diff --git a/timm_tutorial.ipynb b/timm_tutorial.ipynb
new file mode 100644
index 0000000000000000000000000000000000000000..0097dbf3df53f1dca0ee35a8edfff2386adab54d
--- /dev/null
+++ b/timm_tutorial.ipynb
@@ -0,0 +1,1597 @@
+{
+ "cells": [
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "id": "34856f1a",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "['adv_inception_v3',\n",
+ " 'bat_resnext26ts',\n",
+ " 'beit_base_patch16_224',\n",
+ " 'beit_base_patch16_224_in22k',\n",
+ " 'beit_base_patch16_384',\n",
+ " 'beit_large_patch16_224',\n",
+ " 'beit_large_patch16_224_in22k',\n",
+ " 'beit_large_patch16_384',\n",
+ " 'beit_large_patch16_512',\n",
+ " 'botnet26t_256',\n",
+ " 'cait_m36_384',\n",
+ " 'cait_m48_448',\n",
+ " 'cait_s24_224',\n",
+ " 'cait_s24_384',\n",
+ " 'cait_s36_384',\n",
+ " 'cait_xs24_384',\n",
+ " 'cait_xxs24_224',\n",
+ " 'cait_xxs24_384',\n",
+ " 'cait_xxs36_224',\n",
+ " 'cait_xxs36_384',\n",
+ " 'coat_lite_mini',\n",
+ " 'coat_lite_small',\n",
+ " 'coat_lite_tiny',\n",
+ " 'coat_mini',\n",
+ " 'coat_tiny',\n",
+ " 'convit_base',\n",
+ " 'convit_small',\n",
+ " 'convit_tiny',\n",
+ " 'convmixer_768_32',\n",
+ " 'convmixer_1024_20_ks9_p14',\n",
+ " 'convmixer_1536_20',\n",
+ " 'convnext_base',\n",
+ " 'convnext_base_384_in22ft1k',\n",
+ " 'convnext_base_in22ft1k',\n",
+ " 'convnext_base_in22k',\n",
+ " 'convnext_large',\n",
+ " 'convnext_large_384_in22ft1k',\n",
+ " 'convnext_large_in22ft1k',\n",
+ " 'convnext_large_in22k',\n",
+ " 'convnext_small',\n",
+ " 'convnext_tiny',\n",
+ " 'convnext_xlarge_384_in22ft1k',\n",
+ " 'convnext_xlarge_in22ft1k',\n",
+ " 'convnext_xlarge_in22k',\n",
+ " 'crossvit_9_240',\n",
+ " 'crossvit_9_dagger_240',\n",
+ " 'crossvit_15_240',\n",
+ " 'crossvit_15_dagger_240',\n",
+ " 'crossvit_15_dagger_408',\n",
+ " 'crossvit_18_240',\n",
+ " 'crossvit_18_dagger_240',\n",
+ " 'crossvit_18_dagger_408',\n",
+ " 'crossvit_base_240',\n",
+ " 'crossvit_small_240',\n",
+ " 'crossvit_tiny_240',\n",
+ " 'cspdarknet53',\n",
+ " 'cspresnet50',\n",
+ " 'cspresnext50',\n",
+ " 'deit_base_distilled_patch16_224',\n",
+ " 'deit_base_distilled_patch16_384',\n",
+ " 'deit_base_patch16_224',\n",
+ " 'deit_base_patch16_384',\n",
+ " 'deit_small_distilled_patch16_224',\n",
+ " 'deit_small_patch16_224',\n",
+ " 'deit_tiny_distilled_patch16_224',\n",
+ " 'deit_tiny_patch16_224',\n",
+ " 'densenet121',\n",
+ " 'densenet161',\n",
+ " 'densenet169',\n",
