012970473971217871, Test loss: 0.014035550877451897\n", + "Starting epoch 3635\n", + "Train loss: 0.013189703654497863, Test loss: 0.01386790256947279\n", + "Starting epoch 3636\n", + "Train loss: 0.013169575314968825, Test loss: 0.011282789520919323\n", + "Starting epoch 3637\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.013150746356695891, Test loss: 0.012679354287683964\n", + "Starting epoch 3638\n", + "Train loss: 0.013188408631831407, Test loss: 0.012649554759263992\n", + "Starting epoch 3639\n", + "Train loss: 0.01313577862456441, Test loss: 0.01223849132657051\n", + "Starting epoch 3640\n", + "Train loss: 0.013193527162075043, Test loss: 0.011041113175451756\n", + "Starting epoch 3641\n", + "Train loss: 0.013183900471776724, Test loss: 0.02420116402208805\n", + "Starting epoch 3642\n", + "Train loss: 0.013165861126035452, Test loss: 0.02570091187953949\n", + "Starting epoch 3643\n", + "Train loss: 0.013131530731916427, Test loss: 0.024538086727261543\n", + "Starting epoch 3644\n", + "Train loss: 0.013104991652071477, Test loss: 0.014269895851612091\n", + "Starting epoch 3645\n", + "Train loss: 0.013113091662526132, Test loss: 0.014228749088943005\n", + "Starting epoch 3646\n", + "Train loss: 0.013123397659510374, Test loss: 0.01130144577473402\n", + "Starting epoch 3647\n", + "Train loss: 0.013081738967448473, Test loss: 0.012177931144833565\n", + "Starting epoch 3648\n", + "Train loss: 0.013162148855626584, Test loss: 0.011104156263172626\n", + "Starting epoch 3649\n", + "Train loss: 0.013175705205649138, Test loss: 0.019373444840312004\n", + "Starting epoch 3650\n", + "Train loss: 0.013044760096818209, Test loss: 0.01976810209453106\n", + "Starting epoch 3651\n", + "Train loss: 0.013174432031810283, Test loss: 0.024492578580975533\n", + "Starting epoch 3652\n", + "Train loss: 0.013070577159523963, Test loss: 0.018789498135447502\n", + "Starting epoch 3653\n", + "Train loss: 0.013137426637113094, Test loss: 0.012990977615118027\n", + "Starting epoch 3654\n", + "Train loss: 0.013190155737102032, Test loss: 0.013297028839588165\n", + "Starting epoch 3655\n", + "Train loss: 0.013118802551180125, Test loss: 0.012720080092549324\n", + "Starting epoch 3656\n", + "Train loss: 0.01310454063117504, Test loss: 0.013458089902997017\n", + "Starting epoch 3657\n", + "Train loss: 0.013125271722674369, Test loss: 0.013925113715231419\n", + "Starting epoch 3658\n", + "Train loss: 0.013116787187755108, Test loss: 0.012497534044086933\n", + "Starting epoch 3659\n", + "Train loss: 0.01309725059196353, Test loss: 0.016318093985319138\n", + "Starting epoch 3660\n", + "Train loss: 0.013159309606999159, Test loss: 0.013067929074168205\n", + "Starting epoch 3661\n", + "Train loss: 0.013085055034607648, Test loss: 0.013413889333605766\n", + "Starting epoch 3662\n", + "Train loss: 0.013057738430798055, Test loss: 0.011316709220409393\n", + "Starting epoch 3663\n", + "Train loss: 0.012993525192141532, Test loss: 0.015696397051215172\n", + "Starting epoch 3664\n", + "Train loss: 0.013149548750370742, Test loss: 0.01273402664810419\n", + "Starting epoch 3665\n", + "Train loss: 0.013101261276751756, Test loss: 0.011267077177762985\n", + "Starting epoch 3666\n", + "Train loss: 0.013093946296721696, Test loss: 0.011167899705469608\n", + "Starting epoch 3667\n", + "Train loss: 0.013121316246688365, Test loss: 0.013075209222733974\n", + "Starting epoch 3668\n", + "Train loss: 0.01306356368586421, Test loss: 0.01152009703218937\n", + "Starting epoch 3669\n", + "Train loss: 0.013058234211057425, Test loss: 0.013754811137914658\n", + "Starting epoch 3670\n", + "Train loss: 0.013073186986148358, Test loss: 0.021634001284837723\n", + "Starting epoch 3671\n", + "Train loss: 0.013129847943782807, Test loss: 0.02133690007030964\n", + "Starting epoch 3672\n", + "Train loss: 0.013056381829082966, Test loss: 0.022047294303774834\n", + "Starting epoch 3673\n", + "Train loss: 0.013099164981395007, Test loss: 0.026053540408611298\n", + "Starting epoch 3674\n", + "Train loss: 0.013039432689547539, Test loss: 0.02150741219520569\n", + "Starting epoch 3675\n", + "Train loss: 0.013135619163513183, Test loss: 0.023486627265810966\n", + "Starting epoch 3676\n", + "Train loss: 0.013159415610134602, Test loss: 0.02078680694103241\n", + "Starting epoch 3677\n", + "Train loss: 0.013092152699828148, Test loss: 0.01986326463520527\n", + "Starting epoch 3678\n", + "Train loss: 0.013120681531727314, Test loss: 0.02815162017941475\n", + "Starting epoch 3679\n", + "Train loss: 0.013132000807672739, Test loss: 0.021805688738822937\n", + "Starting epoch 3680\n", + "Train loss: 0.013106225319206714, Test loss: 0.020245090126991272\n", + "Starting epoch 3681\n", + "Train loss: 0.013057848513126373, Test loss: 0.022265350446105003\n", + "Starting epoch 3682\n", + "Train loss: 0.013013533130288125, Test loss: 0.02306399866938591\n", + "Starting epoch 3683\n", + "Train loss: 0.013081227634102107, Test loss: 0.024712448939681053\n", + "Starting epoch 3684\n", + "Train loss: 0.013051316924393177, Test loss: 0.026243003085255623\n", + "Starting epoch 3685\n", + "Train loss: 0.013049987833946943, Test loss: 0.021410556510090828\n", + "Starting epoch 3686\n", + "Train loss: 0.013028282430022955, Test loss: 0.012483512982726097\n", + "Starting epoch 3687\n", + "Train loss: 0.013007932137697934, Test loss: 0.012417292222380638\n", + "Starting epoch 3688\n", + "Train loss: 0.012937184777110816, Test loss: 0.015738263726234436\n", + "Starting epoch 3689\n", + "Train loss: 0.013043813761323691, Test loss: 0.01209800411015749\n", + "Starting epoch 3690\n", + "Train loss: 0.012982289269566537, Test loss: 0.013733958825469017\n", + "Starting epoch 3691\n", + "Train loss: 0.013045012820512056, Test loss: 0.011394095607101917\n", + "Starting epoch 3692\n", + "Train loss: 0.013035851623862981, Test loss: 0.012402888387441635\n", + "Starting epoch 3693\n", + "Train loss: 0.012996829841285943, Test loss: 0.011374243535101414\n", + "Starting epoch 3694\n", + "Train loss: 0.013029064312577247, Test loss: 0.0135142607614398\n", + "Starting epoch 3695\n", + "Train loss: 0.013077595736831427, Test loss: 0.012762948870658875\n", + "Starting epoch 3696\n", + "Train loss: 0.01305992079898715, Test loss: 0.011392835527658463\n", + "Starting epoch 3697\n", + "Train loss: 0.013035020213574172, Test loss: 0.011828822083771229\n", + "Starting epoch 3698\n", + "Train loss: 0.01294488124549389, Test loss: 0.015267688781023026\n", + "Starting epoch 3699\n", + "Train loss: 0.01306537326425314, Test loss: 0.013369518332183361\n", + "Starting epoch 3700\n", + "Train loss: 0.013012555968016386, Test loss: 0.01139822881668806\n", + "Starting epoch 3701\n", + "Train loss: 0.012960879504680634, Test loss: 0.012094772420823574\n", + "Starting epoch 3702\n", + "Train loss: 0.013030624687671662, Test loss: 0.012232719920575619\n", + "Starting epoch 3703\n", + "Train loss: 0.012962290141731501, Test loss: 0.011303138919174671\n", + "Starting epoch 3704\n", + "Train loss: 0.012864169944077731, Test loss: 0.011621125973761082\n", + "Starting epoch 3705\n", + "Train loss: 0.013086894191801549, Test loss: 0.013140259310603142\n", + "Starting epoch 3706\n", + "Train loss: 0.013004161771386863, Test loss: 0.011134877800941467\n", + "Starting epoch 3707\n", + "Train loss: 0.012991771027445793, Test loss: 0.010982468724250793\n", + "Starting epoch 3708\n", + "Train loss: 0.012932454999536276, Test loss: 0.012611130252480507\n", + "Starting epoch 3709\n", + "Train loss: 0.01303456723690033, Test loss: 0.012794241309165955\n", + "Starting epoch 3710\n", + "Train loss: 0.012993928007781506, Test loss: 0.02279733121395111\n", + "Starting epoch 3711\n", + "Train loss: 0.012981279343366623, Test loss: 0.02373143658041954\n", + "Starting epoch 3712\n", + "Train loss: 0.013024427834898234, Test loss: 0.02101159654557705\n", + "Starting epoch 3713\n", + "Train loss: 0.012916411962360144, Test loss: 0.02646491676568985\n", + "Starting epoch 3714\n", + "Train loss: 0.013075004126876592, Test loss: 0.02255951426923275\n", + "Starting epoch 3715\n", + "Train loss: 0.013073559273034335, Test loss: 0.02282567508518696\n", + "Starting epoch 3716\n", + "Train loss: 0.013025432284921408, Test loss: 0.019310515373945236\n", + "Starting epoch 3717\n", + "Train loss: 0.012950685769319535, Test loss: 0.02091917395591736\n", + "Starting epoch 3718\n", + "Train loss: 0.013044619560241699, Test loss: 0.026410337537527084\n", + "Starting epoch 3719\n", + "Train loss: 0.012964948434382678, Test loss: 0.011513705365359783\n", + "Starting epoch 3720\n", + "Train loss: 0.013034376557916403, Test loss: 0.01390912476927042\n", + "Starting epoch 3721\n", + "Train loss: 0.013052612449973822, Test loss: 0.013980219140648842\n", + "Starting epoch 3722\n", + "Train loss: 0.012966084610670805, Test loss: 0.012136523611843586\n", + "Starting epoch 3723\n", + "Train loss: 0.012988159768283367, Test loss: 0.012643025256693363\n", + "Starting epoch 3724\n", + "Train loss: 0.012996894400566816, Test loss: 0.010877964086830616\n", + "Starting epoch 3725\n", + "Train loss: 0.012982883285731077, Test loss: 0.012617806904017925\n", + "Starting epoch 3726\n", + "Train loss: 0.013104824963957071, Test loss: 0.014757209457457066\n", + "Starting epoch 3727\n", + "Train loss: 0.012944026049226523, Test loss: 0.011862034909427166\n", + "Starting epoch 3728\n", + "Train loss: 0.01303378963842988, Test loss: 0.014308453537523746\n", + "Starting epoch 3729\n", + "Train loss: 0.013024155031889677, Test loss: 0.013125295750796795\n", + "Starting epoch 3730\n", + "Train loss: 0.013008346557617187, Test loss: 0.011940866708755493\n", + "Starting epoch 3731\n", + "Train loss: 0.012951722219586372, Test loss: 0.011626286432147026\n", + "Starting epoch 3732\n", + "Train loss: 0.012891633305698633, Test loss: 0.01188086997717619\n", + "Starting epoch 3733\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.012887436039745808, Test loss: 0.02429017797112465\n", + "Starting epoch 3734\n", + "Train loss: 0.012967998646199703, Test loss: 0.023975178599357605\n", + "Starting epoch 3735\n", + "Train loss: 0.012995802853256464, Test loss: 0.01874583587050438\n", + "Starting epoch 3736\n", + "Train loss: 0.012954978663474321, Test loss: 0.021374840289354324\n", + "Starting epoch 3737\n", + "Train loss: 0.01297105386853218, Test loss: 0.018084267154335976\n", + "Starting epoch 3738\n", + "Train loss: 0.01291556291282177, Test loss: 0.022079436108469963\n", + "Starting epoch 3739\n", + "Train loss: 0.0129868371412158, Test loss: 0.02220296487212181\n", + "Starting epoch 3740\n", + "Train loss: 0.012992751896381377, Test loss: 0.023137520998716354\n", + "Starting epoch 3741\n", + "Train loss: 0.012881197519600391, Test loss: 0.02358568273484707\n", + "Starting epoch 3742\n", + "Train loss: 0.012996761985123157, Test loss: 0.02077941969037056\n", + "Starting epoch 3743\n", + "Train loss: 0.012986119519919157, Test loss: 0.02379182167351246\n", + "Starting epoch 3744\n", + "Train loss: 0.012955863792449236, Test loss: 0.01925939880311489\n", + "Starting epoch 3745\n", + "Train loss: 0.012932484857738019, Test loss: 0.023538758978247643\n", + "Starting epoch 3746\n", + "Train loss: 0.012923690751194953, Test loss: 0.020676637068390846\n", + "Starting epoch 3747\n", + "Train loss: 0.01297009428963065, Test loss: 0.022727614268660545\n", + "Starting epoch 3748\n", + "Train loss: 0.012999224793165922, Test loss: 0.02300065942108631\n", + "Starting epoch 3749\n", + "Train loss: 0.012851500120013952, Test loss: 0.02335510030388832\n", + "Starting epoch 3750\n", + "Train loss: 0.01294146578758955, Test loss: 0.02037937566637993\n", + "Starting epoch 3751\n", + "Train loss: 0.01290824929252267, Test loss: 0.019759761169552803\n", + "Starting epoch 3752\n", + "Train loss: 0.012863334380090236, Test loss: 0.013393745757639408\n", + "Starting epoch 3753\n", + "Train loss: 0.012996587213128805, Test loss: 0.011411095038056374\n", + "Starting epoch 3754\n", + "Train loss: 0.01289223289117217, Test loss: 0.020982956513762474\n", + "Starting epoch 3755\n", + "Train loss: 0.012988001201301813, Test loss: 0.022138230502605438\n", + "Starting epoch 3756\n", + "Train loss: 0.012828451953828335, Test loss: 0.01958441734313965\n", + "Starting epoch 3757\n", + "Train loss: 0.012978809364140033, Test loss: 0.02005966752767563\n", + "Starting epoch 3758\n", + "Train loss: 0.012978776637464762, Test loss: 0.021259069442749023\n", + "Starting epoch 3759\n", + "Train loss: 0.012940148953348399, Test loss: 0.024374866858124733\n", + "Starting epoch 3760\n", + "Train loss: 0.012926922105252743, Test loss: 0.021930627524852753\n", + "Starting epoch 3761\n", + "Train loss: 0.012820493821054697, Test loss: 0.01999376155436039\n", + "Starting epoch 3762\n", + "Train loss: 0.012969793658703566, Test loss: 0.02196149155497551\n", + "Starting epoch 3763\n", + "Train loss: 0.012940591108053923, Test loss: 0.021349942311644554\n", + "Starting epoch 3764\n", + "Train loss: 0.012860567588359118, Test loss: 0.021895674988627434\n", + "Starting epoch 3765\n", + "Train loss: 0.012860599011182784, Test loss: 0.020986344665288925\n", + "Starting epoch 3766\n", + "Train loss: 0.012950074095278978, Test loss: 0.015807446092367172\n", + "Starting epoch 3767\n", + "Train loss: 0.012756428960710765, Test loss: 0.010645431466400623\n", + "Starting epoch 3768\n", + "Train loss: 0.012934792153537274, Test loss: 0.0124046104028821\n", + "Starting epoch 3769\n", + "Train loss: 0.013015973549336195, Test loss: 0.014413759112358093\n", + "Starting epoch 3770\n", + "Train loss: 0.012926234807819128, Test loss: 0.01383663434535265\n", + "Starting epoch 3771\n", + "Train loss: 0.012985311336815358, Test loss: 0.011764397844672203\n", + "Starting epoch 3772\n", + "Train loss: 0.01290308628231287, Test loss: 0.02207561768591404\n", + "Starting epoch 3773\n", + "Train loss: 0.01293919961899519, Test loss: 0.024988986551761627\n", + "Starting epoch 3774\n", + "Train loss: 0.012928182277828455, Test loss: 0.02494179457426071\n", + "Starting epoch 3775\n", + "Train loss: 0.01293093228712678, Test loss: 0.02100525051355362\n", + "Starting epoch 3776\n", + "Train loss: 0.012833117991685867, Test loss: 0.02021215111017227\n", + "Starting epoch 3777\n", + "Train loss: 0.012957644686102866, Test loss: 0.018803749233484268\n", + "Starting epoch 3778\n", + "Train loss: 0.012919951565563678, Test loss: 0.017411744222044945\n", + "Starting epoch 3779\n", + "Train loss: 0.012883013654500246, Test loss: 0.021485481411218643\n", + "Starting epoch 3780\n", + "Train loss: 0.012858593743294478, Test loss: 0.0255418848246336\n", + "Starting epoch 3781\n", + "Train loss: 0.01285918703302741, Test loss: 0.023193085566163063\n", + "Starting epoch 3782\n", + "Train loss: 0.012879945952445269, Test loss: 0.02363305166363716\n", + "Starting epoch 3783\n", + "Train loss: 0.013022334724664688, Test loss: 0.02567955106496811\n", + "Starting epoch 3784\n", + "Train loss: 0.01289935752749443, Test loss: 0.01994982548058033\n", + "Starting epoch 3785\n", + "Train loss: 0.012946089394390584, Test loss: 0.01727941632270813\n", + "Starting epoch 3786\n", + "Train loss: 0.012934663984924554, Test loss: 0.024169422686100006\n", + "Starting epoch 3787\n", + "Train loss: 0.012891848627477885, Test loss: 0.02152138389647007\n", + "Starting epoch 3788\n", + "Train loss: 0.012914193663746118, Test loss: 0.021499324589967728\n", + "Starting epoch 3789\n", + "Train loss: 0.012849923335015773, Test loss: 0.017627667635679245\n", + "Starting epoch 3790\n", + "Train loss: 0.012869784776121378, Test loss: 0.020973868668079376\n", + "Starting epoch 3791\n", + "Train loss: 0.012888461463153362, Test loss: 0.021603262051939964\n", + "Starting epoch 3792\n", + "Train loss: 0.012905104514211417, Test loss: 0.023945046588778496\n", + "Starting epoch 3793\n", + "Train loss: 0.012842301949858666, Test loss: 0.025414206087589264\n", + "Starting epoch 3794\n", + "Train loss: 0.012802008111029863, Test loss: 0.017817357555031776\n", + "Starting epoch 3795\n", + "Train loss: 0.012827161625027656, Test loss: 0.01881764642894268\n", + "Starting epoch 3796\n", + "Train loss: 0.012819501850754023, Test loss: 0.019474241882562637\n", + "Starting epoch 3797\n", + "Train loss: 0.012830081041902303, Test loss: 0.023882634937763214\n", + "Starting epoch 3798\n", + "Train loss: 0.012771834321320057, Test loss: 0.025481857359409332\n", + "Starting epoch 3799\n", + "Train loss: 0.012821453642100095, Test loss: 0.019027991220355034\n", + "Starting epoch 3800\n", + "Train loss: 0.012824984937906266, Test loss: 0.02375430427491665\n", + "Starting epoch 3801\n", + "Train loss: 0.01283326804637909, Test loss: 0.02075796015560627\n", + "Starting epoch 3802\n", + "Train loss: 0.012778622657060623, Test loss: 0.011633841320872307\n", + "Starting epoch 3803\n", + "Train loss: 0.012883927524089813, Test loss: 0.01204104907810688\n", + "Starting epoch 3804\n", + "Train loss: 0.01287522479891777, Test loss: 0.013310062699019909\n", + "Starting epoch 3805\n", + "Train loss: 0.012833264768123626, Test loss: 0.01384373102337122\n", + "Starting epoch 3806\n", + "Train loss: 0.012845367193222046, Test loss: 0.013198827393352985\n", + "Starting epoch 3807\n", + "Train loss: 0.012734828479588033, Test loss: 0.026164736598730087\n", + "Starting epoch 3808\n", + "Train loss: 0.012848141733556986, Test loss: 0.02052200771868229\n", + "Starting epoch 3809\n", + "Train loss: 0.012842859271913767, Test loss: 0.018336951732635498\n", + "Starting epoch 3810\n", + "Train loss: 0.012878221217542887, Test loss: 0.02227342315018177\n", + "Starting epoch 3811\n", + "Train loss: 0.01291366558521986, Test loss: 0.024536672979593277\n", + "Starting epoch 3812\n", + "Train loss: 0.012958709374070168, Test loss: 0.021730367094278336\n", + "Starting epoch 3813\n", + "Train loss: 0.012893969360738992, Test loss: 0.021862845867872238\n", + "Starting epoch 3814\n", + "Train loss: 0.012863488458096982, Test loss: 0.023560622707009315\n", + "Starting epoch 3815\n", + "Train loss: 0.012922985181212425, Test loss: 0.021677788347005844\n", + "Starting epoch 3816\n", + "Train loss: 0.012745339088141919, Test loss: 0.022365543991327286\n", + "Starting epoch 3817\n", + "Train loss: 0.012872582916170359, Test loss: 0.021941829472780228\n", + "Starting epoch 3818\n", + "Train loss: 0.012829661443829537, Test loss: 0.011053766123950481\n", + "Starting epoch 3819\n", + "Train loss: 