+ " 'densenet201',\n",
+ " 'densenetblur121d',\n",
+ " 'dla34',\n",
+ " 'dla46_c',\n",
+ " 'dla46x_c',\n",
+ " 'dla60',\n",
+ " 'dla60_res2net',\n",
+ " 'dla60_res2next',\n",
+ " 'dla60x',\n",
+ " 'dla60x_c',\n",
+ " 'dla102',\n",
+ " 'dla102x',\n",
+ " 'dla102x2',\n",
+ " 'dla169',\n",
+ " 'dm_nfnet_f0',\n",
+ " 'dm_nfnet_f1',\n",
+ " 'dm_nfnet_f2',\n",
+ " 'dm_nfnet_f3',\n",
+ " 'dm_nfnet_f4',\n",
+ " 'dm_nfnet_f5',\n",
+ " 'dm_nfnet_f6',\n",
+ " 'dpn68',\n",
+ " 'dpn68b',\n",
+ " 'dpn92',\n",
+ " 'dpn98',\n",
+ " 'dpn107',\n",
+ " 'dpn131',\n",
+ " 'eca_botnext26ts_256',\n",
+ " 'eca_halonext26ts',\n",
+ " 'eca_nfnet_l0',\n",
+ " 'eca_nfnet_l1',\n",
+ " 'eca_nfnet_l2',\n",
+ " 'eca_resnet33ts',\n",
+ " 'eca_resnext26ts',\n",
+ " 'ecaresnet26t',\n",
+ " 'ecaresnet50d',\n",
+ " 'ecaresnet50d_pruned',\n",
+ " 'ecaresnet50t',\n",
+ " 'ecaresnet101d',\n",
+ " 'ecaresnet101d_pruned',\n",
+ " 'ecaresnet269d',\n",
+ " 'ecaresnetlight',\n",
+ " 'efficientnet_b0',\n",
+ " 'efficientnet_b1',\n",
+ " 'efficientnet_b1_pruned',\n",
+ " 'efficientnet_b2',\n",
+ " 'efficientnet_b2_pruned',\n",
+ " 'efficientnet_b3',\n",
+ " 'efficientnet_b3_pruned',\n",
+ " 'efficientnet_b4',\n",
+ " 'efficientnet_el',\n",
+ " 'efficientnet_el_pruned',\n",
+ " 'efficientnet_em',\n",
+ " 'efficientnet_es',\n",
+ " 'efficientnet_es_pruned',\n",
+ " 'efficientnet_lite0',\n",
+ " 'efficientnetv2_rw_m',\n",
+ " 'efficientnetv2_rw_s',\n",
+ " 'efficientnetv2_rw_t',\n",
+ " 'ens_adv_inception_resnet_v2',\n",
+ " 'ese_vovnet19b_dw',\n",
+ " 'ese_vovnet39b',\n",
+ " 'fbnetc_100',\n",
+ " 'fbnetv3_b',\n",
+ " 'fbnetv3_d',\n",
+ " 'fbnetv3_g',\n",
+ " 'gc_efficientnetv2_rw_t',\n",
+ " 'gcresnet33ts',\n",
+ " 'gcresnet50t',\n",
+ " 'gcresnext26ts',\n",
+ " 'gcresnext50ts',\n",
+ " 'gernet_l',\n",
+ " 'gernet_m',\n",
+ " 'gernet_s',\n",
+ " 'ghostnet_100',\n",
+ " 'gluon_inception_v3',\n",
+ " 'gluon_resnet18_v1b',\n",
+ " 'gluon_resnet34_v1b',\n",
+ " 'gluon_resnet50_v1b',\n",
+ " 'gluon_resnet50_v1c',\n",
+ " 'gluon_resnet50_v1d',\n",
+ " 'gluon_resnet50_v1s',\n",
+ " 'gluon_resnet101_v1b',\n",
+ " 'gluon_resnet101_v1c',\n",
+ " 'gluon_resnet101_v1d',\n",
+ " 'gluon_resnet101_v1s',\n",
+ " 'gluon_resnet152_v1b',\n",
+ " 'gluon_resnet152_v1c',\n",
+ " 'gluon_resnet152_v1d',\n",
+ " 'gluon_resnet152_v1s',\n",
+ " 'gluon_resnext50_32x4d',\n",
+ " 'gluon_resnext101_32x4d',\n",
+ " 'gluon_resnext101_64x4d',\n",
+ " 'gluon_senet154',\n",
+ " 'gluon_seresnext50_32x4d',\n",
+ " 'gluon_seresnext101_32x4d',\n",
+ " 'gluon_seresnext101_64x4d',\n",
+ " 'gluon_xception65',\n",
+ " 'gmixer_24_224',\n",
+ " 'gmlp_s16_224',\n",
+ " 'halo2botnet50ts_256',\n",
+ " 'halonet26t',\n",
+ " 'halonet50ts',\n",
+ " 'haloregnetz_b',\n",
+ " 'hardcorenas_a',\n",
+ " 'hardcorenas_b',\n",
+ " 'hardcorenas_c',\n",
+ " 'hardcorenas_d',\n",
+ " 'hardcorenas_e',\n",
+ " 'hardcorenas_f',\n",
+ " 'hrnet_w18',\n",
+ " 'hrnet_w18_small',\n",
+ " 'hrnet_w18_small_v2',\n",
+ " 'hrnet_w30',\n",
+ " 'hrnet_w32',\n",
+ " 'hrnet_w40',\n",
+ " 'hrnet_w44',\n",
+ " 'hrnet_w48',\n",
+ " 'hrnet_w64',\n",
+ " 'ig_resnext101_32x8d',\n",
+ " 'ig_resnext101_32x16d',\n",
+ " 'ig_resnext101_32x32d',\n",
+ " 'ig_resnext101_32x48d',\n",
+ " 'inception_resnet_v2',\n",
+ " 'inception_v3',\n",
+ " 'inception_v4',\n",
+ " 'jx_nest_base',\n",
+ " 'jx_nest_small',\n",
+ " 'jx_nest_tiny',\n",
+ " 'lambda_resnet26rpt_256',\n",
+ " 'lambda_resnet26t',\n",
+ " 'lambda_resnet50ts',\n",
+ " 'lamhalobotnet50ts_256',\n",
+ " 'lcnet_050',\n",
+ " 'lcnet_075',\n",
+ " 'lcnet_100',\n",
+ " 'legacy_senet154',\n",
+ " 'legacy_seresnet18',\n",
+ " 'legacy_seresnet34',\n",
+ " 'legacy_seresnet50',\n",
+ " 'legacy_seresnet101',\n",
+ " 'legacy_seresnet152',\n",
+ " 'legacy_seresnext26_32x4d',\n",
+ " 'legacy_seresnext50_32x4d',\n",
+ " 'legacy_seresnext101_32x4d',\n",
+ " 'levit_128',\n",
+ " 'levit_128s',\n",
+ " 'levit_192',\n",
+ " 'levit_256',\n",
+ " 'levit_384',\n",
+ " 'mixer_b16_224',\n",
+ " 'mixer_b16_224_in21k',\n",
+ " 'mixer_b16_224_miil',\n",
+ " 'mixer_b16_224_miil_in21k',\n",
+ " 'mixer_l16_224',\n",
+ " 'mixer_l16_224_in21k',\n",
+ " 'mixnet_l',\n",
+ " 'mixnet_m',\n",
+ " 'mixnet_s',\n",
+ " 'mixnet_xl',\n",
+ " 'mnasnet_100',\n",
+ " 'mnasnet_small',\n",
+ " 'mobilenetv2_050',\n",
+ " 'mobilenetv2_100',\n",
+ " 'mobilenetv2_110d',\n",
+ " 'mobilenetv2_120d',\n",
+ " 'mobilenetv2_140',\n",
+ " 'mobilenetv3_large_100',\n",
+ " 'mobilenetv3_large_100_miil',\n",
+ " 'mobilenetv3_large_100_miil_in21k',\n",
+ " 'mobilenetv3_rw',\n",
+ " 'nasnetalarge',\n",
+ " 'nf_regnet_b1',\n",
+ " 'nf_resnet50',\n",
+ " 'nfnet_l0',\n",
+ " 'pit_b_224',\n",
+ " 'pit_b_distilled_224',\n",
+ " 'pit_s_224',\n",
+ " 'pit_s_distilled_224',\n",
+ " 'pit_ti_224',\n",
+ " 'pit_ti_distilled_224',\n",
+ " 'pit_xs_224',\n",
+ " 'pit_xs_distilled_224',\n",
+ " 'pnasnet5large',\n",
+ " 'regnetx_002',\n",