0.01285547798499465, Test loss: 0.012178720906376839\n", + "Starting epoch 3820\n", + "Train loss: 0.01280987162142992, Test loss: 0.012218085117638111\n", + "Starting epoch 3821\n", + "Train loss: 0.012808237280696631, Test loss: 0.01127533707767725\n", + "Starting epoch 3822\n", + "Train loss: 0.012814556993544102, Test loss: 0.01238689199090004\n", + "Starting epoch 3823\n", + "Train loss: 0.012804133258759976, Test loss: 0.012405597604811192\n", + "Starting epoch 3824\n", + "Train loss: 0.012849376946687698, Test loss: 0.011730262078344822\n", + "Starting epoch 3825\n", + "Train loss: 0.012793120443820954, Test loss: 0.010998348705470562\n", + "Starting epoch 3826\n", + "Train loss: 0.012776877321302891, Test loss: 0.010996672324836254\n", + "Starting epoch 3827\n", + "Train loss: 0.012693460155278445, Test loss: 0.011971861124038696\n", + "Starting epoch 3828\n", + "Train loss: 0.012761752028018237, Test loss: 0.011920087039470673\n", + "Starting epoch 3829\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.012885275147855282, Test loss: 0.011591486632823944\n", + "Starting epoch 3830\n", + "Train loss: 0.012808249555528164, Test loss: 0.01512700691819191\n", + "Starting epoch 3831\n", + "Train loss: 0.012703931387513877, Test loss: 0.01304203923791647\n", + "Starting epoch 3832\n", + "Train loss: 0.012743750736117363, Test loss: 0.012340307235717773\n", + "Starting epoch 3833\n", + "Train loss: 0.01264452513307333, Test loss: 0.01964285783469677\n", + "Starting epoch 3834\n", + "Train loss: 0.012801919057965279, Test loss: 0.010408572852611542\n", + "Starting epoch 3835\n", + "Train loss: 0.01296595012769103, Test loss: 0.011180424131453037\n", + "Starting epoch 3836\n", + "Train loss: 0.012955246921628714, Test loss: 0.010970769450068474\n", + "Starting epoch 3837\n", + "Train loss: 0.012854759283363819, Test loss: 0.010473039932549\n", + "Starting epoch 3838\n", + "Train loss: 0.012882254105061292, Test loss: 0.012047882191836834\n", + "Starting epoch 3839\n", + "Train loss: 0.012782099321484566, Test loss: 0.013917184434831142\n", + "Starting epoch 3840\n", + "Train loss: 0.012814018782228232, Test loss: 0.013013159856200218\n", + "Starting epoch 3841\n", + "Train loss: 0.012644166555255652, Test loss: 0.02244618535041809\n", + "Starting epoch 3842\n", + "Train loss: 0.012837427351623774, Test loss: 0.01903296820819378\n", + "Starting epoch 3843\n", + "Train loss: 0.012801307402551174, Test loss: 0.01907745935022831\n", + "Starting epoch 3844\n", + "Train loss: 0.012780016139149666, Test loss: 0.012165981344878674\n", + "Starting epoch 3845\n", + "Train loss: 0.012808099929243326, Test loss: 0.01029312051832676\n", + "Starting epoch 3846\n", + "Train loss: 0.012816536352038383, Test loss: 0.012951897457242012\n", + "Starting epoch 3847\n", + "Train loss: 0.012773477379232645, Test loss: 0.011570367030799389\n", + "Starting epoch 3848\n", + "Train loss: 0.012808756995946169, Test loss: 0.010938565246760845\n", + "Starting epoch 3849\n", + "Train loss: 0.012760524060577155, Test loss: 0.01168048195540905\n", + "Starting epoch 3850\n", + "Train loss: 0.012735169474035501, Test loss: 0.01161678321659565\n", + "Starting epoch 3851\n", + "Train loss: 0.01279797898605466, Test loss: 0.012508152984082699\n", + "Starting epoch 3852\n", + "Train loss: 0.012762661706656218, Test loss: 0.011902794241905212\n", + "Starting epoch 3853\n", + "Train loss: 0.01273499419912696, Test loss: 0.01089785248041153\n", + "Starting epoch 3854\n", + "Train loss: 0.01273355720564723, Test loss: 0.01144720520824194\n", + "Starting epoch 3855\n", + "Train loss: 0.012715933378785849, Test loss: 0.012417294085025787\n", + "Starting epoch 3856\n", + "Train loss: 0.012617971021682025, Test loss: 0.017653170973062515\n", + "Starting epoch 3857\n", + "Train loss: 0.012820610180497169, Test loss: 0.011857254430651665\n", + "Starting epoch 3858\n", + "Train loss: 0.012852127868682147, Test loss: 0.01192791573703289\n", + "Starting epoch 3859\n", + "Train loss: 0.012794441170990467, Test loss: 0.011521516367793083\n", + "Starting epoch 3860\n", + "Train loss: 0.012797421831637621, Test loss: 0.010904342867434025\n", + "Starting epoch 3861\n", + "Train loss: 0.012697028126567603, Test loss: 0.010776461102068424\n", + "Starting epoch 3862\n", + "Train loss: 0.01273736234754324, Test loss: 0.012115996330976486\n", + "Starting epoch 3863\n", + "Train loss: 0.012849950958043336, Test loss: 0.011546213179826736\n", + "Starting epoch 3864\n", + "Train loss: 0.012700869161635637, Test loss: 0.013694040477275848\n", + "Starting epoch 3865\n", + "Train loss: 0.012773370556533336, Test loss: 0.012549838051199913\n", + "Starting epoch 3866\n", + "Train loss: 0.012749117352068424, Test loss: 0.013771811500191689\n", + "Starting epoch 3867\n", + "Train loss: 0.012793762274086476, Test loss: 0.010421108454465866\n", + "Starting epoch 3868\n", + "Train loss: 0.012773276418447494, Test loss: 0.010672887787222862\n", + "Starting epoch 3869\n", + "Train loss: 0.012747350353747607, Test loss: 0.012757514603435993\n", + "Starting epoch 3870\n", + "Train loss: 0.01275077000260353, Test loss: 0.02236996591091156\n", + "Starting epoch 3871\n", + "Train loss: 0.012679363992065191, Test loss: 0.02213614247739315\n", + "Starting epoch 3872\n", + "Train loss: 0.012726790774613618, Test loss: 0.023518310859799385\n", + "Starting epoch 3873\n", + "Train loss: 0.012853393368422984, Test loss: 0.021317727863788605\n", + "Starting epoch 3874\n", + "Train loss: 0.012743402421474457, Test loss: 0.021588239818811417\n", + "Starting epoch 3875\n", + "Train loss: 0.012680294830352068, Test loss: 0.023731781169772148\n", + "Starting epoch 3876\n", + "Train loss: 0.012677702438086271, Test loss: 0.019237512722611427\n", + "Starting epoch 3877\n", + "Train loss: 0.012775804698467254, Test loss: 0.023710520938038826\n", + "Starting epoch 3878\n", + "Train loss: 0.012733970340341329, Test loss: 0.02369154803454876\n", + "Starting epoch 3879\n", + "Train loss: 0.012714519761502743, Test loss: 0.025257375091314316\n", + "Starting epoch 3880\n", + "Train loss: 0.012679664846509696, Test loss: 0.024016661569476128\n", + "Starting epoch 3881\n", + "Train loss: 0.012654507644474506, Test loss: 0.02157985232770443\n", + "Starting epoch 3882\n", + "Train loss: 0.012726070433855057, Test loss: 0.026521213352680206\n", + "Starting epoch 3883\n", + "Train loss: 0.0127237250469625, Test loss: 0.019401557743549347\n", + "Starting epoch 3884\n", + "Train loss: 0.012656443547457456, Test loss: 0.022879580035805702\n", + "Starting epoch 3885\n", + "Train loss: 0.012666776720434427, Test loss: 0.020141636952757835\n", + "Starting epoch 3886\n", + "Train loss: 0.012682035733014345, Test loss: 0.012231320142745972\n", + "Starting epoch 3887\n", + "Train loss: 0.01269340593367815, Test loss: 0.010610753670334816\n", + "Starting epoch 3888\n", + "Train loss: 0.012742123417556286, Test loss: 0.011013994924724102\n", + "Starting epoch 3889\n", + "Train loss: 0.012723609637469054, Test loss: 0.010491356253623962\n", + "Starting epoch 3890\n", + "Train loss: 0.012682890836149454, Test loss: 0.011663460172712803\n", + "Starting epoch 3891\n", + "Train loss: 0.01270724728703499, Test loss: 0.010755945928394794\n", + "Starting epoch 3892\n", + "Train loss: 0.012584949489682912, Test loss: 0.012802189216017723\n", + "Starting epoch 3893\n", + "Train loss: 0.012730153035372496, Test loss: 0.01076305191963911\n", + "Starting epoch 3894\n", + "Train loss: 0.012719064932316542, Test loss: 0.01412952784448862\n", + "Starting epoch 3895\n", + "Train loss: 0.012701338045299054, Test loss: 0.01140392292290926\n", + "Starting epoch 3896\n", + "Train loss: 0.012628145273774862, Test loss: 0.010830974206328392\n", + "Starting epoch 3897\n", + "Train loss: 0.012638351228088141, Test loss: 0.011645457707345486\n", + "Starting epoch 3898\n", + "Train loss: 0.012697852645069362, Test loss: 0.011286329478025436\n", + "Starting epoch 3899\n", + "Train loss: 0.012655172795057297, Test loss: 0.014409193769097328\n", + "Starting epoch 3900\n", + "Train loss: 0.012696002423763276, Test loss: 0.020328713580965996\n", + "Starting epoch 3901\n", + "Train loss: 0.012695803698152304, Test loss: 0.023958634585142136\n", + "Starting epoch 3902\n", + "Train loss: 0.012697221171110868, Test loss: 0.02247554622590542\n", + "Starting epoch 3903\n", + "Train loss: 0.012681065034121275, Test loss: 0.023961015045642853\n", + "Starting epoch 3904\n", + "Train loss: 0.012718631960451603, Test loss: 0.025743812322616577\n", + "Starting epoch 3905\n", + "Train loss: 0.012780430261045694, Test loss: 0.019510725513100624\n", + "Starting epoch 3906\n", + "Train loss: 0.012649216130375862, Test loss: 0.022845985367894173\n", + "Starting epoch 3907\n", + "Train loss: 0.012721797712147235, Test loss: 0.024121832102537155\n", + "Starting epoch 3908\n", + "Train loss: 0.012645488698035478, Test loss: 0.021482564508914948\n", + "Starting epoch 3909\n", + "Train loss: 0.012681981530040503, Test loss: 0.02144038677215576\n", + "Starting epoch 3910\n", + "Train loss: 0.01272297538816929, Test loss: 0.01794867031276226\n", + "Starting epoch 3911\n", + "Train loss: 0.012635877970606089, Test loss: 0.020641034469008446\n", + "Starting epoch 3912\n", + "Train loss: 0.01260305566713214, Test loss: 0.022855665534734726\n", + "Starting epoch 3913\n", + "Train loss: 0.012684116736054421, Test loss: 0.025798374786973\n", + "Starting epoch 3914\n", + "Train loss: 0.012630815543234348, Test loss: 0.019826790317893028\n", + "Starting epoch 3915\n", + "Train loss: 0.012632880359888077, Test loss: 0.024348370730876923\n", + "Starting epoch 3916\n", + "Train loss: 0.012673378568142652, Test loss: 0.018426749855279922\n", + "Starting epoch 3917\n", + "Train loss: 0.012561951782554388, Test loss: 0.021044578403234482\n", + "Starting epoch 3918\n", + "Train loss: 0.012675755806267262, Test loss: 0.019263464957475662\n", + "Starting epoch 3919\n", + "Train loss: 0.012661884259432554, Test loss: 0.011901509016752243\n", + "Starting epoch 3920\n", + "Train loss: 0.012624828983098269, Test loss: 0.010270426981151104\n", + "Starting epoch 3921\n", + "Train loss: 0.012611444722861052, Test loss: 0.010997715406119823\n", + "Starting epoch 3922\n", + "Train loss: 0.012633294239640236, Test loss: 0.015197329223155975\n", + "Starting epoch 3923\n", + "Train loss: 0.012615319229662419, Test loss: 0.013929333537817001\n", + "Starting epoch 3924\n", + "Train loss: 0.012512860596179962, Test loss: 0.016812629997730255\n", + "Starting epoch 3925\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.012588560432195663, Test loss: 0.02378152869641781\n", + "Starting epoch 3926\n", + "Train loss: 0.01265276739373803, Test loss: 0.0241993498057127\n", + "Starting epoch 3927\n", + "Train loss: 0.012514801118522883, Test loss: 0.02314268983900547\n", + "Starting epoch 3928\n", + "Train loss: 0.012589294463396072, Test loss: 0.01736561395227909\n", + "Starting epoch 3929\n", + "Train loss: 0.012586892135441303, Test loss: 0.024611156433820724\n", + "Starting epoch 3930\n", + "Train loss: 0.01258027384057641, Test loss: 0.02318967692553997\n", + "Starting epoch 3931\n", + "Train loss: 0.012562333699315786, Test loss: 0.018140461295843124\n", + "Starting epoch 3932\n", + "Train loss: 0.01260448932647705, Test loss: 0.023144572973251343\n", + "Starting epoch 3933\n", + "Train loss: 0.012597542274743318, Test loss: 0.019685804843902588\n", + "Starting epoch 3934\n", + "Train loss: 0.012557029612362386, Test loss: 0.020464153960347176\n", + "Starting epoch 3935\n", + "Train loss: 0.012505627423524856, Test loss: 0.022006023675203323\n", + "Starting epoch 3936\n", + "Train loss: 0.012613937817513942, Test loss: 0.02229301445186138\n", + "Starting epoch 3937\n", + "Train loss: 0.012573058400303126, Test loss: 0.021281199529767036\n", + "Starting epoch 3938\n", + "Train loss: 0.012594098374247551, Test loss: 0.020238984376192093\n", + "Starting epoch 3939\n", + "Train loss: 0.012549642939120531, Test loss: 0.016327327117323875\n", + "Starting epoch 3940\n", + "Train loss: 0.012545117381960154, Test loss: 0.013385358266532421\n", + "Starting epoch 3941\n", + "Train loss: 0.012597385533154011, Test loss: 0.010150354355573654\n", + "Starting epoch 3942\n", + "Train loss: 0.012595944497734308, Test loss: 0.010374512523412704\n", + "Starting epoch 3943\n", + "Train loss: 0.012586619183421134, Test loss: 0.011958127841353416\n", + "Starting epoch 3944\n", + "Train loss: 0.012612121999263764, Test loss: 0.01322796382009983\n", + "Starting epoch 3945\n", + "Train loss: 0.012621881049126387, Test loss: 0.009870518930256367\n", + "Starting epoch 3946\n", + "Train loss: 0.012678558938205242, Test loss: 0.011789229698479176\n", + "Starting epoch 3947\n", + "Train loss: 0.012616399601101875, Test loss: 0.011623278260231018\n", + "Starting epoch 3948\n", + "Train loss: 0.01258748373016715, Test loss: 0.01474025472998619\n", + "Starting epoch 3949\n", + "Train loss: 0.012512836549431085, Test loss: 0.010980590246617794\n", + "Starting epoch 3950\n", + "Train loss: 0.012525669876486062, Test loss: 0.0188776645809412\n", + "Starting epoch 3951\n", + "Train loss: 0.012624240219593049, Test loss: 0.022251231595873833\n", + "Starting epoch 3952\n", + "Train loss: 0.012736153807491064, Test loss: 0.022029174491763115\n", + "Starting epoch 3953\n", + "Train loss: 0.012663949392735957, Test loss: 0.024468809366226196\n", + "Starting epoch 3954\n", + "Train loss: 0.012604997102171183, Test loss: 0.020278673619031906\n", + "Starting epoch 3955\n", + "Train loss: 0.012527079172432422, Test loss: 0.0258454829454422\n", + "Starting epoch 3956\n", + "Train loss: 0.012645169608294964, Test loss: 0.019672753289341927\n", + "Starting epoch 3957\n", + "Train loss: 0.012573780491948128, Test loss: 0.020595449954271317\n", + "Starting epoch 3958\n", + "Train loss: 0.012521494533866645, Test loss: 0.01138038095086813\n", + "Starting epoch 3959\n", + "Train loss: 0.012469762731343507, Test loss: 0.011845840141177177\n", + "Starting epoch 3960\n", + "Train loss: 0.012587469518184662, Test loss: 0.023709194734692574\n", + "Starting epoch 3961\n", + "Train loss: 0.01261678233742714, Test loss: 0.019948137924075127\n", + "Starting epoch 3962\n", + "Train loss: 0.012585444543510675, Test loss: 0.020217381417751312\n", + "Starting epoch 3963\n", + "Train loss: 0.012558044008910656, Test loss: 0.02249699831008911\n", + "Starting epoch 3964\n", + "Train loss: 0.01257966760545969, Test loss: 0.022180598229169846\n", + "Starting epoch 3965\n", + "Train loss: 0.012568999268114566, Test loss: 0.022943897172808647\n", + "Starting epoch 3966\n", + "Train loss: 0.01259286118671298, Test loss: 0.02102581039071083\n", + "Starting epoch 3967\n", + "Train loss: 0.012617982607334852, Test loss: 0.022173427045345306\n", + "Starting epoch 3968\n", + "Train loss: 0.01251640686765313, Test loss: 0.022667916491627693\n", + "Starting epoch 3969\n", + "Train loss: 0.012522392682731151, Test loss: 0.0219822246581316\n", + "Starting epoch 3970\n", + "Train loss: 0.01251206960529089, Test loss: 0.019472837448120117\n", + "Starting epoch 3971\n", + "Train loss: 0.012524790819734335, Test loss: 0.02353503927588463\n", + "Starting epoch 3972\n", + "Train loss: 0.012542810197919608, Test loss: 0.02062343806028366\n", + "Starting epoch 3973\n", + "Train loss: 0.012578121032565831, Test loss: 0.024593720212578773\n", + "Starting epoch 3974\n", + "Train loss: 0.01252018142491579, Test loss: 0.027399299666285515\n", + "Starting epoch 3975\n", + "Train loss: 0.012512041218578816, Test loss: 0.017931459471583366\n", + "Starting epoch 3976\n", + "Train loss: 0.01250979671254754, Test loss: 0.012022041715681553\n", + "Starting epoch 3977\n", + "Train loss: 0.012543076612055302, Test loss: 0.011964036151766777\n", + "Starting epoch 3978\n", + "Train loss: 0.012510711681097747, Test loss: 0.012733603827655315\n", + "Starting epoch 3979\n", + "Train loss: 0.012585159670561553, Test loss: 0.01106464397162199\n", + "Starting epoch 3980\n", + "Train loss: 0.012568626180291175, Test loss: 0.0114701297134161\n", + "Starting epoch 3981\n", + "Train loss: 0.012461198326200247, Test loss: 0.01331628393381834\n", + "Starting epoch 3982\n", + "Train loss: 0.012543524838984013, Test loss: 0.012425352819263935\n", + "Starting epoch 3983\n", + "Train loss: 0.012559044323861599, Test loss: 0.010915732942521572\n", + "Starting epoch 3984\n", + "Train loss: 0.012574345152825118, Test loss: 0.012113135308027267\n", + "Starting epoch 3985\n", + "Train loss: 0.012509987745434046, Test loss: 0.014384621754288673\n", + "Starting epoch 3986\n", + "Train loss: 0.01254208842292428, Test loss: 0.011325226165354252\n", + "Starting epoch 3987\n", + "Train loss: 0.01252772957086563, Test loss: 0.012801401317119598\n", + "Starting epoch 3988\n", + "Train loss: 0.012470713667571545, Test loss: 0.012456434778869152\n", + "Starting epoch 3989\n", + "Train loss: 0.012477610092610121, Test loss: 0.011458031833171844\n", + "Starting epoch 3990\n", + "Train loss: 0.012536625135689974, Test loss: 0.011116153560578823\n", + "Starting epoch 3991\n", + "Train loss: 0.012459693029522895, Test loss: 0.011306765489280224\n", + "Starting epoch 3992\n", + "Train loss: 0.012430169973522425, Test loss: 0.012485429644584656\n", + "Starting epoch 3993\n", + "Train loss: 0.012463799249380826, Test loss: 0.01055150292813778\n", + "Starting epoch 3994\n", + "Train loss: 0.01251645153388381, Test loss: 0.010594042018055916\n", + "Starting epoch 3995\n", + "Train loss: 0.01256668908521533, Test loss: 0.01071471069008112\n", + "Starting epoch 3996\n", + "Train loss: 0.012443311382085084, Test loss: 0.011738799512386322\n", + "Starting epoch 3997\n", + "Train loss: 0.01240094618871808, Test loss: 0.025409772992134094\n", + "Starting epoch 3998\n", + "Train loss: 0.012557370699942111, Test loss: 0.023851048201322556\n", + "Starting epoch 3999\n", + "Train loss: 0.012531602382659912, Test loss: 0.018389025703072548\n", + "Starting epoch 4000\n", + "Train loss: 0.012524468675255775, Test loss: 0.019912337884306908\n", + "Starting epoch 4001\n", + "Train loss: 0.012526907399296761, Test loss: 0.024791080504655838\n", + "Starting epoch 4002\n", + "Train loss: 0.01254219725728035, Test loss: 0.02333439327776432\n", + "Starting epoch 4003\n", + "Train loss: 0.012548072896897793, Test loss: 0.01987597532570362\n", + "Starting epoch 4004\n", + "Train loss: 0.012481770310550928, Test loss: 0.022373123094439507\n", + "Starting epoch 4005\n", + "Train loss: 0.01242056541144848, Test loss: 0.022111639380455017\n", + "Starting epoch 4006\n", + "Train loss: 0.012566431444138288, Test loss: 0.021137846633791924\n", + "Starting epoch 4007\n", + "Train loss: 0.012523970510810613, Test loss: 0.021652694791555405\n", + "Starting epoch 4008\n", + "Train loss: 0.01246638996526599, Test loss: 0.019253958016633987\n", + "Starting epoch 4009\n", + "Train loss: 0.012432443425059319, Test loss: 0.019790420308709145\n", + "Starting epoch 4010\n", + "Train loss: 0.012506961803883315, Test loss: 0.02444387786090374\n", + "Starting epoch 4011\n", + "Train loss: 0.012497255075722934, Test loss: 0.023962296545505524\n", + "Starting epoch 4012\n", + "Train loss: 0.012472302988171577, Test loss: 0.02349131740629673\n", + "Starting epoch 4013\n", + "Train loss: 0.012426150273531675, Test loss: 0.020425844937562943\n", + "Starting epoch 4014\n", + "Train loss: 0.012391283642500639, Test loss: 0.02189396321773529\n", + "Starting epoch 4015\n", + "Train loss: 0.012429089453071356, Test loss: 0.02083679474890232\n", + "Starting epoch 4016\n", + "Train loss: 0.012469071261584759, Test loss: 0.0220344215631485\n", + "Starting epoch 4017\n", + "Train loss: 0.012431415859609843, Test loss: 0.017796112224459648\n", + "Starting epoch 4018\n", + "Train loss: 0.012455588020384312, Test loss: 0.020645365118980408\n", + "Starting epoch 4019\n", + "Train loss: 0.012481027152389287, Test loss: 0.02302505634725094\n", + "Starting epoch 4020\n", + "Train loss: 0.012454989627003669, Test loss: 0.02274871990084648\n", + "Starting epoch 4021\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.012421724256128072, Test loss: 0.024355825036764145\n", + "Starting epoch 4022\n", + "Train loss: 0.01250026624649763, Test loss: 0.02241867035627365\n", + "Starting epoch 4023\n", + "Train loss: 0.012427268344908952, Test loss: 0.020182207226753235\n", + "Starting epoch 4024\n", + "Train loss: 0.012431093361228704, Test loss: 0.019053686410188675\n", + "Starting epoch 4025\n", + "Train loss: 0.012473526578396558, Test loss: 0.020721163600683212\n", + "Starting epoch 4026\n", + "Train loss: 0.012501487210392952, Test loss: 0.022296417504549026\n", + "Starting epoch 4027\n", + "Train loss: 0.01255327820777893, Test loss: 0.022570496425032616\n", + "Starting epoch 4028\n", + "Train loss: 0.012424964401870965, Test loss: 0.012479749508202076\n", + "Starting epoch 4029\n", + "Train loss: 0.012483203914016486, Test loss: 0.011283298023045063\n", + "Starting epoch 4030\n", + "Train loss: 0.012465901225805282, Test loss: 0.011033876799046993\n", + "Starting epoch 4031\n", + "Train loss: 0.012431570328772068, Test loss: 0.01098774466663599\n", + "Starting epoch 4032\n", + "Train loss: 0.01244117721915245, Test loss: 0.019580963999032974\n", + "Starting epoch 4033\n", + "Train loss: 0.012450698614120483, Test loss: 0.02139567956328392\n", + "Starting epoch 4034\n", + "Train loss: 0.012450916897505522, Test loss: 0.021693388000130653\n", + "Starting epoch 4035\n", + "Train loss: 0.012474016956984997, Test loss: 0.017575077712535858\n", + "Starting epoch 4036\n", + "Train loss: 0.012429971303790808, Test loss: 0.022600412368774414\n", + "Starting epoch 4037\n", + "Train loss: 0.012454854138195515, Test loss: 0.01791662909090519\n", + "Starting epoch 4038\n", + "Train loss: 0.012372246775776148, Test loss: 0.01803106628358364\n", + "Starting epoch 4039\n", + "Train loss: 0.012410436049103737, Test loss: 0.02288452722132206\n", + "Starting epoch 4040\n", + "Train loss: 0.01245788598433137, Test loss: 0.022929668426513672\n", + "Starting epoch 4041\n", + "Train loss: 0.012318142652511598, Test loss: 0.016562826931476593\n", + "Starting epoch 4042\n", + "Train loss: 0.012450771015137434, Test loss: 0.021616147831082344\n", + "Starting epoch 4043\n", + "Train loss: 0.01256330756470561, Test loss: 0.021136829629540443\n", + "Starting epoch 4044\n", + "Train loss: 0.01248987140133977, Test loss: 0.019544627517461777\n", + "Starting epoch 4045\n", + "Train loss: 0.012458249516785145, Test loss: 0.021508527919650078\n", + "Starting epoch 4046\n", + "Train loss: 0.012422417234629392, Test loss: 0.020421374589204788\n", + "Starting epoch 4047\n", + "Train loss: 0.012418063506484032, Test loss: 0.020791690796613693\n", + "Starting epoch 4048\n", + "Train loss: 0.012402639538049699, Test loss: 0.02233058027923107\n", + "Starting epoch 4049\n", + "Train loss: 0.012372613921761512, Test loss: 0.01246999204158783\n", + "Starting epoch 4050\n", + "Train loss: 0.012392992731183767, Test loss: 0.025467026978731155\n", + "Starting epoch 4051\n", + "Train loss: 0.01243696378543973, Test loss: 0.021026745438575745\n", + "Starting epoch 4052\n", + "Train loss: 0.012426523938775063, Test loss: 0.023209912702441216\n", + "Starting epoch 4053\n", + "Train loss: 0.012395304683595896, Test loss: 0.020609110593795776\n", + "Starting epoch 4054\n", + "Train loss: 0.012359931580722332, Test loss: 0.022208889946341515\n", + "Starting epoch 4055\n", + "Train loss: 0.012355324365198612, Test loss: 0.022234540432691574\n", + "Starting epoch 4056\n", + "Train loss: 0.012382080517709256, Test loss: 0.023261722177267075\n", + "Starting epoch 4057\n", + "Train loss: 0.012386158909648656, Test loss: 0.024401480332016945\n", + "Starting epoch 4058\n", + "Train loss: 0.012276442907750606, Test loss: 0.017260253429412842\n", + "Starting epoch 4059\n", + "Train loss: 0.012368306368589401, Test loss: 0.02100956067442894\n", + "Starting epoch 4060\n", + "Train loss: 0.012389214169234037, Test loss: 0.023767486214637756\n", + "Starting epoch 4061\n", + "Train loss: 0.012295669671148061, Test loss: 0.02059188112616539\n", + "Starting epoch 4062\n", + "Train loss: 0.012343278676271439, Test loss: 0.013438425026834011\n", + "Starting epoch 4063\n", + "Train loss: 0.012419492565095424, Test loss: 0.012492852285504341\n", + "Starting epoch 4064\n", + "Train loss: 0.012394060753285885, Test loss: 0.012393043376505375\n", + "Starting epoch 4065\n", + "Train loss: 0.01239870447665453, Test loss: 0.014395751990377903\n", + "Starting epoch 4066\n", + "Train loss: 0.012316631209105254, Test loss: 0.011656266637146473\n", + "Starting epoch 4067\n", + "Train loss: 0.01230641946196556, Test loss: 0.014702991582453251\n", + "Starting epoch 4068\n", + "Train loss: 0.012416412103921175, Test loss: 0.0114049781113863\n", + "Starting epoch 4069\n", + "Train loss: 0.012391276061534881, Test loss: 0.014148017391562462\n", + "Starting epoch 4070\n", + "Train loss: 0.012336515802890063, Test loss: 0.010085390880703926\n", + "Starting epoch 4071\n", + "Train loss: 0.012326225079596043, Test loss: 0.018530724570155144\n", + "Starting epoch 4072\n", + "Train loss: 0.012374585065990686, Test loss: 0.018331529572606087\n", + "Starting epoch 4073\n", + "Train loss: 0.01234459986910224, Test loss: 0.020774956792593002\n", + "Starting epoch 4074\n", + "Train loss: 0.01232242899015546, Test loss: 0.021836208179593086\n", + "Starting epoch 4075\n", + "Train loss: 0.01231015494093299, Test loss: 0.019664978608489037\n", + "Starting epoch 4076\n", + "Train loss: 0.012381917219609023, Test loss: 0.022605739533901215\n", + "Starting epoch 4077\n", + "Train loss: 0.012306338921189309, Test loss: 0.0213219802826643\n", + "Starting epoch 4078\n", + "Train loss: 0.012315900456160307, Test loss: 0.02129622921347618\n", + "Starting epoch 4079\n", + "Train loss: 0.012294458895921708, Test loss: 0.019839078187942505\n", + "Starting epoch 4080\n", + "Train loss: 0.012303944136947393, Test loss: 0.010942569933831692\n", + "Starting epoch 4081\n", + "Train loss: 0.012284519877284765, Test loss: 0.011206986382603645\n", + "Starting epoch 4082\n", + "Train loss: 0.012337619625031947, Test loss: 0.011816030368208885\n", + "Starting epoch 4083\n", + "Train loss: 0.01232810752466321, Test loss: 0.013326352462172508\n", + "Starting epoch 4084\n", + "Train loss: 0.012293158285319804, Test loss: 0.013796250335872173\n", + "Starting epoch 4085\n", + "Train loss: 0.01232674954459071, Test loss: 0.011235426180064678\n", + "Starting epoch 4086\n", + "Train loss: 0.012278542015701532, Test loss: 0.010812824591994286\n", + "Starting epoch 4087\n", + "Train loss: 0.012402644716203212, Test loss: 0.011401773430407047\n", + "Starting epoch 4088\n", + "Train loss: 0.012227973826229572, Test loss: 0.01148188579827547\n", + "Starting epoch 4089\n", + "Train loss: 0.012368816491216421, Test loss: 0.013060886412858963\n", + "Starting epoch 4090\n", + "Train loss: 0.012284269370138646, Test loss: 0.011593044735491276\n", + "Starting epoch 4091\n", + "Train loss: 0.012346257027238607, Test loss: 0.011685144156217575\n", + "Starting epoch 4092\n", + "Train loss: 0.012410497833043337, Test loss: 0.010960358195006847\n", + "Starting epoch 4093\n", + "Train loss: 0.012326347008347512, Test loss: 0.011805038899183273\n", + "Starting epoch 4094\n", + "Train loss: 0.012323156259953975, Test loss: 0.00956928264349699\n", + "Starting epoch 4095\n", + "Train loss: 0.012278890293091536, Test loss: 0.014821802265942097\n", + "Starting epoch 4096\n", + "Train loss: 0.012382928878068925, Test loss: 0.010564824566245079\n", + "Starting epoch 4097\n", + "Train loss: 0.012339805886149406, Test loss: 0.011339659802615643\n", + "Starting epoch 4098\n", + "Train loss: 0.012318069934844971, Test loss: 0.010721155442297459\n", + "Starting epoch 4099\n", + "Train loss: 0.014415041860193014, Test loss: 0.023331215605139732\n", + "Starting epoch 4100\n", + "Train loss: 0.013827028442174196, Test loss: 0.01781369000673294\n", + "Starting epoch 4101\n", + "Train loss: 0.012497053276747465, Test loss: 0.023250581696629524\n", + "Starting epoch 4102\n", + "Train loss: 0.012429081965237856, Test loss: 0.025517648085951805\n", + "Starting epoch 4103\n", + "Train loss: 0.012339083310216666, Test loss: 0.018474025651812553\n", + "Starting epoch 4104\n", + "Train loss: 0.012449358068406581, Test loss: 0.022641897201538086\n", + "Starting epoch 4105\n", + "Train loss: 0.01241886030882597, Test loss: 0.02103336714208126\n", + "Starting epoch 4106\n", + "Train loss: 0.012366295456886292, Test loss: 0.022055426612496376\n", + "Starting epoch 4107\n", + "Train loss: 0.01228659376502037, Test loss: 0.021938344463706017\n", + "Starting epoch 4108\n", + "Train loss: 0.012360660694539546, Test loss: 0.02124459482729435\n", + "Starting epoch 4109\n", + "Train loss: 0.01235422806814313, Test loss: 0.022487768903374672\n", + "Starting epoch 4110\n", + "Train loss: 0.012212421614676713, Test loss: 0.015760499984025955\n", + "Starting epoch 4111\n", + "Train loss: 0.012205541506409645, Test loss: 0.019086573272943497\n", + "Starting epoch 4112\n", + "Train loss: 0.01235119417309761, Test loss: 0.022148730233311653\n", + "Starting epoch 4113\n", + "Train loss: 0.012376351170241833, Test loss: 0.02050481177866459\n", + "Starting epoch 4114\n", + "Train loss: 0.012287011463195085, Test loss: 0.024128733202815056\n", + "Starting epoch 4115\n", + "Train loss: 0.012225917521864176, Test loss: 0.011380781419575214\n", + "Starting epoch 4116\n", + "Train loss: 0.012325191795825958, Test loss: 0.01161632314324379\n", + "Starting epoch 4117\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.012346604783087968, Test loss: 0.011974656954407692\n", + "Starting epoch 4118\n", + "Train loss: 0.012347576767206192, Test loss: 0.012557470239698887\n", + "Starting epoch 4119\n", + "Train loss: 0.012236240189522504, Test loss: 0.010797512717545033\n", + "Starting epoch 4120\n", + "Train loss: 0.01230518391355872, Test loss: 0.011490694247186184\n", + "Starting epoch 4121\n", + "Train loss: 0.01226296804845333, Test loss: 0.021656999364495277\n", + "Starting epoch 4122\n", + "Train loss: 0.012298166546970605, Test loss: 0.021392736583948135\n", + "Starting epoch 4123\n", + "Train loss: 0.012274085264652967, Test loss: 0.023779820650815964\n", + "Starting epoch 4124\n", + "Train loss: 0.012217366434633732, Test loss: 0.010965133085846901\n", + "Starting epoch 4125\n", + "Train loss: 0.012233949843794107, Test loss: 0.011909302324056625\n", + "Starting epoch 4126\n", + "Train loss: 0.012252450957894324, Test loss: 0.011831799522042274\n", + "Starting epoch 4127\n", + "Train loss: 0.012327962219715119, Test loss: 0.013409928418695927\n", + "Starting epoch 4128\n", + "Train loss: 0.012190339770168066, Test loss: 0.013631378300487995\n", + "Starting epoch 4129\n", + "Train loss: 0.012195932976901531, Test loss: 0.011012750677764416\n", + "Starting epoch 4130\n", + "Train loss: 0.012234253641217947, Test loss: 0.010935928672552109\n", + "Starting epoch 4131\n", + "Train loss: 0.01227225111797452, Test loss: 0.012883746065199375\n", + "Starting epoch 4132\n", + "Train loss: 0.012243798524141312, Test loss: 0.010856124572455883\n", + "Starting epoch 4133\n", + "Train loss: 0.012268432471901179, Test loss: 0.011245124973356724\n", + "Starting epoch 4134\n", + "Train loss: 0.012261144537478685, Test loss: 0.012026195414364338\n", + "Starting epoch 4135\n", + "Train loss: 0.012154908087104559, Test loss: 0.012108869850635529\n", + "Starting epoch 4136\n", + "Train loss: 0.012270989567041398, Test loss: 0.013944494538009167\n", + "Starting epoch 4137\n", + "Train loss: 0.012152463607490062, Test loss: 0.011039367876946926\n", + "Starting epoch 4138\n", + "Train loss: 0.012269120924174785, Test loss: 0.012404412031173706\n", + "Starting epoch 4139\n", + "Train loss: 0.012218359392136335, Test loss: 0.011480685323476791\n", + "Starting epoch 4140\n", + "Train loss: 0.012231394033879042, Test loss: 0.019695119932293892\n", + "Starting epoch 4141\n", + "Train loss: 0.012140713855624199, Test loss: 0.017746577039361\n", + "Starting epoch 4142\n", + "Train loss: 0.012249498385936022, Test loss: 0.022262437269091606\n", + "Starting epoch 4143\n", + "Train loss: 0.012225225828588008, Test loss: 0.02113231271505356\n", + "Starting epoch 4144\n", + "Train loss: 0.012231253404170275, Test loss: 0.02328345738351345\n", + "Starting epoch 4145\n", + "Train loss: 0.012226494923233987, Test loss: 0.015026113949716091\n", + "Starting epoch 4146\n", + "Train loss: 0.012161169946193696, Test loss: 0.02346387691795826\n", + "Starting epoch 4147\n", + "Train loss: 0.01228803351521492, Test loss: 0.020835891366004944\n", + "Starting epoch 4148\n", + "Train loss: 0.012187549844384193, Test loss: 0.02285509556531906\n", + "Starting epoch 4149\n", + "Train loss: 0.012245803773403167, Test loss: 0.020752359181642532\n", + "Starting epoch 4150\n", + "Train loss: 0.01226930070668459, Test loss: 0.019839230924844742\n", + "Starting epoch 4151\n", + "Train loss: 0.012194615975022315, Test loss: 0.018792351707816124\n", + "Starting epoch 4152\n", + "Train loss: 0.012150033097714185, Test loss: 0.02124713733792305\n", + "Starting epoch 4153\n", + "Train loss: 0.012221220787614584, Test loss: 0.020884571596980095\n", + "Starting epoch 4154\n", + "Train loss: 0.012189752459526061, Test loss: 0.02250959537923336\n", + "Starting epoch 4155\n", + "Train loss: 0.012152222450822592, Test loss: 0.011770540848374367\n", + "Starting epoch 4156\n", + "Train loss: 0.012193573359400033, Test loss: 0.011179818771779537\n", + "Starting epoch 4157\n", + "Train loss: 0.012181501761078834, Test loss: 0.01130808237940073\n", + "Starting epoch 4158\n", + "Train loss: 0.012175047267228366, Test loss: 0.012035851366817951\n", + "Starting epoch 4159\n", + "Train loss: 0.012340571824461221, Test loss: 0.012169325724244118\n", + "Starting epoch 4160\n", + "Train loss: 0.012244439739733935, Test loss: 0.011698590591549873\n", + "Starting epoch 4161\n", + "Train loss: 0.012227270267903805, Test loss: 0.010522688739001751\n", + "Starting epoch 4162\n", + "Train loss: 0.012219071798026561, Test loss: 0.01187115628272295\n", + "Starting epoch 4163\n", + "Train loss: 0.012119725085794926, Test loss: 0.011348732747137547\n", + "Starting epoch 4164\n", + "Train loss: 0.012268929220736026, Test loss: 0.012996520847082138\n", + "Starting epoch 4165\n", + "Train loss: 0.012157584130764008, Test loss: 0.012579347938299179\n", + "Starting epoch 4166\n", + "Train loss: 0.012146954350173473, Test loss: 0.011501472443342209\n", + "Starting epoch 4167\n", + "Train loss: 0.012221597358584404, Test loss: 0.013811946846544743\n", + "Starting epoch 4168\n", + "Train loss: 0.012127835098654031, Test loss: 0.02403273992240429\n", + "Starting epoch 4169\n", + "Train loss: 0.012233353946357965, Test loss: 0.01653159223496914\n", + "Starting epoch 4170\n", + "Train loss: 0.012106818873435258, Test loss: 0.022996369749307632\n", + "Starting epoch 4171\n", + "Train loss: 0.01223453089594841, Test loss: 0.020605584606528282\n", + "Starting epoch 4172\n", + "Train loss: 0.012145185042172671, Test loss: 0.011242113076150417\n", + "Starting epoch 4173\n", + "Train loss: 0.012332647293806075, Test loss: 0.02139398455619812\n", + "Starting epoch 4174\n", + "Train loss: 0.012300862539559603, Test loss: 0.021342214196920395\n", + "Starting epoch 4175\n", + "Train loss: 0.012264504209160804, Test loss: 0.021365631371736526\n", + "Starting epoch 4176\n", + "Train loss: 0.012201856411993504, Test loss: 0.011348056606948376\n", + "Starting epoch 4177\n", + "Train loss: 0.012215812262147664, Test loss: 0.012474162504076958\n", + "Starting epoch 4178\n", + "Train loss: 0.012202929630875587, Test loss: 0.012800930067896843\n", + "Starting epoch 4179\n", + "Train loss: 0.012171632163226604, Test loss: 0.011936523951590061\n", + "Starting epoch 4180\n", + "Train loss: 0.01226169541478157, Test loss: 0.010826869867742062\n", + "Starting epoch 4181\n", + "Train loss: 0.012202050313353538, Test loss: 0.012876884080469608\n", + "Starting epoch 4182\n", + "Train loss: 0.012121893782168627, Test loss: 0.010035284794867039\n", + "Starting epoch 4183\n", + "Train loss: 0.012106944117695093, Test loss: 0.011368317529559135\n", + "Starting epoch 4184\n", + "Train loss: 0.012310805711895228, Test loss: 0.011230727657675743\n", + "Starting epoch 4185\n", + "Train loss: 0.01216790558770299, Test loss: 0.011680932715535164\n", + "Starting epoch 4186\n", + "Train loss: 0.012164471056312322, Test loss: 0.010772721841931343\n", + "Starting epoch 4187\n", + "Train loss: 0.01221997307613492, Test loss: 0.012851184234023094\n", + "Starting epoch 4188\n", + "Train loss: 0.012161746509373189, Test loss: 0.010808208025991917\n", + "Starting epoch 4189\n", + "Train loss: 0.01215314382687211, Test