+ " 'regnetx_004',\n",
+ " 'regnetx_006',\n",
+ " 'regnetx_008',\n",
+ " 'regnetx_016',\n",
+ " 'regnetx_032',\n",
+ " 'regnetx_040',\n",
+ " 'regnetx_064',\n",
+ " 'regnetx_080',\n",
+ " 'regnetx_120',\n",
+ " 'regnetx_160',\n",
+ " 'regnetx_320',\n",
+ " 'regnety_002',\n",
+ " 'regnety_004',\n",
+ " 'regnety_006',\n",
+ " 'regnety_008',\n",
+ " 'regnety_016',\n",
+ " 'regnety_032',\n",
+ " 'regnety_040',\n",
+ " 'regnety_064',\n",
+ " 'regnety_080',\n",
+ " 'regnety_120',\n",
+ " 'regnety_160',\n",
+ " 'regnety_320',\n",
+ " 'regnetz_b16',\n",
+ " 'regnetz_c16',\n",
+ " 'regnetz_d8',\n",
+ " 'regnetz_d32',\n",
+ " 'regnetz_e8',\n",
+ " 'repvgg_a2',\n",
+ " 'repvgg_b0',\n",
+ " 'repvgg_b1',\n",
+ " 'repvgg_b1g4',\n",
+ " 'repvgg_b2',\n",
+ " 'repvgg_b2g4',\n",
+ " 'repvgg_b3',\n",
+ " 'repvgg_b3g4',\n",
+ " 'res2net50_14w_8s',\n",
+ " 'res2net50_26w_4s',\n",
+ " 'res2net50_26w_6s',\n",
+ " 'res2net50_26w_8s',\n",
+ " 'res2net50_48w_2s',\n",
+ " 'res2net101_26w_4s',\n",
+ " 'res2next50',\n",
+ " 'resmlp_12_224',\n",
+ " 'resmlp_12_224_dino',\n",
+ " 'resmlp_12_distilled_224',\n",
+ " 'resmlp_24_224',\n",
+ " 'resmlp_24_224_dino',\n",
+ " 'resmlp_24_distilled_224',\n",
+ " 'resmlp_36_224',\n",
+ " 'resmlp_36_distilled_224',\n",
+ " 'resmlp_big_24_224',\n",
+ " 'resmlp_big_24_224_in22ft1k',\n",
+ " 'resmlp_big_24_distilled_224',\n",
+ " 'resnest14d',\n",
+ " 'resnest26d',\n",
+ " 'resnest50d',\n",
+ " 'resnest50d_1s4x24d',\n",
+ " 'resnest50d_4s2x40d',\n",
+ " 'resnest101e',\n",
+ " 'resnest200e',\n",
+ " 'resnest269e',\n",
+ " 'resnet18',\n",
+ " 'resnet18d',\n",
+ " 'resnet26',\n",
+ " 'resnet26d',\n",
+ " 'resnet26t',\n",
+ " 'resnet32ts',\n",
+ " 'resnet33ts',\n",
+ " 'resnet34',\n",
+ " 'resnet34d',\n",
+ " 'resnet50',\n",
+ " 'resnet50_gn',\n",
+ " 'resnet50d',\n",
+ " 'resnet51q',\n",
+ " 'resnet61q',\n",
+ " 'resnet101',\n",
+ " 'resnet101d',\n",
+ " 'resnet152',\n",
+ " 'resnet152d',\n",
+ " 'resnet200d',\n",
+ " 'resnetblur50',\n",
+ " 'resnetrs50',\n",
+ " 'resnetrs101',\n",
+ " 'resnetrs152',\n",
+ " 'resnetrs200',\n",
+ " 'resnetrs270',\n",
+ " 'resnetrs350',\n",
+ " 'resnetrs420',\n",
+ " 'resnetv2_50',\n",
+ " 'resnetv2_50x1_bit_distilled',\n",
+ " 'resnetv2_50x1_bitm',\n",
+ " 'resnetv2_50x1_bitm_in21k',\n",
+ " 'resnetv2_50x3_bitm',\n",
+ " 'resnetv2_50x3_bitm_in21k',\n",
+ " 'resnetv2_101',\n",
+ " 'resnetv2_101x1_bitm',\n",
+ " 'resnetv2_101x1_bitm_in21k',\n",