loss: 0.01260766014456749\n", + "Starting epoch 4190\n", + "Train loss: 0.012116262465715408, Test loss: 0.011213183403015137\n", + "Starting epoch 4191\n", + "Train loss: 0.012150989845395088, Test loss: 0.01233639195561409\n", + "Starting epoch 4192\n", + "Train loss: 0.012146351523697377, Test loss: 0.01143808476626873\n", + "Starting epoch 4193\n", + "Train loss: 0.012124222237616777, Test loss: 0.012266326695680618\n", + "Starting epoch 4194\n", + "Train loss: 0.012147581540048122, Test loss: 0.012545785866677761\n", + "Starting epoch 4195\n", + "Train loss: 0.01209733374416828, Test loss: 0.012028804048895836\n", + "Starting epoch 4196\n", + "Train loss: 0.01213813802227378, Test loss: 0.011680759489536285\n", + "Starting epoch 4197\n", + "Train loss: 0.012070858031511306, Test loss: 0.012385970912873745\n", + "Starting epoch 4198\n", + "Train loss: 0.012119567375630141, Test loss: 0.020847683772444725\n", + "Starting epoch 4199\n", + "Train loss: 0.012145249284803867, Test loss: 0.02049531228840351\n", + "Starting epoch 4200\n", + "Train loss: 0.012075096629559993, Test loss: 0.024910587817430496\n", + "Starting epoch 4201\n", + "Train loss: 0.01214232698082924, Test loss: 0.020727990195155144\n", + "Starting epoch 4202\n", + "Train loss: 0.01217579672113061, Test loss: 0.021286658942699432\n", + "Starting epoch 4203\n", + "Train loss: 0.01204925900325179, Test loss: 0.012339964509010315\n", + "Starting epoch 4204\n", + "Train loss: 0.0122148441337049, Test loss: 0.022634439170360565\n", + "Starting epoch 4205\n", + "Train loss: 0.012166645471006632, Test loss: 0.0215445626527071\n", + "Starting epoch 4206\n", + "Train loss: 0.012057035230100155, Test loss: 0.01078453753143549\n", + "Starting epoch 4207\n", + "Train loss: 0.012132402081042527, Test loss: 0.020668333396315575\n", + "Starting epoch 4208\n", + "Train loss: 0.012052871268242598, Test loss: 0.012188442051410675\n", + "Starting epoch 4209\n", + "Train loss: 0.012249129377305507, Test loss: 0.021062562242150307\n", + "Starting epoch 4210\n", + "Train loss: 0.012238431852310896, Test loss: 0.019350068643689156\n", + "Starting epoch 4211\n", + "Train loss: 0.012190677970647813, Test loss: 0.019870249554514885\n", + "Starting epoch 4212\n", + "Train loss: 0.012140074651688338, Test loss: 0.020219027996063232\n", + "Starting epoch 4213\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.01207819165661931, Test loss: 0.022244097664952278\n", + "Starting epoch 4214\n", + "Train loss: 0.012090999055653811, Test loss: 0.01113820169121027\n", + "Starting epoch 4215\n", + "Train loss: 0.012118311021476985, Test loss: 0.011509462259709835\n", + "Starting epoch 4216\n", + "Train loss: 0.012179187703877688, Test loss: 0.010739369317889214\n", + "Starting epoch 4217\n", + "Train loss: 0.012145201042294503, Test loss: 0.010016446001827717\n", + "Starting epoch 4218\n", + "Train loss: 0.011987934559583665, Test loss: 0.013887215405702591\n", + "Starting epoch 4219\n", + "Train loss: 0.012101615145802498, Test loss: 0.024394916370511055\n", + "Starting epoch 4220\n", + "Train loss: 0.012145720720291138, Test loss: 0.023863239213824272\n", + "Starting epoch 4221\n", + "Train loss: 0.012106169164180756, Test loss: 0.020121973007917404\n", + "Starting epoch 4222\n", + "Train loss: 0.012031511981040239, Test loss: 0.02174992300570011\n", + "Starting epoch 4223\n", + "Train loss: 0.012095932494848966, Test loss: 0.019265463575720787\n", + "Starting epoch 4224\n", + "Train loss: 0.012021228689700365, Test loss: 0.026746395975351334\n", + "Starting epoch 4225\n", + "Train loss: 0.012076886594295502, Test loss: 0.017346566542983055\n", + "Starting epoch 4226\n", + "Train loss: 0.01210465122014284, Test loss: 0.010100891813635826\n", + "Starting epoch 4227\n", + "Train loss: 0.012124884016811848, Test loss: 0.015218665823340416\n", + "Starting epoch 4228\n", + "Train loss: 0.012085852343589068, Test loss: 0.011084534227848053\n", + "Starting epoch 4229\n", + "Train loss: 0.012166772168129683, Test loss: 0.010521983727812767\n", + "Starting epoch 4230\n", + "Train loss: 0.01213056594133377, Test loss: 0.011856526136398315\n", + "Starting epoch 4231\n", + "Train loss: 0.012101553045213223, Test loss: 0.011518171057105064\n", + "Starting epoch 4232\n", + "Train loss: 0.012049212902784347, Test loss: 0.010476069524884224\n", + "Starting epoch 4233\n", + "Train loss: 0.012020433656871318, Test loss: 0.015016875229775906\n", + "Starting epoch 4234\n", + "Train loss: 0.01208143288269639, Test loss: 0.014646482653915882\n", + "Starting epoch 4235\n", + "Train loss: 0.012089429702609778, Test loss: 0.011678236536681652\n", + "Starting epoch 4236\n", + "Train loss: 0.01218665722757578, Test loss: 0.010639411397278309\n", + "Starting epoch 4237\n", + "Train loss: 0.012094216346740723, Test loss: 0.009955613873898983\n", + "Starting epoch 4238\n", + "Train loss: 0.012110248748213053, Test loss: 0.01085885800421238\n", + "Starting epoch 4239\n", + "Train loss: 0.01209892962127924, Test loss: 0.012646077200770378\n", + "Starting epoch 4240\n", + "Train loss: 0.012082293946295977, Test loss: 0.01341546606272459\n", + "Starting epoch 4241\n", + "Train loss: 0.01203711086884141, Test loss: 0.011058167554438114\n", + "Starting epoch 4242\n", + "Train loss: 0.012024741023778915, Test loss: 0.011026792228221893\n", + "Starting epoch 4243\n", + "Train loss: 0.012048248499631882, Test loss: 0.010248756967484951\n", + "Starting epoch 4244\n", + "Train loss: 0.012024400979280472, Test loss: 0.011513634584844112\n", + "Starting epoch 4245\n", + "Train loss: 0.012033263146877289, Test loss: 0.010754303075373173\n", + "Starting epoch 4246\n", + "Train loss: 0.012016190402209758, Test loss: 0.012418544851243496\n", + "Starting epoch 4247\n", + "Train loss: 0.012002525497227906, Test loss: 0.011084917932748795\n", + "Starting epoch 4248\n", + "Train loss: 0.011976948082447052, Test loss: 0.012651936151087284\n", + "Starting epoch 4249\n", + "Train loss: 0.01200058612972498, Test loss: 0.011131472885608673\n", + "Starting epoch 4250\n", + "Train loss: 0.01201055120676756, Test loss: 0.011083678342401981\n", + "Starting epoch 4251\n", + "Train loss: 0.012033022586256265, Test loss: 0.012087777256965637\n", + "Starting epoch 4252\n", + "Train loss: 0.011955739539116621, Test loss: 0.012747758999466896\n", + "Starting epoch 4253\n", + "Train loss: 0.012073248233646154, Test loss: 0.010546451434493065\n", + "Starting epoch 4254\n", + "Train loss: 0.012104565352201462, Test loss: 0.011180202476680279\n", + "Starting epoch 4255\n", + "Train loss: 0.012031689696013927, Test loss: 0.011165031231939793\n", + "Starting epoch 4256\n", + "Train loss: 0.011989536322653293, Test loss: 0.011225122958421707\n", + "Starting epoch 4257\n", + "Train loss: 0.011930504906922578, Test loss: 0.013073590584099293\n", + "Starting epoch 4258\n", + "Train loss: 0.012023835219442844, Test loss: 0.011289265006780624\n", + "Starting epoch 4259\n", + "Train loss: 0.012050361167639493, Test loss: 0.010772423818707466\n", + "Starting epoch 4260\n", + "Train loss: 0.01207342827692628, Test loss: 0.010602396912872791\n", + "Starting epoch 4261\n", + "Train loss: 0.01205658219754696, Test loss: 0.01210835576057434\n", + "Starting epoch 4262\n", + "Train loss: 0.01205407341942191, Test loss: 0.011132092215120792\n", + "Starting epoch 4263\n", + "Train loss: 0.011952140536159277, Test loss: 0.009536990895867348\n", + "Starting epoch 4264\n", + "Train loss: 0.011850009690970182, Test loss: 0.011035604402422905\n", + "Starting epoch 4265\n", + "Train loss: 0.011998005341738463, Test loss: 0.024264825507998466\n", + "Starting epoch 4266\n", + "Train loss: 0.012073729671537876, Test loss: 0.02001360058784485\n", + "Starting epoch 4267\n", + "Train loss: 0.012121955268085002, Test loss: 0.01921500451862812\n", + "Starting epoch 4268\n", + "Train loss: 0.012067520227283239, Test loss: 0.0205758698284626\n", + "Starting epoch 4269\n", + "Train loss: 0.012047969680279493, Test loss: 0.02197849191725254\n", + "Starting epoch 4270\n", + "Train loss: 0.01200153885409236, Test loss: 0.01649688370525837\n", + "Starting epoch 4271\n", + "Train loss: 0.011955674961209297, Test loss: 0.023125825449824333\n", + "Starting epoch 4272\n", + "Train loss: 0.011967693138867617, Test loss: 0.02131819538772106\n", + "Starting epoch 4273\n", + "Train loss: 0.011988760717213155, Test loss: 0.020357981324195862\n", + "Starting epoch 4274\n", + "Train loss: 0.01204956091940403, Test loss: 0.022652165964245796\n", + "Starting epoch 4275\n", + "Train loss: 0.011978960800915957, Test loss: 0.023736992850899696\n", + "Starting epoch 4276\n", + "Train loss: 0.01199278088286519, Test loss: 0.016854792833328247\n", + "Starting epoch 4277\n", + "Train loss: 0.011975176203995942, Test loss: 0.0189408827573061\n", + "Starting epoch 4278\n", + "Train loss: 0.011979218311607838, Test loss: 0.016514385119080544\n", + "Starting epoch 4279\n", + "Train loss: 0.011965181604027748, Test loss: 0.01959126442670822\n", + "Starting epoch 4280\n", + "Train loss: 0.011950979977846145, Test loss: 0.02097390778362751\n", + "Starting epoch 4281\n", + "Train loss: 0.012005682326853276, Test loss: 0.023603809997439384\n", + "Starting epoch 4282\n", + "Train loss: 0.012001551035791636, Test loss: 0.02009539306163788\n", + "Starting epoch 4283\n", + "Train loss: 0.011944413743913173, Test loss: 0.022020602598786354\n", + "Starting epoch 4284\n", + "Train loss: 0.011911687795072795, Test loss: 0.025020157918334007\n", + "Starting epoch 4285\n", + "Train loss: 0.011963980980217457, Test loss: 0.018677357584238052\n", + "Starting epoch 4286\n", + "Train loss: 0.01198837000876665, Test loss: 0.019430672749876976\n", + "Starting epoch 4287\n", + "Train loss: 0.0119148557074368, Test loss: 0.012210730463266373\n", + "Starting epoch 4288\n", + "Train loss: 0.01196265995502472, Test loss: 0.012889232486486435\n", + "Starting epoch 4289\n", + "Train loss: 0.01206696456298232, Test loss: 0.011283011175692081\n", + "Starting epoch 4290\n", + "Train loss: 0.012013184670358897, Test loss: 0.010955486446619034\n", + "Starting epoch 4291\n", + "Train loss: 0.011953068785369397, Test loss: 0.02072145603597164\n", + "Starting epoch 4292\n", + "Train loss: 0.011965151764452458, Test loss: 0.020950568839907646\n", + "Starting epoch 4293\n", + "Train loss: 0.011957227624952793, Test loss: 0.020106831565499306\n", + "Starting epoch 4294\n", + "Train loss: 0.011986920554190874, Test loss: 0.02257026731967926\n", + "Starting epoch 4295\n", + "Train loss: 0.011951432563364506, Test loss: 0.020862175151705742\n", + "Starting epoch 4296\n", + "Train loss: 0.011917166616767646, Test loss: 0.011959921568632126\n", + "Starting epoch 4297\n", + "Train loss: 0.011938585676252842, Test loss: 0.010572580620646477\n", + "Starting epoch 4298\n", + "Train loss: 0.012028227355331182, Test loss: 0.011456713080406189\n", + "Starting epoch 4299\n", + "Train loss: 0.011909768003970384, Test loss: 0.00938858650624752\n", + "Starting epoch 4300\n", + "Train loss: 0.011938616391271353, Test loss: 0.011112083680927753\n", + "Starting epoch 4301\n", + "Train loss: 0.01193693796172738, Test loss: 0.011422183364629745\n", + "Starting epoch 4302\n", + "Train loss: 0.013413220457732678, Test loss: 0.012039653956890106\n", + "Starting epoch 4303\n", + "Train loss: 0.013830297831445933, Test loss: 0.012831076048314571\n", + "Starting epoch 4304\n", + "Train loss: 0.01227016992866993, Test loss: 0.021650051698088646\n", + "Starting epoch 4305\n", + "Train loss: 0.012271190397441387, Test loss: 0.022421013563871384\n", + "Starting epoch 4306\n", + "Train loss: 0.012176300622522831, Test loss: 0.021209923550486565\n", + "Starting epoch 4307\n", + "Train loss: 0.012103465106338262, Test loss: 0.023576103150844574\n", + "Starting epoch 4308\n", + "Train loss: 0.012021175790578128, Test loss: 0.021699529141187668\n", + "Starting epoch 4309\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.012018333319574595, Test loss: 0.022986024618148804\n", + "Starting epoch 4310\n", + "Train loss: 0.011955867614597082, Test loss: 0.023704826831817627\n", + "Starting epoch 4311\n", + "Train loss: 0.01194892730563879, Test loss: 0.020287219434976578\n", + "Starting epoch 4312\n", + "Train loss: 0.011929182559251785, Test loss: 0.019903739914298058\n", + "Starting epoch 4313\n", + "Train loss: 0.011935645993798972, Test loss: 0.01760723628103733\n", + "Starting epoch 4314\n", + "Train loss: 0.012043988462537528, Test loss: 0.022156259045004845\n", + "Starting epoch 4315\n", + "Train loss: 0.012054080124944449, Test loss: 0.018619947135448456\n", + "Starting epoch 4316\n", + "Train loss: 0.011973652187734843, Test loss: 0.020179735496640205\n", + "Starting epoch 4317\n", + "Train loss: 0.011896523423492908, Test loss: 0.018909327685832977\n", + "Starting epoch 4318\n", + "Train loss: 0.01191521018743515, Test loss: 0.011211621575057507\n", + "Starting epoch 4319\n", + "Train loss: 0.011937210839241742, Test loss: 0.01117775123566389\n", + "Starting epoch 4320\n", + "Train loss: 0.011957618817687034, Test loss: 0.01272883452475071\n", + "Starting epoch 4321\n", + "Train loss: 0.011878011580556631, Test loss: 0.02014046721160412\n", + "Starting epoch 4322\n", + "Train loss: 0.011980686001479626, Test loss: 0.02053026854991913\n", + "Starting epoch 4323\n", + "Train loss: 0.011945093329995871, Test loss: 0.023283910006284714\n", + "Starting epoch 4324\n", + "Train loss: 0.011912998780608176, Test loss: 0.023477083072066307\n", + "Starting epoch 4325\n", + "Train loss: 0.01190895402804017, Test loss: 0.020149026066064835\n", + "Starting epoch 4326\n", + "Train loss: 0.01522939845919609, Test loss: 0.014960262924432755\n", + "Starting epoch 4327\n", + "Train loss: 0.01591642018407583, Test loss: 0.020286312326788902\n", + "Starting epoch 4328\n", + "Train loss: 0.013130051121115684, Test loss: 0.016283048316836357\n", + "Starting epoch 4329\n", + "Train loss: 0.01243195280432701, Test loss: 0.011149575933814049\n", + "Starting epoch 4330\n", + "Train loss: 0.012293483652174473, Test loss: 0.011439966969192028\n", + "Starting epoch 4331\n", + "Train loss: 0.0121330994553864, Test loss: 0.010997052304446697\n", + "Starting epoch 4332\n", + "Train loss: 0.012257701810449361, Test loss: 0.009985586628317833\n", + "Starting epoch 4333\n", + "Train loss: 0.012157102935016156, Test loss: 0.010828331112861633\n", + "Starting epoch 4334\n", + "Train loss: 0.012030509524047375, Test loss: 0.011728598736226559\n", + "Starting epoch 4335\n", + "Train loss: 0.012093030139803887, Test loss: 0.010317803360521793\n", + "Starting epoch 4336\n", + "Train loss: 0.01209097532555461, Test loss: 0.010052403435111046\n", + "Starting epoch 4337\n", + "Train loss: 0.011976988427340985, Test loss: 0.01135446596890688\n", + "Starting epoch 4338\n", + "Train loss: 0.011930536590516566, Test loss: 0.010623445734381676\n", + "Starting epoch 4339\n", + "Train loss: 0.011949306670576334, Test loss: 0.011701167561113834\n", + "Starting epoch 4340\n", + "Train loss: 0.01192726330831647, Test loss: 0.011169913224875927\n", + "Starting epoch 4341\n", + "Train loss: 0.011924691945314408, Test loss: 0.009968002326786518\n", + "Starting epoch 4342\n", + "Train loss: 0.011790753025561571, Test loss: 0.017145859077572823\n", + "Starting epoch 4343\n", + "Train loss: 0.011886756904423236, Test loss: 0.01947522535920143\n", + "Starting epoch 4344\n", + "Train loss: 0.011849880144000053, Test loss: 0.01832471415400505\n", + "Starting epoch 4345\n", + "Train loss: 0.01191223444417119, Test loss: 0.02403796836733818\n", + "Starting epoch 4346\n", + "Train loss: 0.011735152583569289, Test loss: 0.02172820456326008\n", + "Starting epoch 4347\n", + "Train loss: 0.01198168095201254, Test loss: 0.021824317052960396\n", + "Starting epoch 4348\n", + "Train loss: 0.011934872809797525, Test loss: 0.02306642383337021\n", + "Starting epoch 4349\n", + "Train loss: 0.011744203083217143, Test loss: 0.020627476274967194\n", + "Starting epoch 4350\n", + "Train loss: 0.011953374613076448, Test loss: 0.021941883489489555\n", + "Starting epoch 4351\n", + "Train loss: 0.011853212043642997, Test loss: 0.025208432227373123\n", + "Starting epoch 4352\n", + "Train loss: 0.011874822359532118, Test loss: 0.01938655599951744\n", + "Starting epoch 4353\n", + "Train loss: 0.011765497401356698, Test loss: 0.019472667947411537\n", + "Starting epoch 4354\n", + "Train loss: 0.011981873214244843, Test loss: 0.022462919354438782\n", + "Starting epoch 4355\n", + "Train loss: 0.011867767851799726, Test loss: 0.02010282129049301\n", + "Starting epoch 4356\n", + "Train loss: 0.011855372320860624, Test loss: 0.020964063704013824\n", + "Starting epoch 4357\n", + "Train loss: 0.011816510707139969, Test loss: 0.017913449555635452\n", + "Starting epoch 4358\n", + "Train loss: 0.011801526620984078, Test loss: 0.017815260216593742\n", + "Starting epoch 4359\n", + "Train loss: 0.011761323194950818, Test loss: 0.019729653373360634\n", + "Starting epoch 4360\n", + "Train loss: 0.011803838070482015, Test loss: 0.020247580483555794\n", + "Starting epoch 4361\n", + "Train loss: 0.011826137732714414, Test loss: 0.02087121270596981\n", + "Starting epoch 4362\n", + "Train loss: 0.01178081214427948, Test loss: 0.01820509321987629\n", + "Starting epoch 4363\n", + "Train loss: 0.011736841313540936, Test loss: 0.02228914201259613\n", + "Starting epoch 4364\n", + "Train loss: 0.011789177227765322, Test loss: 0.01950191520154476\n", + "Starting epoch 4365\n", + "Train loss: 0.01180281575769186, Test loss: 0.01637946255505085\n", + "Starting epoch 4366\n", + "Train loss: 0.011763213723897934, Test loss: 0.018269402906298637\n", + "Starting epoch 4367\n", + "Train loss: 0.011810167804360389, Test loss: 0.022438380867242813\n", + "Starting epoch 4368\n", + "Train loss: 0.011782312132418156, Test loss: 0.020827049389481544\n", + "Starting epoch 4369\n", + "Train loss: 0.01179568698629737, Test loss: 0.021513840183615685\n", + "Starting epoch 4370\n", + "Train loss: 0.011794966738671064, Test loss: 0.017756309360265732\n", + "Starting epoch 4371\n", + "Train loss: 0.011727482490241527, Test loss: 0.02000461146235466\n", + "Starting epoch 4372\n", + "Train loss: 0.011772380284965038, Test loss: 0.02047029323875904\n", + "Starting epoch 4373\n", + "Train loss: 0.011762444898486137, Test loss: 0.020101265981793404\n", + "Starting epoch 4374\n", + "Train loss: 0.01180083842948079, Test loss: 0.017280245199799538\n", + "Starting epoch 4375\n", + "Train loss: 0.011764730457216502, Test loss: 0.02188759855926037\n", + "Starting epoch 4376\n", + "Train loss: 0.011795772407203912, Test loss: 0.018293779343366623\n", + "Starting epoch 4377\n", + "Train loss: 0.011787792444229126, Test loss: 0.017527418211102486\n", + "Starting epoch 4378\n", + "Train loss: 0.01178733704611659, Test loss: 0.02064279280602932\n", + "Starting epoch 4379\n", + "Train loss: 0.011761167850345373, Test loss: 0.023492639884352684\n", + "Starting epoch 4380\n", + "Train loss: 0.011667147297412157, Test loss: 0.019731979817152023\n", + "Starting epoch 4381\n", + "Train loss: 0.011712643895298242, Test loss: 0.019413143396377563\n", + "Starting epoch 4382\n", + "Train loss: 0.011716578416526317, Test loss: 0.020870978012681007\n", + "Starting epoch 4383\n", + "Train loss: 0.011786021739244461, Test loss: 0.02036423236131668\n", + "Starting epoch 4384\n", + "Train loss: 0.01178757295012474, Test loss: 0.01925116404891014\n", + "Starting epoch 4385\n", + "Train loss: 0.01177046837285161, Test loss: 0.020886998623609543\n", + "Starting epoch 4386\n", + "Train loss: 0.011722645703703165, Test loss: 0.019354848191142082\n", + "Starting epoch 4387\n", + "Train loss: 0.011746967509388924, Test loss: 0.02019415982067585\n", + "Starting epoch 4388\n", + "Train loss: 0.01183115953579545, Test loss: 0.019071271643042564\n", + "Starting epoch 4389\n", + "Train loss: 0.011838594917207956, Test loss: 0.02385210432112217\n", + "Starting epoch 4390\n", + "Train loss: 0.0117006297968328, Test loss: 0.02211766690015793\n", + "Starting epoch 4391\n", + "Train loss: 0.011865116283297538, Test loss: 0.017051152884960175\n", + "Starting epoch 4392\n", + "Train loss: 0.011817776188254357, Test loss: 0.021824464201927185\n", + "Starting epoch 4393\n", + "Train loss: 0.011762524731457233, Test loss: 0.019949346780776978\n", + "Starting epoch 4394\n", + "Train loss: 0.011738350614905357, Test loss: 0.01455700397491455\n", + "Starting epoch 4395\n", + "Train loss: 0.0117892200127244, Test loss: 0.01098135206848383\n", + "Starting epoch 4396\n", + "Train loss: 0.011778285615146161, Test loss: 0.021842094138264656\n", + "Starting epoch 4397\n", + "Train loss: 0.011878576558083297, Test loss: 0.02020924724638462\n", + "Starting epoch 4398\n", + "Train loss: 0.01181599885225296, Test loss: 0.013534068129956722\n", + "Starting epoch 4399\n", + "Train loss: 0.011799909286201, Test loss: 0.02040730230510235\n", + "Starting epoch 4400\n", + "Train loss: 0.011791416853666305, Test loss: 0.01785309985280037\n", + "Starting epoch 4401\n", + "Train loss: 0.01175606895238161, Test loss: 0.020563708618283272\n", + "Starting epoch 4402\n", + "Train loss: 0.01171889754012227, Test loss: 0.02391151525080204\n", + "Starting epoch 4403\n", + "Train loss: 0.011707060523331165, Test loss: 0.010636589489877224\n", + "Starting epoch 4404\n", + "Train loss: 0.011733161509037018, Test loss: 0.011456654407083988\n", + "Starting epoch 4405\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.011751420851796866, Test loss: 0.010988809168338776\n", + "Starting epoch 4406\n", + "Train loss: 0.011751287840306759, Test loss: 0.012067648582160473\n", + "Starting epoch 4407\n", + "Train loss: 0.011752632576972247, Test loss: 0.010960903950035572\n", + "Starting epoch 4408\n", + "Train loss: 0.011697792466729879, Test loss: 0.011089419014751911\n", + "Starting epoch 4409\n", + "Train loss: 0.011694274470210075, Test loss: 0.012026555836200714\n", + "Starting epoch 4410\n", + "Train loss: 0.011743658669292927, Test loss: 0.01143211405724287\n", + "Starting epoch 4411\n", + "Train loss: 0.01168901501223445, Test loss: 0.010612710379064083\n", + "Starting epoch 4412\n", + "Train loss: 0.011751311589032412, Test loss: 0.011621850542724133\n", + "Starting epoch 4413\n", + "Train loss: 0.011789113152772188, Test loss: 0.011930488981306553\n", + "Starting epoch 4414\n", + "Train loss: 0.011744809616357089, Test loss: 0.013016238808631897\n", + "Starting epoch 4415\n", + "Train loss: 0.011816771794110536, Test loss: 0.011003734543919563\n", + "Starting epoch 4416\n", + "Train loss: 0.011824587602168321, Test loss: 0.010469581000506878\n", + "Starting epoch 4417\n", + "Train loss: 0.011701768804341555, Test loss: 0.010764958336949348\n", + "Starting epoch 4418\n", + "Train loss: 0.011680966392159462, Test loss: 0.011690412648022175\n", + "Starting epoch 4419\n", + "Train loss: 0.01177499882876873, Test loss: 0.010759168304502964\n", + "Starting epoch 4420\n", + "Train loss: 0.011700345315039157, Test loss: 0.010674986056983471\n", + "Starting epoch 4421\n", + "Train loss: 0.011750382706522941, Test loss: 0.019441599026322365\n", + "Starting epoch 4422\n", + "Train loss: 0.011832358315587044, Test loss: 0.011190813034772873\n", + "Starting epoch 4423\n", + "Train loss: 0.011744111366569996, Test loss: 0.011648531071841717\n", + "Starting epoch 4424\n", + "Train loss: 0.011809318270534276, Test loss: 0.011655583046376705\n", + "Starting epoch 4425\n", + "Train loss: 0.012671357188373804, Test loss: 0.016971327364444733\n", + "Starting epoch 4426\n", + "Train loss: 0.011873631812632084, Test loss: 0.016522986814379692\n", + "Starting epoch 4427\n", + "Train loss: 0.011865464132279157, Test loss: 0.01698145642876625\n", + "Starting epoch 4428\n", + "Train loss: 0.011742176674306392, Test loss: 0.010938701219856739\n", + "Starting epoch 4429\n", + "Train loss: 0.011891483180224895, Test loss: 0.010954158380627632\n", + "Starting epoch 4430\n", + "Train loss: 0.011851938292384148, Test loss: 0.010806255042552948\n", + "Starting epoch 4431\n", + "Train loss: 0.01179672395810485, Test loss: 0.01075215544551611\n", + "Starting epoch 4432\n", + "Train loss: 0.011632369793951512, Test loss: 0.01062279287725687\n", + "Starting epoch 4433\n", + "Train loss: 0.011810723822563887, Test loss: 0.012271476909518242\n", + "Starting epoch 4434\n", + "Train loss: 0.011743293218314648, Test loss: 0.019994419068098068\n", + "Starting epoch 4435\n", + "Train loss: 0.011767163425683975, Test loss: 0.018979983404278755\n", + "Starting epoch 4436\n", + "Train loss: 0.011769819278270007, Test loss: 0.011282026767730713\n", + "Starting epoch 4437\n", + "Train loss: 0.011844610292464496, Test loss: 0.010548424907028675\n", + "Starting epoch 4438\n", + "Train loss: 0.011748903542757034, Test loss: 0.01095257792621851\n", + "Starting epoch 4439\n", + "Train loss: 0.0116914676502347, Test loss: 0.010431617498397827\n", + "Starting epoch 4440\n", + "Train loss: 0.011765319406986236, Test loss: 0.012129864655435085\n", + "Starting epoch 4441\n", + "Train loss: 0.011688242424279452, Test loss: 0.011169254779815674\n", + "Starting epoch 4442\n", + "Train loss: 0.011655483860522509, Test loss: 0.010694555938243866\n", + "Starting epoch 4443\n", + "Train loss: 0.011658106967806815, Test loss: 0.0110899917781353\n", + "Starting epoch 4444\n", + "Train loss: 0.011652841735631227, Test loss: 0.01118314266204834\n", + "Starting epoch 4445\n", + "Train loss: 0.011654445305466651, Test loss: 0.014303691685199738\n", + "Starting epoch 4446\n", + "Train loss: 0.011664811242371797, Test loss: 0.009838924743235111\n", + "Starting epoch 4447\n", + "Train loss: 0.011627014875411987, Test loss: 0.0127258924767375\n", + "Starting epoch 4448\n", + "Train loss: 0.011669111829251051, Test loss: 0.010638946667313576\n", + "Starting epoch 4449\n", + "Train loss: 0.011712958719581366, Test loss: 0.012667566537857056\n", + "Starting epoch 4450\n", + "Train loss: 0.011596644259989262, Test loss: 0.010691138915717602\n", + "Starting epoch 4451\n", + "Train loss: 0.011645614895969629, Test loss: 0.010408082976937294\n", + "Starting epoch 4452\n", + "Train loss: 0.011620463691651822, Test loss: 0.010158888064324856\n", + "Starting epoch 4453\n", + "Train loss: 0.01164411410689354, Test loss: 0.010609078221023083\n", + "Starting epoch 4454\n", + "Train loss: 0.011601350139826536, Test loss: 0.020366255193948746\n", + "Starting epoch 4455\n", + "Train loss: 0.011600878592580556, Test loss: 0.023281672969460487\n", + "Starting epoch 4456\n", + "Train loss: 0.011585549898445607, Test loss: 0.018167566508054733\n", + "Starting epoch 4457\n", + "Train loss: 0.01164164012297988, Test loss: 0.020475436002016068\n", + "Starting epoch 4458\n", + "Train loss: 0.011609673891216516, Test loss: 0.0213017538189888\n", + "Starting epoch 4459\n", + "Train loss: 0.011528613623231649, Test loss: 0.015040713362395763\n", + "Starting epoch 4460\n", + "Train loss: 0.011815149690955878, Test loss: 0.014489351771771908\n", + "Starting epoch 4461\n", + "Train loss: 0.01187489164993167, Test loss: 0.012014641426503658\n", + "Starting epoch 4462\n", + "Train loss: 0.011823412384837867, Test loss: 0.011910013854503632\n", + "Starting epoch 4463\n", + "Train loss: 0.011800298597663642, Test loss: 0.011829713359475136\n", + "Starting epoch 4464\n", + "Train loss: 0.011746814176440239, Test loss: 0.00998206902295351\n", + "Starting epoch 4465\n", + "Train loss: 0.011735223978757859, Test loss: 0.011571746319532394\n", + "Starting epoch 4466\n", + "Train loss: 0.01160666298121214, Test loss: 0.010226062498986721\n", + "Starting epoch 4467\n", + "Train loss: 0.01164545889943838, Test loss: 0.015117025002837181\n", + "Starting epoch 4468\n", + "Train loss: 0.011695060543715954, Test loss: 0.010720634832978249\n", + "Starting epoch 4469\n", + "Train loss: 0.011650251857936382, Test loss: 0.01124437153339386\n", + "Starting epoch 4470\n", + "Train loss: 0.011672834120690822, Test loss: 0.011022518388926983\n", + "Starting epoch 4471\n", + "Train loss: 0.011554866135120391, Test loss: 0.011519168503582478\n", + "Starting epoch 4472\n", + "Train loss: 0.011706685777753592, Test loss: 0.011283451691269875\n", + "Starting epoch 4473\n", + "Train loss: 0.01168004745617509, Test loss: 0.01117036398500204\n", + "Starting epoch 4474\n", + "Train loss: 0.011684112101793289, Test loss: 0.009858246892690659\n", + "Starting epoch 4475\n", + "Train loss: 0.01166427504271269, Test loss: 0.010347861796617508\n", + "Starting epoch 4476\n", + "Train loss: 0.011641682982444763, Test loss: 0.009871234185993671\n", + "Starting epoch 4477\n", + "Train loss: 0.011553616598248481, Test loss: 0.011160945519804955\n", + "Starting epoch 4478\n", + "Train loss: 0.011595897264778615, Test loss: 0.012727443128824234\n", + "Starting epoch 4479\n", + "Train loss: 0.011636161226779223, Test loss: 0.011916250921785831\n", + "Starting epoch 4480\n", + "Train loss: 0.011662437375634908, Test loss: 0.011340221390128136\n", + "Starting epoch 4481\n", + "Train loss: 0.011663116700947284, Test loss: 0.009551026858389378\n", + "Starting epoch 4482\n", + "Train loss: 0.011556448265910148, Test loss: 0.010879919864237309\n", + "Starting epoch 4483\n", + "Train loss: 0.011586965005844831, Test loss: 0.010589850135147572\n", + "Starting epoch 4484\n", + "Train loss: 0.012012687101960183, Test loss: 0.013982781209051609\n", + "Starting epoch 4485\n", + "Train loss: 0.011684680003672837, Test loss: 0.012198596261441708\n", + "Starting epoch 4486\n", + "Train loss: 0.01170420179143548, Test loss: 0.010360799729824066\n", + "Starting epoch 4487\n", + "Train loss: 0.011637838166207075, Test loss: 0.013795129954814911\n", + "Starting epoch 4488\n", + "Train loss: 0.011593162193894386, Test loss: 0.01059787254780531\n", + "Starting epoch 4489\n", + "Train loss: 0.01167556444182992, Test loss: 0.01096472330391407\n", + "Starting epoch 4490\n", + "Train loss: 0.011648204885423183, Test loss: 0.011423921212553978\n", + "Starting epoch 4491\n", + "Train loss: 0.011646295189857483, Test loss: 0.010676747187972069\n", + "Starting epoch 4492\n", + "Train loss: 0.011523684617131949, Test loss: 0.018238231539726257\n", + "Starting epoch 4493\n", + "Train loss: 0.011596977077424526, Test loss: 0.016502054408192635\n", + "Starting epoch 4494\n", + "Train loss: 0.01161217801272869, Test loss: 0.019845107570290565\n", + "Starting epoch 4495\n", + "Train loss: 0.011584955360740423, Test loss: 0.01856808178126812\n", + "Starting epoch 4496\n", + "Train loss: 0.011529915276914835, Test loss: 0.018715983256697655\n", + "Starting epoch 4497\n", + "Train loss: 0.011597808301448822, Test loss: 0.019147144630551338\n", + "Starting epoch 4498\n", + "Train loss: 0.011578247640281916, Test loss: 0.02078869193792343\n", + "Starting epoch 4499\n", + "Train loss: 0.01157254133373499, Test loss: 0.017491720616817474\n", + "Starting epoch 4500\n", + "Train loss: 0.011533647011965513, Test loss: 0.020022954791784286\n", + "Starting epoch 4501\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.011555122639983893, Test loss: 0.020763512700796127\n", + "Starting epoch 4502\n", + "Train loss: 0.011284705642610788, Test loss: 0.011462301947176456\n", + "Starting epoch 4503\n", + "Train loss: 0.01159784061834216, Test loss: 0.010365184396505356\n", + "Starting epoch 4504\n", + "Train loss: 0.011581802051514387, Test loss: 0.01170391496270895\n", + "Starting epoch 4505\n", + "Train loss: 0.011496900636702776, Test loss: 0.015995141118764877\n", + "Starting epoch 4506\n", + "Train loss: 0.011605063173919915, Test loss: 0.02039930783212185\n", + "Starting epoch 4507\n", + "Train loss: 0.011537615898996592, Test loss: 0.02126133441925049\n", + "Starting epoch 4508\n", + "Train loss: 0.011532463673502207, Test loss: 0.016018595546483994\n", + "Starting epoch 4509\n", + "Train loss: 0.011526990961283446, Test loss: 0.020740177482366562\n", + "Starting epoch 4510\n", + "Train loss: 0.011561912186443805, Test loss: 0.02013457380235195\n", + "Starting epoch 4511\n", + "Train loss: 0.011488204915076494, Test loss: 0.010483822785317898\n", + "Starting epoch 4512\n", + "Train loss: 0.011611829716712237, Test loss: 0.01011221669614315\n", + "Starting epoch 4513\n", + "Train loss: 0.011547773964703084, Test loss: 0.01217712927609682\n", + "Starting epoch 4514\n", + "Train loss: 0.011623284220695496, Test loss: 0.011194209568202496\n", + "Starting epoch 4515\n", + "Train loss: 0.011623149309307336, Test loss: 0.010881267488002777\n", + "Starting epoch 4516\n", + "Train loss: 0.011757619380950927, Test loss: 0.010675960220396519\n", + "Starting epoch 4517\n", + "Train loss: 0.011644510626792908, Test loss: 0.011034680530428886\n", + "Starting epoch 4518\n", + "Train loss: 0.011623571421951055, Test loss: 0.011456510052084923\n", + "Starting epoch 4519\n", + "Train loss: 0.011583450902253389, Test loss: 0.013148204423487186\n", + "Starting epoch 4520\n", + "Train loss: 0.01158820889890194, Test loss: 0.010667076334357262\n", + "Starting epoch 4521\n", + "Train loss: 0.011570926252752543, Test loss: 0.012032643891870975\n", + "Starting epoch 4522\n", + "Train loss: 0.01154626753181219, Test loss: 0.012250043451786041\n", + "Starting epoch 4523\n", + "Train loss: 0.011499590501189231, Test loss: 0.011223693378269672\n", + "Starting epoch 4524\n", + "Train loss: 0.011549609806388616, Test loss: 0.011557111516594887\n", + "Starting epoch 4525\n", + "Train loss: 0.011472428496927024, Test loss: 0.0109446682035923\n", + "Starting epoch 4526\n", + "Train loss: 0.011362543255090713, Test loss: 0.01825459487736225\n", + "Starting epoch 4527\n", + "Train loss: 0.011572167798876763, Test loss: 0.02121727354824543\n", + "Starting epoch 4528\n", + "Train loss: 0.011559970937669277, Test loss: 0.015218539163470268\n", + "Starting epoch 4529\n", + "Train loss: 0.011438221856951713, Test loss: 0.018267463892698288\n", + "Starting epoch 4530\n", + "Train loss: 0.011528879832476378, Test loss: 0.020810279995203018\n", + "Starting epoch 4531\n", + "Train loss: 0.011535875257104635, Test loss: 0.019526146352291107\n", + "Starting epoch 4532\n", + "Train loss: 0.011459150705486537, Test loss: 0.019656384363770485\n", + "Starting epoch 4533\n", + "Train loss: 0.011477020271122455, Test loss: 0.015944888815283775\n", + "Starting epoch 4534\n", + "Train loss: 0.01148381533101201, Test loss: 0.02289782278239727\n", + "Starting epoch 4535\n", + "Train loss: 0.01146626701578498, Test loss: 0.01820092275738716\n", + "Starting epoch 4536\n", + "Train loss: 0.011469545383006334, Test loss: 0.01875334233045578\n", + "Starting epoch 4537\n", + "Train loss: 0.011493526380509139, Test loss: 0.01763847842812538\n", + "Starting epoch 4538\n", + "Train loss: 0.011434089746326209, Test loss: 0.01987217180430889\n", + "Starting epoch 4539\n", + "Train loss: 0.01149669935926795, Test loss: 0.02090114913880825\n", + "Starting epoch 4540\n", + "Train loss: 0.011443242859095336, Test loss: 0.02158983424305916\n", + "Starting epoch 4541\n", + "Train loss: 0.011429041102528572, Test loss: 0.02308199740946293\n", + "Starting epoch 4542\n", + "Train loss: 0.011459491811692715, Test loss: 0.018587609753012657\n", + "Starting epoch 4543\n", + "Train loss: 0.011441950406879186, Test loss: 0.017382023856043816\n", + "Starting epoch 4544\n", + "Train loss: 0.011432963702827693, Test loss: 0.019730309024453163\n", + "Starting epoch 4545\n", + "Train loss: 0.011431371290236712, Test loss: 0.020327456295490265\n", + "Starting epoch 4546\n", + "Train loss: 0.011436081808060407, Test loss: 0.021951992064714432\n", + "Starting epoch 4547\n", + "Train loss: 0.011432014368474484, Test loss: 0.01888214983046055\n", + "Starting epoch 4548\n", + "Train loss: 0.01139073984697461, Test loss: 0.020350594073534012\n", + "Starting epoch 4549\n", + "Train loss: 0.011339336112141609, Test loss: 0.021621860563755035\n", + "Starting epoch 4550\n", + "Train loss: 0.01149597903713584, Test loss: 0.02057134546339512\n", + "Starting epoch 4551\n", + "Train loss: 0.011507910899817944, Test loss: 0.018950464203953743\n", + "Starting epoch 4552\n", + "Train loss: 0.011470561865717172, Test loss: 0.01870688796043396\n", + "Starting epoch 4553\n", + "Train loss: 0.011430213134735823, Test loss: 0.0206444188952446\n", + "Starting epoch 4554\n", + "Train loss: 0.011436357516795398, Test loss: 0.01929016225039959\n", + "Starting epoch 4555\n", + "Train loss: 0.011528606992214918, Test loss: 0.017602358013391495\n", + "Starting epoch 4556\n", + "Train loss: 0.0115173445828259, Test loss: 0.019254038110375404\n", + "Starting epoch 4557\n", + "Train loss: 0.011451238729059697, Test loss: 0.020745418965816498\n", + "Starting epoch 4558\n", + "Train loss: 0.01149893194437027, Test loss: 0.01871541701257229\n", + "Starting epoch 4559\n", + "Train loss: 0.011492964792996645, Test loss: 0.01800873503088951\n", + "Starting