+ " 'resnetv2_101x3_bitm',\n",
+ " 'resnetv2_101x3_bitm_in21k',\n",
+ " 'resnetv2_152x2_bit_teacher',\n",
+ " 'resnetv2_152x2_bit_teacher_384',\n",
+ " 'resnetv2_152x2_bitm',\n",
+ " 'resnetv2_152x2_bitm_in21k',\n",
+ " 'resnetv2_152x4_bitm',\n",
+ " 'resnetv2_152x4_bitm_in21k',\n",
+ " 'resnext26ts',\n",
+ " 'resnext50_32x4d',\n",
+ " 'resnext50d_32x4d',\n",
+ " 'resnext101_32x8d',\n",
+ " 'rexnet_100',\n",
+ " 'rexnet_130',\n",
+ " 'rexnet_150',\n",
+ " 'rexnet_200',\n",
+ " 'sebotnet33ts_256',\n",
+ " 'sehalonet33ts',\n",
+ " 'selecsls42b',\n",
+ " 'selecsls60',\n",
+ " 'selecsls60b',\n",
+ " 'semnasnet_075',\n",
+ " 'semnasnet_100',\n",
+ " 'seresnet33ts',\n",
+ " 'seresnet50',\n",
+ " 'seresnet152d',\n",
+ " 'seresnext26d_32x4d',\n",
+ " 'seresnext26t_32x4d',\n",
+ " 'seresnext26ts',\n",
+ " 'seresnext50_32x4d',\n",
+ " 'skresnet18',\n",
+ " 'skresnet34',\n",
+ " 'skresnext50_32x4d',\n",
+ " 'spnasnet_100',\n",
+ " 'ssl_resnet18',\n",
+ " 'ssl_resnet50',\n",
+ " 'ssl_resnext50_32x4d',\n",
+ " 'ssl_resnext101_32x4d',\n",
+ " 'ssl_resnext101_32x8d',\n",
+ " 'ssl_resnext101_32x16d',\n",
+ " 'swin_base_patch4_window7_224',\n",
+ " 'swin_base_patch4_window7_224_in22k',\n",
+ " 'swin_base_patch4_window12_384',\n",
+ " 'swin_base_patch4_window12_384_in22k',\n",
+ " 'swin_large_patch4_window7_224',\n",
+ " 'swin_large_patch4_window7_224_in22k',\n",
+ " 'swin_large_patch4_window12_384',\n",
+ " 'swin_large_patch4_window12_384_in22k',\n",
+ " 'swin_small_patch4_window7_224',\n",
+ " 'swin_tiny_patch4_window7_224',\n",
+ " 'swsl_resnet18',\n",
+ " 'swsl_resnet50',\n",
+ " 'swsl_resnext50_32x4d',\n",
+ " 'swsl_resnext101_32x4d',\n",
+ " 'swsl_resnext101_32x8d',\n",
+ " 'swsl_resnext101_32x16d',\n",
+ " 'tf_efficientnet_b0',\n",
+ " 'tf_efficientnet_b0_ap',\n",
+ " 'tf_efficientnet_b0_ns',\n",
+ " 'tf_efficientnet_b1',\n",
+ " 'tf_efficientnet_b1_ap',\n",
+ " 'tf_efficientnet_b1_ns',\n",
+ " 'tf_efficientnet_b2',\n",
+ " 'tf_efficientnet_b2_ap',\n",
+ " 'tf_efficientnet_b2_ns',\n",
+ " 'tf_efficientnet_b3',\n",
+ " 'tf_efficientnet_b3_ap',\n",
+ " 'tf_efficientnet_b3_ns',\n",
+ " 'tf_efficientnet_b4',\n",
+ " 'tf_efficientnet_b4_ap',\n",
+ " 'tf_efficientnet_b4_ns',\n",
+ " 'tf_efficientnet_b5',\n",
+ " 'tf_efficientnet_b5_ap',\n",
+ " 'tf_efficientnet_b5_ns',\n",
+ " 'tf_efficientnet_b6',\n",
+ " 'tf_efficientnet_b6_ap',\n",
+ " 'tf_efficientnet_b6_ns',\n",
+ " 'tf_efficientnet_b7',\n",