epoch 4560\n", + "Train loss: 0.011422332860529422, Test loss: 0.01884935051202774\n", + "Starting epoch 4561\n", + "Train loss: 0.011434880699962378, Test loss: 0.020296521484851837\n", + "Starting epoch 4562\n", + "Train loss: 0.011455824039876461, Test loss: 0.01450782548636198\n", + "Starting epoch 4563\n", + "Train loss: 0.011439287289977073, Test loss: 0.018805470317602158\n", + "Starting epoch 4564\n", + "Train loss: 0.01143094576895237, Test loss: 0.01973351277410984\n", + "Starting epoch 4565\n", + "Train loss: 0.011398352459073067, Test loss: 0.020671604201197624\n", + "Starting epoch 4566\n", + "Train loss: 0.011463390905410051, Test loss: 0.018052034080028534\n", + "Starting epoch 4567\n", + "Train loss: 0.01141427306458354, Test loss: 0.01889912597835064\n", + "Starting epoch 4568\n", + "Train loss: 0.0114428785815835, Test loss: 0.01959344372153282\n", + "Starting epoch 4569\n", + "Train loss: 0.011420282814651728, Test loss: 0.020876074209809303\n", + "Starting epoch 4570\n", + "Train loss: 0.01147136565297842, Test loss: 0.0206462349742651\n", + "Starting epoch 4571\n", + "Train loss: 0.011438844781368971, Test loss: 0.02076081931591034\n", + "Starting epoch 4572\n", + "Train loss: 0.01143084904178977, Test loss: 0.019768869504332542\n", + "Starting epoch 4573\n", + "Train loss: 0.011441917736083268, Test loss: 0.016280336305499077\n", + "Starting epoch 4574\n", + "Train loss: 0.011349425856024027, Test loss: 0.017324315384030342\n", + "Starting epoch 4575\n", + "Train loss: 0.017012288477271795, Test loss: 0.022883960977196693\n", + "Starting epoch 4576\n", + "Train loss: 0.0167709187977016, Test loss: 0.016192635521292686\n", + "Starting epoch 4577\n", + "Train loss: 0.014307827856391668, Test loss: 0.017149480059742928\n", + "Starting epoch 4578\n", + "Train loss: 0.012570017892867326, Test loss: 0.01772172376513481\n", + "Starting epoch 4579\n", + "Train loss: 0.012066222690045834, Test loss: 0.017664607614278793\n", + "Starting epoch 4580\n", + "Train loss: 0.011917160265147686, Test loss: 0.01768619753420353\n", + "Starting epoch 4581\n", + "Train loss: 0.01185671441257, Test loss: 0.018521836027503014\n", + "Starting epoch 4582\n", + "Train loss: 0.01179768079891801, Test loss: 0.022791173309087753\n", + "Starting epoch 4583\n", + "Train loss: 0.011715843696147204, Test loss: 0.019381079822778702\n", + "Starting epoch 4584\n", + "Train loss: 0.011699939724057912, Test loss: 0.017733220010995865\n", + "Starting epoch 4585\n", + "Train loss: 0.01160982348024845, Test loss: 0.02063794992864132\n", + "Starting epoch 4586\n", + "Train loss: 0.011615312304347753, Test loss: 0.01912059262394905\n", + "Starting epoch 4587\n", + "Train loss: 0.012646719124168157, Test loss: 0.012018028646707535\n", + "Starting epoch 4588\n", + "Train loss: 0.011515355557203294, Test loss: 0.02161695621907711\n", + "Starting epoch 4589\n", + "Train loss: 0.01155133755877614, Test loss: 0.01915047876536846\n", + "Starting epoch 4590\n", + "Train loss: 0.011548949535936118, Test loss: 0.01924203708767891\n", + "Starting epoch 4591\n", + "Train loss: 0.011457481179386378, Test loss: 0.01930973306298256\n", + "Starting epoch 4592\n", + "Train loss: 0.01142144551500678, Test loss: 0.018726112321019173\n", + "Starting epoch 4593\n", + "Train loss: 0.01139573508873582, Test loss: 0.02016577497124672\n", + "Starting epoch 4594\n", + "Train loss: 0.011383435111492873, Test loss: 0.010963344015181065\n", + "Starting epoch 4595\n", + "Train loss: 0.011381699703633785, Test loss: 0.011905457824468613\n", + "Starting epoch 4596\n", + "Train loss: 0.011448086667805911, Test loss: 0.011576521210372448\n", + "Starting epoch 4597\n", + "Train loss: 0.011387490667402744, Test loss: 0.010152129456400871\n", + "Starting epoch 4598\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.011336114518344402, Test loss: 0.01006277371197939\n", + "Starting epoch 4599\n", + "Train loss: 0.011300844550132751, Test loss: 0.012335509061813354\n", + "Starting epoch 4600\n", + "Train loss: 0.011362929567694665, Test loss: 0.010932382196187973\n", + "Starting epoch 4601\n", + "Train loss: 0.011297276858240366, Test loss: 0.010287836194038391\n", + "Starting epoch 4602\n", + "Train loss: 0.011329960227012634, Test loss: 0.010049916803836823\n", + "Starting epoch 4603\n", + "Train loss: 0.011329613253474235, Test loss: 0.012435918673872948\n", + "Starting epoch 4604\n", + "Train loss: 0.011293733958154917, Test loss: 0.010762170888483524\n", + "Starting epoch 4605\n", + "Train loss: 0.011320531237870454, Test loss: 0.009748956188559532\n", + "Starting epoch 4606\n", + "Train loss: 0.011310527194291353, Test loss: 0.01173222716897726\n", + "Starting epoch 4607\n", + "Train loss: 0.011382991448044777, Test loss: 0.01012270338833332\n", + "Starting epoch 4608\n", + "Train loss: 0.011367028281092643, Test loss: 0.010027172975242138\n", + "Starting epoch 4609\n", + "Train loss: 0.011281807962805033, Test loss: 0.011409255675971508\n", + "Starting epoch 4610\n", + "Train loss: 0.011339045334607363, Test loss: 0.010067308321595192\n", + "Starting epoch 4611\n", + "Train loss: 0.011297774016857148, Test loss: 0.010365204885601997\n", + "Starting epoch 4612\n", + "Train loss: 0.011300209090113639, Test loss: 0.010132260620594025\n", + "Starting epoch 4613\n", + "Train loss: 0.011272859890013934, Test loss: 0.011933078989386559\n", + "Starting epoch 4614\n", + "Train loss: 0.011269498839974404, Test loss: 0.010026748292148113\n", + "Starting epoch 4615\n", + "Train loss: 0.012110788095742463, Test loss: 0.019685856997966766\n", + "Starting epoch 4616\n", + "Train loss: 0.011483576856553554, Test loss: 0.02128494344651699\n", + "Starting epoch 4617\n", + "Train loss: 0.011437146384268999, Test loss: 0.019703632220625877\n", + "Starting epoch 4618\n", + "Train loss: 0.011405764799565076, Test loss: 0.016589131206274033\n", + "Starting epoch 4619\n", + "Train loss: 0.011348026711493731, Test loss: 0.017097700387239456\n", + "Starting epoch 4620\n", + "Train loss: 0.011365377102047205, Test loss: 0.01685056835412979\n", + "Starting epoch 4621\n", + "Train loss: 0.011298820339143277, Test loss: 0.018878454342484474\n", + "Starting epoch 4622\n", + "Train loss: 0.011327582560479641, Test loss: 0.018684301525354385\n", + "Starting epoch 4623\n", + "Train loss: 0.011336475275456906, Test loss: 0.0181695818901062\n", + "Starting epoch 4624\n", + "Train loss: 0.011259011812508107, Test loss: 0.015490666963160038\n", + "Starting epoch 4625\n", + "Train loss: 0.011280220318585634, Test loss: 0.018069176003336906\n", + "Starting epoch 4626\n", + "Train loss: 0.011282462663948535, Test loss: 0.021794497966766357\n", + "Starting epoch 4627\n", + "Train loss: 0.011259652264416218, Test loss: 0.014100895263254642\n", + "Starting epoch 4628\n", + "Train loss: 0.011188840717077255, Test loss: 0.020971808582544327\n", + "Starting epoch 4629\n", + "Train loss: 0.011309200339019299, Test loss: 0.019235549494624138\n", + "Starting epoch 4630\n", + "Train loss: 0.011329701263457536, Test loss: 0.01773972250521183\n", + "Starting epoch 4631\n", + "Train loss: 0.011240372881293297, Test loss: 0.02098257839679718\n", + "Starting epoch 4632\n", + "Train loss: 0.011279711164534093, Test loss: 0.019423944875597954\n", + "Starting epoch 4633\n", + "Train loss: 0.01121498629450798, Test loss: 0.020528271794319153\n", + "Starting epoch 4634\n", + "Train loss: 0.011249720323830844, Test loss: 0.018483567982912064\n", + "Starting epoch 4635\n", + "Train loss: 0.011244853921234609, Test loss: 0.018268249928951263\n", + "Starting epoch 4636\n", + "Train loss: 0.011212655212730169, Test loss: 0.02013484761118889\n", + "Starting epoch 4637\n", + "Train loss: 0.01126533443108201, Test loss: 0.01719895750284195\n", + "Starting epoch 4638\n", + "Train loss: 0.011260459758341312, Test loss: 0.01781488209962845\n", + "Starting epoch 4639\n", + "Train loss: 0.011111400239169598, Test loss: 0.018864063546061516\n", + "Starting epoch 4640\n", + "Train loss: 0.011300483960658312, Test loss: 0.020469265058636665\n", + "Starting epoch 4641\n", + "Train loss: 0.011332272682338954, Test loss: 0.01910105161368847\n", + "Starting epoch 4642\n", + "Train loss: 0.011292869858443737, Test loss: 0.018142078071832657\n", + "Starting epoch 4643\n", + "Train loss: 0.011278483308851719, Test loss: 0.02132508158683777\n", + "Starting epoch 4644\n", + "Train loss: 0.011330030430108308, Test loss: 0.01804198883473873\n", + "Starting epoch 4645\n", + "Train loss: 0.011292329765856266, Test loss: 0.017016137018799782\n", + "Starting epoch 4646\n", + "Train loss: 0.011270531192421914, Test loss: 0.018583647906780243\n", + "Starting epoch 4647\n", + "Train loss: 0.01127310024574399, Test loss: 0.01898740977048874\n", + "Starting epoch 4648\n", + "Train loss: 0.011254717297852039, Test loss: 0.020929934456944466\n", + "Starting epoch 4649\n", + "Train loss: 0.011215651910752057, Test loss: 0.01917997933924198\n", + "Starting epoch 4650\n", + "Train loss: 0.011237328480929136, Test loss: 0.015353706665337086\n", + "Starting epoch 4651\n", + "Train loss: 0.011298519838601351, Test loss: 0.021604595705866814\n", + "Starting epoch 4652\n", + "Train loss: 0.011217558830976486, Test loss: 0.018646948039531708\n", + "Starting epoch 4653\n", + "Train loss: 0.011203049216419458, Test loss: 0.01708127185702324\n", + "Starting epoch 4654\n", + "Train loss: 0.011388680562376977, Test loss: 0.012349582277238369\n", + "Starting epoch 4655\n", + "Train loss: 0.011314226482063532, Test loss: 0.013094612397253513\n", + "Starting epoch 4656\n", + "Train loss: 0.011136895790696145, Test loss: 0.020429758355021477\n", + "Starting epoch 4657\n", + "Train loss: 0.011293455846607686, Test loss: 0.01811208575963974\n", + "Starting epoch 4658\n", + "Train loss: 0.011299614626914263, Test loss: 0.019060825929045677\n", + "Starting epoch 4659\n", + "Train loss: 0.011233653854578734, Test loss: 0.018572768196463585\n", + "Starting epoch 4660\n", + "Train loss: 0.011218113042414189, Test loss: 0.022061878815293312\n", + "Starting epoch 4661\n", + "Train loss: 0.01122415840625763, Test loss: 0.017290081828832626\n", + "Starting epoch 4662\n", + "Train loss: 0.011209021899849176, Test loss: 0.019390573725104332\n", + "Starting epoch 4663\n", + "Train loss: 0.011183451022952795, Test loss: 0.022508494555950165\n", + "Starting epoch 4664\n", + "Train loss: 0.011192955151200295, Test loss: 0.0215251874178648\n", + "Starting epoch 4665\n", + "Train loss: 0.011178689487278461, Test loss: 0.01930978335440159\n", + "Starting epoch 4666\n", + "Train loss: 0.011204576529562473, Test loss: 0.020820219069719315\n", + "Starting epoch 4667\n", + "Train loss: 0.011266901530325413, Test loss: 0.017125878483057022\n", + "Starting epoch 4668\n", + "Train loss: 0.011188504733145237, Test loss: 0.018384797498583794\n", + "Starting epoch 4669\n", + "Train loss: 0.011194557771086694, Test loss: 0.017808042466640472\n", + "Starting epoch 4670\n", + "Train loss: 0.011144142951816321, Test loss: 0.02150966413319111\n", + "Starting epoch 4671\n", + "Train loss: 0.0111974236369133, Test loss: 0.02291993610560894\n", + "Starting epoch 4672\n", + "Train loss: 0.01118228917941451, Test loss: 0.01514597050845623\n", + "Starting epoch 4673\n", + "Train loss: 0.011163477730005979, Test loss: 0.020481083542108536\n", + "Starting epoch 4674\n", + "Train loss: 0.011328430771827698, Test loss: 0.019167441874742508\n", + "Starting epoch 4675\n", + "Train loss: 0.011239432319998742, Test loss: 0.017047980800271034\n", + "Starting epoch 4676\n", + "Train loss: 0.011172140426933765, Test loss: 0.019472908228635788\n", + "Starting epoch 4677\n", + "Train loss: 0.011241902448236942, Test loss: 0.014880357310175896\n", + "Starting epoch 4678\n", + "Train loss: 0.01119201773777604, Test loss: 0.012660476379096508\n", + "Starting epoch 4679\n", + "Train loss: 0.011222502812743188, Test loss: 0.010862062685191631\n", + "Starting epoch 4680\n", + "Train loss: 0.011226024571806192, Test loss: 0.013253905810415745\n", + "Starting epoch 4681\n", + "Train loss: 0.011197775825858116, Test loss: 0.009498918429017067\n", + "Starting epoch 4682\n", + "Train loss: 0.011071128789335489, Test loss: 0.010289106518030167\n", + "Starting epoch 4683\n", + "Train loss: 0.011117386687546969, Test loss: 0.01167314499616623\n", + "Starting epoch 4684\n", + "Train loss: 0.011215630862861871, Test loss: 0.011490363627672195\n", + "Starting epoch 4685\n", + "Train loss: 0.011141076125204563, Test loss: 0.01097904797643423\n", + "Starting epoch 4686\n", + "Train loss: 0.011200700961053371, Test loss: 0.010936321690678596\n", + "Starting epoch 4687\n", + "Train loss: 0.011118472646921873, Test loss: 0.011209775693714619\n", + "Starting epoch 4688\n", + "Train loss: 0.011184995882213116, Test loss: 0.010955396108329296\n", + "Starting epoch 4689\n", + "Train loss: 0.011147639881819486, Test loss: 0.015175661072134972\n", + "Starting epoch 4690\n", + "Train loss: 0.01109602950513363, Test loss: 0.01038400363177061\n", + "Starting epoch 4691\n", + "Train loss: 0.011100291255861521, Test loss: 0.010809739120304585\n", + "Starting epoch 4692\n", + "Train loss: 0.011079300306737422, Test loss: 0.011920087039470673\n", + "Starting epoch 4693\n", + "Train loss: 0.01113598445430398, Test loss: 0.011861774139106274\n", + "Starting epoch 4694\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.011146914642304183, Test loss: 0.011464406736195087\n", + "Starting epoch 4695\n", + "Train loss: 0.011102070640772581, Test loss: 0.014693391509354115\n", + "Starting epoch 4696\n", + "Train loss: 0.011132353655993939, Test loss: 0.009738869033753872\n", + "Starting epoch 4697\n", + "Train loss: 0.011258995607495308, Test loss: 0.011633637361228466\n", + "Starting epoch 4698\n", + "Train loss: 0.011223022360354662, Test loss: 0.011861789971590042\n", + "Starting epoch 4699\n", + "Train loss: 0.011177255157381296, Test loss: 0.011958570219576359\n", + "Starting epoch 4700\n", + "Train loss: 0.011155514921993017, Test loss: 0.0113062160089612\n", + "Starting epoch 4701\n", + "Train loss: 0.011029124800115824, Test loss: 0.015767117962241173\n", + "Starting epoch 4702\n", + "Train loss: 0.011250243093818425, Test loss: 0.018697179853916168\n", + "Starting epoch 4703\n", + "Train loss: 0.011214474383741617, Test loss: 0.0186581090092659\n", + "Starting epoch 4704\n", + "Train loss: 0.011219683643430471, Test loss: 0.018616901710629463\n", + "Starting epoch 4705\n", + "Train loss: 0.011182791013270616, Test loss: 0.0193357914686203\n", + "Starting epoch 4706\n", + "Train loss: 0.01116786265745759, Test loss: 0.01828479953110218\n", + "Starting epoch 4707\n", + "Train loss: 0.01114117942750454, Test loss: 0.017242245376110077\n", + "Starting epoch 4708\n", + "Train loss: 0.011162792667746543, Test loss: 0.01885993778705597\n", + "Starting epoch 4709\n", + "Train loss: 0.01107546839863062, Test loss: 0.019030530005693436\n", + "Starting epoch 4710\n", + "Train loss: 0.011093908343464137, Test loss: 0.01645028404891491\n", + "Starting epoch 4711\n", + "Train loss: 0.011102585252374411, Test loss: 0.02283419482409954\n", + "Starting epoch 4712\n", + "Train loss: 0.01108234589919448, Test loss: 0.014501725323498249\n", + "Starting epoch 4713\n", + "Train loss: 0.011095272190868855, Test loss: 0.018221870064735413\n", + "Starting epoch 4714\n", + "Train loss: 0.011054479200392962, Test loss: 0.019731078296899796\n", + "Starting epoch 4715\n", + "Train loss: 0.011078998185694218, Test loss: 0.01522731687873602\n", + "Starting epoch 4716\n", + "Train loss: 0.011114455331116914, Test loss: 0.019349005073308945\n", + "Starting epoch 4717\n", + "Train loss: 0.011127618178725242, Test loss: 0.01633620448410511\n", + "Starting epoch 4718\n", + "Train loss: 0.011135117448866368, Test loss: 0.0184008926153183\n", + "Starting epoch 4719\n", + "Train loss: 0.011094216015189886, Test loss: 0.019636085256934166\n", + "Starting epoch 4720\n", + "Train loss: 0.011137496028095484, Test loss: 0.019529227167367935\n", + "Starting epoch 4721\n", + "Train loss: 0.01107540886849165, Test loss: 0.019006595015525818\n", + "Starting epoch 4722\n", + "Train loss: 0.011034633293747901, Test loss: 0.01916591078042984\n", + "Starting epoch 4723\n", + "Train loss: 0.011166917160153389, Test loss: 0.010706320405006409\n", + "Starting epoch 4724\n", + "Train loss: 0.01109702030196786, Test loss: 0.020499855279922485\n", + "Starting epoch 4725\n", + "Train loss: 0.011129058469086885, Test loss: 0.02062167413532734\n", + "Starting epoch 4726\n", + "Train loss: 0.011179920975118876, Test loss: 0.019378356635570526\n", + "Starting epoch 4727\n", + "Train loss: 0.011124588698148728, Test loss: 0.022280432283878326\n", + "Starting epoch 4728\n", + "Train loss: 0.011130222957581282, Test loss: 0.017742061987519264\n", + "Starting epoch 4729\n", + "Train loss: 0.011059139035642146, Test loss: 0.01883317343890667\n", + "Starting epoch 4730\n", + "Train loss: 0.011088795531541108, Test loss: 0.019945669919252396\n", + "Starting epoch 4731\n", + "Train loss: 0.011090529728680848, Test loss: 0.018237272277474403\n", + "Starting epoch 4732\n", + "Train loss: 0.011014087367802858, Test loss: 0.01388685405254364\n", + "Starting epoch 4733\n", + "Train loss: 0.010998062621802092, Test loss: 0.02234513871371746\n", + "Starting epoch 4734\n", + "Train loss: 0.011082039903849364, Test loss: 0.017527367919683456\n", + "Starting epoch 4735\n", + "Train loss: 0.011032562367618085, Test loss: 0.018450438976287842\n", + "Starting epoch 4736\n", + "Train loss: 0.011029092222452163, Test loss: 0.018103355541825294\n", + "Starting epoch 4737\n", + "Train loss: 0.010977731645107269, Test loss: 0.018071135506033897\n", + "Starting epoch 4738\n", + "Train loss: 0.01104793393984437, Test loss: 0.020135818049311638\n", + "Starting epoch 4739\n", + "Train loss: 0.011055402178317308, Test loss: 0.014547741040587425\n", + "Starting epoch 4740\n", + "Train loss: 0.011006950456649066, Test loss: 0.01865422911942005\n", + "Starting epoch 4741\n", + "Train loss: 0.011326636262238026, Test loss: 0.01526787132024765\n", + "Starting epoch 4742\n", + "Train loss: 0.01119898185133934, Test loss: 0.018016446381807327\n", + "Starting epoch 4743\n", + "Train loss: 0.011155985537916422, Test loss: 0.019607380032539368\n", + "Starting epoch 4744\n", + "Train loss: 0.011167353615164757, Test loss: 0.01897444576025009\n", + "Starting epoch 4745\n", + "Train loss: 0.011105505228042602, Test loss: 0.01028548926115036\n", + "Starting epoch 4746\n", + "Train loss: 0.011020343080163001, Test loss: 0.010474077425897121\n", + "Starting epoch 4747\n", + "Train loss: 0.011210430637001991, Test loss: 0.011060282588005066\n", + "Starting epoch 4748\n", + "Train loss: 0.011124442238360644, Test loss: 0.011802421882748604\n", + "Starting epoch 4749\n", + "Train loss: 0.011101960875093938, Test loss: 0.010437398217618465\n", + "Starting epoch 4750\n", + "Train loss: 0.011092823483049869, Test loss: 0.010775590315461159\n", + "Starting epoch 4751\n", + "Train loss: 0.011086754072457551, Test loss: 0.009618010371923447\n", + "Starting epoch 4752\n", + "Train loss: 0.011091059688478708, Test loss: 0.011358678340911865\n", + "Starting epoch 4753\n", + "Train loss: 0.011041427440941333, Test loss: 0.009854373522102833\n", + "Starting epoch 4754\n", + "Train loss: 0.01106504138559103, Test loss: 0.011501404456794262\n", + "Starting epoch 4755\n", + "Train loss: 0.011010106261819601, Test loss: 0.011440168134868145\n", + "Starting epoch 4756\n", + "Train loss: 0.011041187867522239, Test loss: 0.010995439253747463\n", + "Starting epoch 4757\n", + "Train loss: 0.01102346109226346, Test loss: 0.010612045414745808\n", + "Starting epoch 4758\n", + "Train loss: 0.011019506882876158, Test loss: 0.011929458938539028\n", + "Starting epoch 4759\n", + "Train loss: 0.01104608241468668, Test loss: 0.010525810532271862\n", + "Starting epoch 4760\n", + "Train loss: 0.010988269634544849, Test loss: 0.01196039654314518\n", + "Starting epoch 4761\n", + "Train loss: 0.011052627749741077, Test loss: 0.012061609886586666\n", + "Starting epoch 4762\n", + "Train loss: 0.01102582762017846, Test loss: 0.01163732260465622\n", + "Starting epoch 4763\n", + "Train loss: 0.011058241482824088, Test loss: 0.010106590576469898\n", + "Starting epoch 4764\n", + "Train loss: 0.01101275073364377, Test loss: 0.009756878949701786\n", + "Starting epoch 4765\n", + "Train loss: 0.01105624619871378, Test loss: 0.010805772617459297\n", + "Starting epoch 4766\n", + "Train loss: 0.011008690781891347, Test loss: 0.010867021977901459\n", + "Starting epoch 4767\n", + "Train loss: 0.010866182632744312, Test loss: 0.010568570345640182\n", + "Starting epoch 4768\n", + "Train loss: 0.011038017347455025, Test loss: 0.00986794475466013\n", + "Starting epoch 4769\n", + "Train loss: 0.011021670810878277, Test loss: 0.01027961540967226\n", + "Starting epoch 4770\n", + "Train loss: 0.01097784711048007, Test loss: 0.009435025043785572\n", + "Starting epoch 4771\n", + "Train loss: 0.011084529478102923, Test loss: 0.012196226045489311\n", + "Starting epoch 4772\n", + "Train loss: 0.011247405465692282, Test loss: 0.018971798941493034\n", + "Starting epoch 4773\n", + "Train loss: 0.011145707219839096, Test loss: 0.019642645493149757\n", + "Starting epoch 4774\n", + "Train loss: 0.011100951936095954, Test loss: 0.018518073484301567\n", + "Starting epoch 4775\n", + "Train loss: 0.01106103977188468, Test loss: 0.01642543449997902\n", + "Starting epoch 4776\n", + "Train loss: 0.010995579361915588, Test loss: 0.019110053777694702\n", + "Starting epoch 4777\n", + "Train loss: 0.01101245328783989, Test loss: 0.015298278070986271\n", + "Starting epoch 4778\n", + "Train loss: 0.01093408402055502, Test loss: 0.016367176547646523\n", + "Starting epoch 4779\n", + "Train loss: 0.010991424471139908, Test loss: 0.017264187335968018\n", + "Starting epoch 4780\n", + "Train loss: 0.010899598728865385, Test loss: 0.020289111882448196\n", + "Starting epoch 4781\n", + "Train loss: 0.011063688825815916, Test loss: 0.019106505438685417\n", + "Starting epoch 4782\n", + "Train loss: 0.011012885831296443, Test loss: 0.010870961472392082\n", + "Starting epoch 4783\n", + "Train loss: 0.011075201127678156, Test loss: 0.009732143953442574\n", + "Starting epoch 4784\n", + "Train loss: 0.01105020372197032, Test loss: 0.010082603432238102\n", + "Starting epoch 4785\n", + "Train loss: 0.010937809608876706, Test loss: 0.012050543911755085\n", + "Starting epoch 4786\n", + "Train loss: 0.010987789537757635, Test loss: 0.014134452678263187\n", + "Starting epoch 4787\n", + "Train loss: 0.011055818684399128, Test loss: 0.01063569076359272\n", + "Starting epoch 4788\n", + "Train loss: 0.011024851948022843, Test loss: 0.010214193724095821\n", + "Starting epoch 4789\n", + "Train loss: 0.011010763980448247, Test loss: 0.010599314235150814\n", + "Starting epoch 4790\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.010855319183319808, Test loss: 0.013456474989652634\n", + "Starting epoch 4791\n", + "Train loss: 0.011071961298584938, Test loss: 0.01889204978942871\n", + "Starting epoch 4792\n", + "Train loss: 0.011043932754546404, Test loss: 0.016609281301498413\n", + "Starting epoch 4793\n", + "Train loss: 0.01093702357262373, Test loss: 0.015872234478592873\n", + "Starting epoch 4794\n", + "Train loss: 0.011021107733249665, Test loss: 0.018268633633852005\n", + "Starting epoch 4795\n", + "Train loss: 0.010993605460971594, Test loss: 0.02044605277478695\n", + "Starting epoch 4796\n", + "Train loss: 0.010952798146754504, Test loss: 0.01824643835425377\n", + "Starting epoch 4797\n", + "Train loss: 0.01094065060839057, Test loss: 0.01607632264494896\n", + "Starting epoch 4798\n", + "Train loss: 0.010831632800400257, Test loss: 0.012388869188725948\n", + "Starting epoch 4799\n", + "Train loss: 0.010882694385945796, Test loss: 0.01852799579501152\n", + "Starting epoch 4800\n", + "Train loss: 0.01093070639297366, Test loss: 0.01639692485332489\n", + "Starting epoch 4801\n", + "Train loss: 0.010983990095555783, Test loss: 0.018082527443766594\n", + "Starting epoch 4802\n", + "Train loss: 0.010629936959594488, Test loss: 0.016188034787774086\n", + "Starting epoch 4803\n", + "Train loss: 0.010325907561928034, Test loss: 0.017072616145014763\n", + "Starting epoch 4804\n", + "Train loss: 0.01004792319610715, Test loss: 0.016944926232099533\n", + "Starting epoch 4805\n", + "Train loss: 0.009746061600744725, Test loss: 0.017016835510730743\n", + "Starting epoch 4806\n", + "Train loss: 0.009541791025549173, Test loss: 0.018093828111886978\n", + "Starting epoch 4807\n", + "Train loss: 0.009291969388723374, Test loss: 0.014853736385703087\n", + "Starting epoch 4808\n", + "Train loss: 0.009053609669208526, Test loss: 0.015262969769537449\n", + "Starting epoch 4809\n", + "Train loss: 0.008732924349606036, Test loss: 0.013564794324338436\n", + "Starting epoch 4810\n", + "Train loss: 0.008632190804928542, Test loss: 0.017133474349975586\n", + "Starting epoch 4811\n", + "Train loss: 0.00841658585704863, Test loss: 0.01480699609965086\n", + "Starting epoch 4812\n", + "Train loss: 0.008168088989332318, Test loss: 0.014911491423845291\n", + "Starting epoch 4813\n", + "Train loss: 0.008003834392875433, Test loss: 0.01592859998345375\n", + "Starting epoch 4814\n", + "Train loss: 0.007840873608365655, Test loss: 0.014569918625056744\n", + "Starting epoch 4815\n", + "Train loss: 0.007611669404432178, Test loss: 0.009985296986997128\n", + "Starting epoch 4816\n", + "Train loss: 0.007546777268871665, Test loss: 0.010187351144850254\n", + "Starting epoch 4817\n", + "Train loss: 0.007355133164674043, Test loss: 0.010480129159986973\n", + "Starting epoch 4818\n", + "Train loss: 0.0072776813711971046, Test loss: 0.010049836710095406\n", + "Starting epoch 4819\n", + "Train loss: 0.0071106002014130355, Test loss: 0.009402519091963768\n", + "Starting epoch 4820\n", + "Train loss: 0.0070042199920862915, Test loss: 0.009979739785194397\n", + "Starting epoch 4821\n", + "Train loss: 0.006837070593610406, Test loss: 0.01041752565652132\n", + "Starting epoch 4822\n", + "Train loss: 0.006761934440582991, Test loss: 0.010932459495961666\n", + "Starting epoch 4823\n", + "Train loss: 0.006674799406901002, Test loss: 0.013398060575127602\n", + "Starting epoch 4824\n", + "Train loss: 0.006591356843709945, Test loss: 0.012051018886268139\n", + "Starting epoch 4825\n", + "Train loss: 0.006517683323472738, Test loss: 0.010973180644214153\n", + "Starting epoch 4826\n", + "Train loss: 0.006441647754982114, Test loss: 0.01202926691621542\n", + "Starting epoch 4827\n", + "Train loss: 0.0064179361611604695, Test loss: 0.010684797540307045\n", + "Starting epoch 4828\n", + "Train loss: 0.006398076377809048, Test loss: 0.01037360168993473\n", + "Starting epoch 4829\n", + "Train loss: 0.006266996348276734, Test loss: 0.010217303410172462\n", + "Starting epoch 4830\n", + "Train loss: 0.006231273459270596, Test loss: 0.012845108285546303\n", + "Starting epoch 4831\n", + "Train loss: 0.0062115717865526675, Test loss: 0.01097707636654377\n", + "Starting epoch 4832\n", + "Train loss: 0.006154038654640317, Test loss: 0.01139618456363678\n", + "Starting epoch 4833\n", + "Train loss: 0.006136774281039834, Test loss: 0.011411957442760468\n", + "Starting epoch 4834\n", + "Train loss: 0.0060658676084131, Test loss: 0.011610784567892551\n", + "Starting epoch 4835\n", + "Train loss: 0.006073982957750559, Test loss: 0.01290134247392416\n", + "Starting epoch 4836\n", + "Train loss: 0.0060434587672352795, Test loss: 0.011298892088234425\n", + "Starting epoch 4837\n", + "Train loss: 0.0059551964234560726, Test loss: 0.011900185607373714\n", + "Starting epoch 4838\n", + "Train loss: 0.006020016269758344, Test loss: 0.01185629516839981\n", + "Starting epoch 4839\n", + "Train loss: 0.005937121044844389, Test loss: 0.010929269716143608\n", + "Starting epoch 4840\n", + "Train loss: 0.005925969369709492, Test loss: 0.010960000567138195\n", + "Starting epoch 4841\n", + "Train loss: 0.005882435385137797, Test loss: 0.011656849645078182\n", + "Starting epoch 4842\n", + "Train loss: 0.005909838369116187, Test loss: 0.011748834513127804\n", + "Starting epoch 4843\n", + "Train loss: 0.005916123744100332, Test loss: 0.011756714433431625\n", + "Starting epoch 4844\n", + "Train loss: 0.005880301967263222, Test loss: 0.011432289145886898\n", + "Starting epoch 4845\n", + "Train loss: 0.005849395142868161, Test loss: 0.010603220202028751\n", + "Starting epoch 4846\n", + "Train loss: 0.005840325439348817, Test loss: 0.011874445714056492\n", + "Starting epoch 4847\n", + "Train loss: 0.005872956877574324, Test loss: 0.011629600077867508\n", + "Starting epoch 4848\n", + "Train loss: 0.005834519760683179, Test loss: 0.01027233898639679\n", + "Starting epoch 4849\n", + "Train loss: 0.0058215472940355536, Test loss: 0.011491606011986732\n", + "Starting epoch 4850\n", + "Train loss: 0.005796104269102216, Test loss: 0.011114201508462429\n", + "Starting epoch 4851\n", + "Train loss: 0.005812344728037715, Test loss: 0.011457152664661407\n", + "Starting epoch 4852\n", + "Train loss: 0.005774286864325404, Test loss: 0.010803265497088432\n", + "Starting epoch 4853\n", + "Train loss: 0.005771110020577908, Test loss: 0.010983822867274284\n", + "Starting epoch 4854\n", + "Train loss: 0.005774932708591223, Test loss: 0.010674556717276573\n", + "Starting epoch 4855\n", + "Train loss: 0.005756314210593701, Test loss: 0.011407052166759968\n", + "Starting epoch 4856\n", + "Train loss: 0.005739161921665073, Test loss: 0.012471873313188553\n", + "Starting epoch 4857\n", + "Train loss: 0.005738161345943809, Test loss: 0.011065762490034103\n", + "Starting epoch 4858\n", + "Train loss: 0.005731256371363997, Test loss: 0.010696991346776485\n", + "Starting epoch 4859\n", + "Train loss: 0.0057215981092303995, Test loss: 0.010702046565711498\n", + "Starting epoch 4860\n", + "Train loss: 0.005717146592214704, Test loss: 0.010348280891776085\n", + "Starting epoch 4861\n", + "Train loss: 0.005696412278339267, Test loss: 0.010341242887079716\n", + "Starting epoch 4862\n", + "Train loss: 0.005699597122147679, Test loss: 0.0105077950283885\n", + "Starting epoch 4863\n", + "Train loss: 0.0056737281940877436, Test loss: 0.0106028001755476\n", + "Starting epoch 4864\n", + "Train loss: 0.005676120491698384, Test loss: 0.010792059823870659\n", + "Starting epoch 4865\n", + "Train loss: 0.005665965778753161, Test loss: 0.010081027634441853\n", + "Starting epoch 4866\n", + "Train loss: 0.005667938683182001, Test loss: 0.010581733658909798\n", + "Starting epoch 4867\n", + "Train loss: 0.005655496018007398, Test loss: 0.011094029992818832\n", + "Starting epoch 4868\n", + "Train loss: 0.00564336427487433, Test loss: 0.010775186121463776\n", + "Starting epoch 4869\n", + "Train loss: 0.005654685897752643, Test loss: 0.010978036560118198\n", + "Starting epoch 4870\n", + "Train loss: 0.005627121133729815, Test loss: 0.011472536250948906\n", + "Starting epoch 4871\n", + "Train loss: 0.005642647827044129, Test loss: 0.011087444610893726\n", + "Starting epoch 4872\n", + "Train loss: 0.00562272603623569, Test loss: 0.01086514350026846\n", + "Starting epoch 4873\n", + "Train loss: 0.005615093894302845, Test loss: 0.010212347842752934\n", + "Starting epoch 4874\n", + "Train loss: 0.005603255005553365, Test loss: 0.010306409560143948\n", + "Starting epoch 4875\n", + "Train loss: 0.005595422945916652, Test loss: 0.010921783745288849\n", + "Starting epoch 4876\n", + "Train loss: 0.005598830301314592, Test loss: 0.010849091224372387\n", + "Starting epoch 4877\n", + "Train loss: 0.005589636834338307, Test loss: 0.010609407909214497\n", + "Starting epoch 4878\n", + "Train loss: 0.005580998789519072, Test loss: 0.010012233629822731\n", + "Starting epoch 4879\n", + "Train loss: 0.005589303122833371, Test loss: 0.010711988434195518\n", + "Starting epoch 4880\n", + "Train loss: 0.005571669759228825, Test loss: 0.01046763826161623\n", + "Starting epoch 4881\n", + "Train loss: 0.005570782693102956, Test loss: 0.010172453708946705\n", + "Starting epoch 4882\n", + "Train loss: 0.005565641261637211, Test loss: 0.010611888021230698\n", + "Starting epoch 4883\n", + "Train loss: 0.0055597048997879026, Test loss: 0.010478549636900425\n", + "Starting epoch 4884\n", + "Train loss: 0.005546159418299794, Test loss: 0.01091237273067236\n", + "Starting epoch 4885\n", + "Train loss: 0.005550660798326135, Test loss: 0.01076491642743349\n", + "Starting epoch 4886\n", + "Train loss: 0.005540180979296565, Test loss: 0.010779723525047302\n", + "Starting epoch 4887\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.005530780972912907, Test loss: 0.010337239131331444\n", + "Starting epoch 4888\n", + "Train loss: 0.005531076490879059, Test loss: 0.010532006621360779\n", + "Starting epoch 4889\n", + "Train loss: 0.005529063362628222, Test loss: 0.010416942648589611\n", + "Starting epoch 4890\n", + "Train loss: 0.00551729685626924, Test loss: 0.011010256595909595\n", + "Starting epoch 4891\n", + "Train loss: 0.005518454862758517, Test loss: 0.010333499871194363\n", + "Starting epoch 4892\n", + "Train loss: 0.005517315249890089, Test loss: 0.010963116772472858\n", + "Starting epoch 4893\n", + "Train loss: 0.005503158383071423, Test loss: 0.010845910757780075\n", + "Starting epoch 4894\n", + "Train loss: 0.005500844717025757, Test loss: 0.010463018901646137\n", + "Starting epoch 4895\n", + "Train loss: 0.0054989133588969705, Test loss: 0.01029361691325903\n", + "Starting epoch 4896\n", + "Train loss: 0.0055012981779873375, Test loss: 0.010844498872756958\n", + "Starting epoch 4897\n", + "Train loss: 0.005491301054134965, Test loss: 0.01073505636304617\n", + "Starting epoch 4898\n", + "Train loss: 0.005483919205144048, Test loss: 0.010625282302498817\n", + "Starting epoch 4899\n", + "Train loss: 0.005476063247770071, Test loss: 0.0107562355697155\n", + "Starting epoch 4900\n", + "Train loss: 0.0054757459461689, Test loss: 0.01106924470514059\n", + "Starting epoch 4901\n", + "Train loss: 0.0054730873461812735, Test loss: 0.01052839495241642\n", + "Starting epoch 4902\n", + "Train loss: 0.005475671440362931, Test loss: 0.010895703919231892\n", + "Starting epoch 4903\n", + "Train loss: 0.005467313239350915, Test loss: 0.010424620471894741\n", + "Starting epoch 4904\n", + "Train loss: 0.005462872115895152, Test loss: 0.010291139595210552\n", + "Starting epoch 4905\n", + "Train loss: 0.005462222285568714, Test loss: 0.010431974194943905\n", + "Starting epoch 4906\n", + "Train loss: 0.005455615893006325, Test loss: 0.01049802079796791\n", + "Starting epoch 4907\n", + "Train loss: 0.005450124423950911, Test loss: 0.011024059727787971\n", + "Starting epoch 4908\n", + "Train loss: 0.005448705293238163, Test loss: 0.010527790524065495\n", + "Starting epoch 4909\n", + "Train loss: 0.005444733528420329, Test loss: 0.010935275815427303\n", + "Starting epoch 4910\n", + "Train loss: 0.00544103323481977, Test loss: 0.010126611217856407\n", + "Starting epoch 4911\n", + "Train loss: 0.00543952090665698, Test loss: 0.01013887394219637\n", + "Starting epoch 4912\n", + "Train loss: 0.005436023129150271, Test loss: 0.010261396877467632\n", + "Starting epoch 4913\n", + "Train loss: 0.005432120291516185, Test loss: 0.011080198921263218\n", + "Starting epoch 4914\n", + "Train loss: 0.005430754348635674, Test loss: 0.010740724392235279\n", + "Starting epoch 4915\n", + "Train loss: 0.005424748063087463, Test loss: 0.01116224192082882\n", + "Starting epoch 4916\n", + "Train loss: 0.005421430682763457, Test loss: 0.010753141716122627\n", + "Starting epoch 4917\n", + "Train loss: 0.0054190968535840515, Test loss: 0.010314512066543102\n", + "Starting epoch 4918\n", + "Train loss: 0.005420206477865576, Test loss: 0.010798243805766106\n", + "Starting epoch 4919\n", + "Train loss: 0.005412580613046884, Test loss: 0.01054325234144926\n", + "Starting epoch 4920\n", + "Train loss: 0.005412103701382875, Test loss: 0.01080111414194107\n", + "Starting epoch 4921\n", + "Train loss: 0.005408049505203963, Test loss: 0.010671384632587433\n", + "Starting epoch 4922\n", + "Train loss: 0.005404898775741458, Test loss: 0.010596890933811665\n", + "Starting epoch 4923\n", + "Train loss: 0.005399924162775278, Test loss: 0.010726659558713436\n", + "Starting epoch 4924\n", + "Train loss: 0.005395846525207162, Test loss: 0.010566896758973598\n", + "Starting epoch 4925\n", + "Train loss: 0.005397768160328269, Test loss: 0.01027072872966528\n", + "Starting epoch 4926\n", + "Train loss: 0.00539627319201827, Test loss: 0.010119826532900333\n", + "Starting epoch 4927\n", + "Train loss: 0.005394056960940361, Test loss: 0.01084162574261427\n", + "Starting epoch 4928\n", + "Train loss: 0.005388357629999519, Test loss: 0.01027104165405035\n", + "Starting epoch 4929\n", + "Train loss: 0.00538684175349772, Test loss: 0.010568777099251747\n", + "Starting