+ " 'tf_efficientnet_b7_ap',\n",
+ " 'tf_efficientnet_b7_ns',\n",
+ " 'tf_efficientnet_b8',\n",
+ " 'tf_efficientnet_b8_ap',\n",
+ " 'tf_efficientnet_cc_b0_4e',\n",
+ " 'tf_efficientnet_cc_b0_8e',\n",
+ " 'tf_efficientnet_cc_b1_8e',\n",
+ " 'tf_efficientnet_el',\n",
+ " 'tf_efficientnet_em',\n",
+ " 'tf_efficientnet_es',\n",
+ " 'tf_efficientnet_l2_ns',\n",
+ " 'tf_efficientnet_l2_ns_475',\n",
+ " 'tf_efficientnet_lite0',\n",
+ " 'tf_efficientnet_lite1',\n",
+ " 'tf_efficientnet_lite2',\n",
+ " 'tf_efficientnet_lite3',\n",
+ " 'tf_efficientnet_lite4',\n",
+ " 'tf_efficientnetv2_b0',\n",
+ " 'tf_efficientnetv2_b1',\n",
+ " 'tf_efficientnetv2_b2',\n",
+ " 'tf_efficientnetv2_b3',\n",
+ " 'tf_efficientnetv2_l',\n",
+ " 'tf_efficientnetv2_l_in21ft1k',\n",
+ " 'tf_efficientnetv2_l_in21k',\n",
+ " 'tf_efficientnetv2_m',\n",
+ " 'tf_efficientnetv2_m_in21ft1k',\n",
+ " 'tf_efficientnetv2_m_in21k',\n",
+ " 'tf_efficientnetv2_s',\n",
+ " 'tf_efficientnetv2_s_in21ft1k',\n",
+ " 'tf_efficientnetv2_s_in21k',\n",
+ " 'tf_efficientnetv2_xl_in21ft1k',\n",
+ " 'tf_efficientnetv2_xl_in21k',\n",
+ " 'tf_inception_v3',\n",
+ " 'tf_mixnet_l',\n",
+ " 'tf_mixnet_m',\n",
+ " 'tf_mixnet_s',\n",
+ " 'tf_mobilenetv3_large_075',\n",
+ " 'tf_mobilenetv3_large_100',\n",
+ " 'tf_mobilenetv3_large_minimal_100',\n",
+ " 'tf_mobilenetv3_small_075',\n",
+ " 'tf_mobilenetv3_small_100',\n",
+ " 'tf_mobilenetv3_small_minimal_100',\n",
+ " 'tinynet_a',\n",
+ " 'tinynet_b',\n",
+ " 'tinynet_c',\n",
+ " 'tinynet_d',\n",
+ " 'tinynet_e',\n",
+ " 'tnt_s_patch16_224',\n",
+ " 'tresnet_l',\n",
+ " 'tresnet_l_448',\n",
+ " 'tresnet_m',\n",
+ " 'tresnet_m_448',\n",
+ " 'tresnet_m_miil_in21k',\n",
+ " 'tresnet_xl',\n",
+ " 'tresnet_xl_448',\n",
+ " 'tv_densenet121',\n",
+ " 'tv_resnet34',\n",
+ " 'tv_resnet50',\n",
+ " 'tv_resnet101',\n",
+ " 'tv_resnet152',\n",
+ " 'tv_resnext50_32x4d',\n",
+ " 'twins_pcpvt_base',\n",
+ " 'twins_pcpvt_large',\n",
+ " 'twins_pcpvt_small',\n",
+ " 'twins_svt_base',\n",
+ " 'twins_svt_large',\n",
+ " 'twins_svt_small',\n",
+ " 'vgg11',\n",
+ " 'vgg11_bn',\n",
+ " 'vgg13',\n",
+ " 'vgg13_bn',\n",
+ " 'vgg16',\n",
+ " 'vgg16_bn',\n",
+ " 'vgg19',\n",
+ " 'vgg19_bn',\n",
+ " 'visformer_small',\n",
+ " 'vit_base_patch8_224',\n",
+ " 'vit_base_patch8_224_in21k',\n",
+ " 'vit_base_patch16_224',\n",
+ " 'vit_base_patch16_224_in21k',\n",
+ " 'vit_base_patch16_224_miil',\n",
+ " 'vit_base_patch16_224_miil_in21k',\n",
+ " 'vit_base_patch16_384',\n",
+ " 'vit_base_patch16_sam_224',\n",
+ " 'vit_base_patch32_224',\n",
+ " 'vit_base_patch32_224_in21k',\n",
+ " 'vit_base_patch32_384',\n",
+ " 'vit_base_patch32_sam_224',\n",
+ " 'vit_base_r50_s16_224_in21k',\n",
+ " 'vit_base_r50_s16_384',\n",
+ " 'vit_huge_patch14_224_in21k',\n",
+ " 'vit_large_patch16_224',\n",
+ " 'vit_large_patch16_224_in21k',\n",
+ " 'vit_large_patch16_384',\n",
+ " 'vit_large_patch32_224_in21k',\n",
+ " 'vit_large_patch32_384',\n",
+ " 'vit_large_r50_s32_224',\n",
+ " 'vit_large_r50_s32_224_in21k',\n",
+ " 'vit_large_r50_s32_384',\n",
+ " 'vit_small_patch16_224',\n",
+ " 'vit_small_patch16_224_in21k',\n",
+ " 'vit_small_patch16_384',\n",
+ " 'vit_small_patch32_224',\n",
+ " 'vit_small_patch32_224_in21k',\n",
+ " 'vit_small_patch32_384',\n",
+ " 'vit_small_r26_s32_224',\n",
+ " 'vit_small_r26_s32_224_in21k',\n",
+ " 'vit_small_r26_s32_384',\n",
+ " 'vit_tiny_patch16_224',\n",
+ " 'vit_tiny_patch16_224_in21k',\n",
+ " 'vit_tiny_patch16_384',\n",
+ " 'vit_tiny_r_s16_p8_224',\n",
+ " 'vit_tiny_r_s16_p8_224_in21k',\n",
+ " 'vit_tiny_r_s16_p8_384',\n",
+ " 'wide_resnet50_2',\n",
+ " 'wide_resnet101_2',\n",
+ " 'xception',\n",
+ " 'xception41',\n",
+ " 'xception65',\n",
+ " 'xception71',\n",
+ " 'xcit_large_24_p8_224',\n",
+ " 'xcit_large_24_p8_224_dist',\n",
+ " 'xcit_large_24_p8_384_dist',\n",
+ " 'xcit_large_24_p16_224',\n",
+ " 'xcit_large_24_p16_224_dist',\n",
+ " 'xcit_large_24_p16_384_dist',\n",
+ " 'xcit_medium_24_p8_224',\n",
+ " 'xcit_medium_24_p8_224_dist',\n",
+ " 'xcit_medium_24_p8_384_dist',\n",
+ " 'xcit_medium_24_p16_224',\n",
+ " 'xcit_medium_24_p16_224_dist',\n",
+ " 'xcit_medium_24_p16_384_dist',\n",
+ " 'xcit_nano_12_p8_224',\n",
+ " 'xcit_nano_12_p8_224_dist',\n",
+ " 'xcit_nano_12_p8_384_dist',\n",
+ " 'xcit_nano_12_p16_224',\n",
+ " 'xcit_nano_12_p16_224_dist',\n",
+ " 'xcit_nano_12_p16_384_dist',\n",
+ " 'xcit_small_12_p8_224',\n",
+ " 'xcit_small_12_p8_224_dist',\n",
+ " 'xcit_small_12_p8_384_dist',\n",
+ " 'xcit_small_12_p16_224',\n",
+ " 'xcit_small_12_p16_224_dist',\n",
+ " 'xcit_small_12_p16_384_dist',\n",
+ " 'xcit_small_24_p8_224',\n",
+ " 'xcit_small_24_p8_224_dist',\n",
+ " 'xcit_small_24_p8_384_dist',\n",
+ " 'xcit_small_24_p16_224',\n",
+ " 'xcit_small_24_p16_224_dist',\n",
+ " 'xcit_small_24_p16_384_dist',\n",
+ " 'xcit_tiny_12_p8_224',\n",
+ " 'xcit_tiny_12_p8_224_dist',\n",
+ " 'xcit_tiny_12_p8_384_dist',\n",
+