epoch 4930\n", + "Train loss: 0.005383910685777664, Test loss: 0.010533872991800308\n", + "Starting epoch 4931\n", + "Train loss: 0.005379454893991351, Test loss: 0.010426877997815609\n", + "Starting epoch 4932\n", + "Train loss: 0.0053792398981750015, Test loss: 0.010414482094347477\n", + "Starting epoch 4933\n", + "Train loss: 0.005376246841624379, Test loss: 0.010442623868584633\n", + "Starting epoch 4934\n", + "Train loss: 0.005376306064426899, Test loss: 0.010468910448253155\n", + "Starting epoch 4935\n", + "Train loss: 0.005372341079637408, Test loss: 0.010391500778496265\n", + "Starting epoch 4936\n", + "Train loss: 0.005370226856321096, Test loss: 0.01081398781388998\n", + "Starting epoch 4937\n", + "Train loss: 0.005369216548278928, Test loss: 0.010458000004291534\n", + "Starting epoch 4938\n", + "Train loss: 0.0053686152305454015, Test loss: 0.01081135869026184\n", + "Starting epoch 4939\n", + "Train loss: 0.005363725041970611, Test loss: 0.010421941988170147\n", + "Starting epoch 4940\n", + "Train loss: 0.005364422835409641, Test loss: 0.010469271801412106\n", + "Starting epoch 4941\n", + "Train loss: 0.005359434150159359, Test loss: 0.010714690200984478\n", + "Starting epoch 4942\n", + "Train loss: 0.00535893764346838, Test loss: 0.010458264499902725\n", + "Starting epoch 4943\n", + "Train loss: 0.005357075082138181, Test loss: 0.010330396704375744\n", + "Starting epoch 4944\n", + "Train loss: 0.0053544534090906384, Test loss: 0.010568440891802311\n", + "Starting epoch 4945\n", + "Train loss: 0.005354270469397306, Test loss: 0.0107954703271389\n", + "Starting epoch 4946\n", + "Train loss: 0.005353619884699583, Test loss: 0.01078149676322937\n", + "Starting epoch 4947\n", + "Train loss: 0.005350228296592831, Test loss: 0.010495063848793507\n", + "Starting epoch 4948\n", + "Train loss: 0.005348162529990077, Test loss: 0.010594329796731472\n", + "Starting epoch 4949\n", + "Train loss: 0.005344534451141953, Test loss: 0.010747290216386318\n", + "Starting epoch 4950\n", + "Train loss: 0.005342968618497253, Test loss: 0.010708040557801723\n", + "Starting epoch 4951\n", + "Train loss: 0.005343422638252378, Test loss: 0.010572397150099277\n", + "Starting epoch 4952\n", + "Train loss: 0.005343777434900403, Test loss: 0.010393599979579449\n", + "Starting epoch 4953\n", + "Train loss: 0.005339768985286355, Test loss: 0.01050754263997078\n", + "Starting epoch 4954\n", + "Train loss: 0.0053376935143023725, Test loss: 0.010467072948813438\n", + "Starting epoch 4955\n", + "Train loss: 0.005336545715108514, Test loss: 0.010487574152648449\n", + "Starting epoch 4956\n", + "Train loss: 0.005334506593644619, Test loss: 0.010467235930263996\n", + "Starting epoch 4957\n", + "Train loss: 0.005334467692300678, Test loss: 0.010664188303053379\n", + "Starting epoch 4958\n", + "Train loss: 0.005332835204899311, Test loss: 0.010452860966324806\n", + "Starting epoch 4959\n", + "Train loss: 0.005330664655193686, Test loss: 0.010600280947983265\n", + "Starting epoch 4960\n", + "Train loss: 0.005330358790233731, Test loss: 0.010300944559276104\n", + "Starting epoch 4961\n", + "Train loss: 0.005328937098383903, Test loss: 0.010384603403508663\n", + "Starting epoch 4962\n", + "Train loss: 0.005325970705598593, Test loss: 0.010612331330776215\n", + "Starting epoch 4963\n", + "Train loss: 0.005325369630008936, Test loss: 0.010420041158795357\n", + "Starting epoch 4964\n", + "Train loss: 0.005323780067265034, Test loss: 0.010461384430527687\n", + "Starting epoch 4965\n", + "Train loss: 0.0053228932898491624, Test loss: 0.010404945351183414\n", + "Starting epoch 4966\n", + "Train loss: 0.005323521932587028, Test loss: 0.010251132771372795\n", + "Starting epoch 4967\n", + "Train loss: 0.005321688298135996, Test loss: 0.01050847303122282\n", + "Starting epoch 4968\n", + "Train loss: 0.005319724921137094, Test loss: 0.010546536184847355\n", + "Starting epoch 4969\n", + "Train loss: 0.005318057686090469, Test loss: 0.01040374394506216\n", + "Starting epoch 4970\n", + "Train loss: 0.005319320540875197, Test loss: 0.010568936355412006\n", + "Starting epoch 4971\n", + "Train loss: 0.005316520920023322, Test loss: 0.010518447495996952\n", + "Starting epoch 4972\n", + "Train loss: 0.005315682142972946, Test loss: 0.01071170438081026\n", + "Starting epoch 4973\n", + "Train loss: 0.005314224557951093, Test loss: 0.010273834690451622\n", + "Starting epoch 4974\n", + "Train loss: 0.00531303628347814, Test loss: 0.010280814953148365\n", + "Starting epoch 4975\n", + "Train loss: 0.005312586268410087, Test loss: 0.01055058091878891\n", + "Starting epoch 4976\n", + "Train loss: 0.005311036556959152, Test loss: 0.01044107973575592\n", + "Starting epoch 4977\n", + "Train loss: 0.005311138434335589, Test loss: 0.010450078174471855\n", + "Starting epoch 4978\n", + "Train loss: 0.005310363862663508, Test loss: 0.010529114864766598\n", + "Starting epoch 4979\n", + "Train loss: 0.005308455713093281, Test loss: 0.010348459705710411\n", + "Starting epoch 4980\n", + "Train loss: 0.005308244889602065, Test loss: 0.010278983041644096\n", + "Starting epoch 4981\n", + "Train loss: 0.005306980898603797, Test loss: 0.01028419379144907\n", + "Starting epoch 4982\n", + "Train loss: 0.005305837914347649, Test loss: 0.010467007756233215\n", + "Starting epoch 4983\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train loss: 0.005306165423244238, Test loss: 0.010504936799407005\n", + "Starting epoch 4984\n", + "Train loss: 0.005304618664085865, Test loss: 0.010680776089429855\n", + "Starting epoch 4985\n", + "Train loss: 0.005304562589153647, Test loss: 0.010692150332033634\n", + "Starting epoch 4986\n", + "Train loss: 0.005302222529426217, Test loss: 0.010440037585794926\n", + "Starting epoch 4987\n", + "Train loss: 0.0053019133489578965, Test loss: 0.01027199998497963\n", + "Starting epoch 4988\n", + "Train loss: 0.0053009532019495965, Test loss: 0.01019427552819252\n", + "Starting epoch 4989\n", + "Train loss: 0.005300992596894503, Test loss: 0.010575433261692524\n", + "Starting epoch 4990\n", + "Train loss: 0.005300719328224659, Test loss: 0.010436242446303368\n", + "Starting epoch 4991\n", + "Train loss: 0.0052988437749445436, Test loss: 0.010337332263588905\n", + "Starting epoch 4992\n", + "Train loss: 0.0052985707204788925, Test loss: 0.010436883196234703\n", + "Starting epoch 4993\n", + "Train loss: 0.005298254629597068, Test loss: 0.010424872860312462\n", + "Starting epoch 4994\n", + "Train loss: 0.005297139473259449, Test loss: 0.010505380108952522\n", + "Starting epoch 4995\n", + "Train loss: 0.005296492595225573, Test loss: 0.010471492074429989\n", + "Starting epoch 4996\n", + "Train loss: 0.005296541694551706, Test loss: 0.010349889285862446\n", + "Starting epoch 4997\n", + "Train loss: 0.00529556067660451, Test loss: 0.010400085709989071\n", + "Starting epoch 4998\n", + "Train loss: 0.005294842012226582, Test loss: 0.010290353558957577\n", + "Starting epoch 4999\n", + "Train loss: 0.0052946430444717404, Test loss: 0.010369248688220978\n", + "Starting epoch 5000\n", + "Train loss: 0.005293236412107945, Test loss: 0.010419058613479137\n" + ] + } + ], + "source": [ + "# first we train the network on the full space (once only).\n", + "\n", + "xgrid, ygrid, xsamp, ysamp, xtest, ytest, N, new_ranges = get_data(int(1e6), 0, sample_ranges)\n", + "\n", + "device = 'cuda:0'\n", + "\n", + "mlp = MLP().to(device)\n", + "\n", + "mlp = prepare_model(mlp,xsamp,np.log(ysamp),xgrid,np.log(ygrid)) # train" + ] + }, + { + "cell_type": "code", + "execution_count": 60, + "id": "4d14417c", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "6.215183519911282\n" + ] + } + ], + "source": [ + "xmeanstd = [xsamp.mean(),xsamp.std()]\n", + "ymeanstd = [np.log(ysamp).mean(),np.log(ysamp).std()]\n", + "\n", + "test_input = torch.Tensor(xtest)\n", + "normed_input = norm_inputs(test_input, ref_mean=xmeanstd[0], ref_std=xmeanstd[1]).float().to(device)\n", + "\n", + "n_batches = 10\n", + "\n", + "with torch.no_grad():\n", + " out = []\n", + " for i in range(n_batches):\n", + " output = mlp(normed_input[i * ytest.size // n_batches: (i + 1) * ytest.size // n_batches])\n", + " out.append(output.detach().cpu().numpy())\n", + "\n", + "\n", + "output = np.concatenate(out)\n", + "out_unnorm = unnorm(output, ref_mean=ymeanstd[0], ref_std=ymeanstd[1]).flatten()\n", + "out_unnorm = np.exp(out_unnorm)\n", + "\n", + "print(np.sqrt(np.mean((ytest - out_unnorm)**2)))" + ] + }, + { + "cell_type": "code", + "execution_count": 61, + "id": "4d3dfda1", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Power of two: 0\n", + "[[4.90308999 7.69897 ]\n", + " [4.69897 7.69897 ]\n", + " [0.01 0.99 ]\n", + " [0.01 0.5 ]\n", + " [0.5 0.99 ]\n", + " [0. 3.14159265]\n", + " [0. 6.28318531]\n", + " [0. 3.14159265]\n", + " [0.5 4. ]]\n", + "Linear + NN done\n", + "map_coords done\n", + "Power of two: 1\n", + "[[5.60205999 7. ]\n", + " [5.44897 6.94897 ]\n", + " [0.255 0.745 ]\n", + " [0.1325 0.3775 ]\n", + " [0.6225 0.8675 ]\n", + " [0.78539816 2.35619449]\n", + " [1.57079633 4.71238898]\n", + " [0.78539816 2.35619449]\n", + " [1.375 3.125 ]]\n" + ] + }, + { + "ename": "KeyboardInterrupt", + "evalue": "", + "output_type": "error", + "traceback": [ + "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", + "\u001b[0;31mKeyboardInterrupt\u001b[0m Traceback (most recent call last)", + "\u001b[0;32m/tmp/ipykernel_25522/1733873130.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m 17\u001b[0m \u001b[0mnn_interp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mRegularGridInterpolator\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mxdata_in\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_in\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmethod\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'nearest'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mbounds_error\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 18\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 19\u001b[0;31m \u001b[0mnew_values\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minterp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mxtest\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 20\u001b[0m \u001b[0mnew_values2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnn_interp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mxtest\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 21\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;32m~/anaconda3/envs/hi_torch/lib/python3.9/site-packages/scipy/interpolate/interpolate.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, xi, method)\u001b[0m\n\u001b[1;32m 2512\u001b[0m \u001b[0mindices\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnorm_distances\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mout_of_bounds\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_find_indices\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mxi\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mT\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 2513\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mmethod\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m\"linear\"\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2514\u001b[0;31m result = self._evaluate_linear(indices,\n\u001b[0m\u001b[1;32m 2515\u001b[0m \u001b[0mnorm_distances\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 2516\u001b[0m out_of_bounds)\n", + "\u001b[0;32m~/anaconda3/envs/hi_torch/lib/python3.9/site-packages/scipy/interpolate/interpolate.py\u001b[0m in \u001b[0;36m_evaluate_linear\u001b[0;34m(self, indices, norm_distances, out_of_bounds)\u001b[0m\n\u001b[1;32m 2536\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mei\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0myi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0medge_indices\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindices\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnorm_distances\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 2537\u001b[0m \u001b[0mweight\u001b[0m \u001b[0;34m*=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwhere\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mei\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0myi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0myi\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2538\u001b[0;31m \u001b[0mvalues\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0masarray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0medge_indices\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mweight\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mvslice\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 2539\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mvalues\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 2540\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;31mKeyboardInterrupt\u001b[0m: " + ] + } + ], + "source": [ + "for exponent in np.arange(0,6):\n", + " print('Power of two: ',exponent)\n", + " xgrid, ygrid, xsamp, ysamp, xtest, ytest, N, new_ranges = get_data(int(1e6), exponent, sample_ranges)\n", + "\n", + " np.save(full_path.format(f'{exponent}_xtest.npy'),arr=xtest)\n", + " np.save(full_path.format(f'{exponent}_ytest.npy'),arr=ytest)\n", + " \n", + " xdata_in = ()\n", + " for i in range(xgrid.shape[1]):\n", + " xdata_in += (np.unique(xgrid[:,i]),)\n", + " yresh = (N,)\n", + " for i in range(8):\n", + " yresh += (N,)\n", + " y_in = ygrid.reshape(yresh)\n", + "\n", + " interp = RegularGridInterpolator(xdata_in, y_in, method='linear',bounds_error=False)\n", + " nn_interp = RegularGridInterpolator(xdata_in, y_in, method='nearest',bounds_error=False)\n", + " \n", + " new_values = interp(xtest)\n", + " new_values2 = nn_interp(xtest)\n", + " \n", + " print('Linear + NN done')\n", + " \n", + " np.save(full_path.format(f'{exponent}_linear_out.npy'),arr=new_values)\n", + " np.save(full_path.format(f'{exponent}_nearest_out.npy'),arr=new_values2)\n", + " \n", + " xtest_in = []\n", + "\n", + " norms = [listy[1] - listy[0] for listy in new_ranges]\n", + "\n", + " for i in range(ytest.size):\n", + " sub = []\n", + " for j in range(9):\n", + " sub.append((N-1)*(xtest[i, j] - new_ranges[j][0])/norms[j])\n", + " xtest_in.append(sub)\n", + "\n", + " xtest_in = np.asarray(xtest_in).T\n", + "\n", + " map_values = map_coordinates(cp.array(y_in), cp.array(xtest_in), prefilter=True).get()\n", + " print('map_coords done')\n", + " np.save(full_path.format(f'{exponent}_map_out.npy'),arr=map_values)\n", + " \n", + " device = 'cuda:0'\n", + " \n", + " test_input = torch.Tensor(xtest)\n", + " normed_input = norm_inputs(test_input, ref_mean=xmeanstd[0],ref_std=xmeanstd[1]).float().to(device)\n", + "\n", + " with torch.no_grad():\n", + " out = []\n", + " for i in range(20):\n", + " output = mlp(normed_input[i * ytest.size // 20 : (i+1)*ytest.size // 20])\n", + " out.append(output.detach().cpu().numpy())\n", + "\n", + " output = np.concatenate(out)\n", + "\n", + " out_unnorm = np.exp(unnorm(output, ref_mean=ymeanstd[0],ref_std=ymeanstd[1]))\n", + " out_truth = out_unnorm.flatten()\n", + " \n", + " np.save(full_path.format(f'{exponent}_network_out.npy'),arr=out_truth)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "f4481771", + "metadata": {}, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3 (ipykernel)", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.9.6" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +} diff --git a/mock_data/mock_data_function_1/notebooks/halving_size_plot.ipynb b/mock_data/mock_data_function_1/notebooks/halving_size_plot.ipynb new file mode 100644 index 0000000000000000000000000000000000000000..36da18e523510865f56a7b0cf2eefc885971fae7 --- /dev/null +++ b/mock_data/mock_data_function_1/notebooks/halving_size_plot.ipynb @@ -0,0 +1,103 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": 8, + "id": "dd1f6d19", + "metadata": {}, + "outputs": [], + "source": [ + "import numpy as np\n", + "import matplotlib.pyplot as plt" + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "id": "abc2e918", + "metadata": { + "scrolled": false + }, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "<Figure size 1440x720 with 6 Axes>" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots(nrows=3, ncols=2, figsize=(20,10), facecolor=\"w\")\n", + "ax = ax.flatten()\n", + "\n", + "i = 0\n", + "for n in range(0,6):\n", + " ydata = np.load(f'./_halving_data/{n}_ytest.npy')\n", + " \n", + " lin = np.load(f'./_halving_data/{n}_linear_out.npy')\n", + " near = np.load(f'./_halving_data/{n}_nearest_out.npy')\n", + " mcoord = np.load(f'./_halving_data/{n}_map_out.npy')\n", + " network = np.load(f'./_halving_data/{n}_network_out.npy') \n", + "\n", + " # Absolute difference\n", + "# ax[i].hist(np.log10(abs(lin-ydata)), bins='auto', label='Linear Interp.',histtype='step',density=True, color=\"blue\")\n", + "# ax[i].hist(np.log10(abs(near-ydata)), bins='auto', label='Nearest Neighbour',histtype='step',density=True, color=\"orange\")\n", + "# ax[i].hist(np.log10(abs(mcoord-ydata)), bins='auto', label='n-Cubic Spine interp.',histtype='step',density=True, color=\"green\")\n", + "# ax[i].hist(np.log10(abs(network-ydata)), bins='auto',label='NN', histtype='step', density=True, color=\"red\")\n", + " \n", + " # Relative difference\n", + "# if n > 5:\n", + " ax[i].hist(np.log10(near/ydata), bins='auto', label='Nearest Neighbour',histtype='step',density=False, color='orange')\n", + "# if n > 2:\n", + "# temp = mcoord/ydata\n", + " ax[i].hist(np.log10(mcoord/ydata), bins='auto', label='n-Cubic Spine interp.',histtype='step',density=False, color='green')\n", + " ax[i].hist(np.log10(lin/ydata), bins='auto', label='Linear Interp.',histtype='step',density=False, color='blue')\n", + " \n", + " ax[i].hist(np.log10(network/ydata), bins='auto',label='NN', histtype='step', density=False, color='red')\n", + "\n", + " \n", + " ax[i].legend(loc='upper left')\n", + " ax[i].set_title(f'After {n} halvings')\n", + " i += 1\n", + "\n", + "fig.suptitle('Network vs. Interpolators as a function of prior range halvings - log10|diff|',fontsize=24)\n", + "#plt.xlabel('log10(|diff|) between truth and interpolated')\n", + "\n", + "plt.show()\n" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "0e249162", + "metadata": {}, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3 (ipykernel)", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.9.6" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +}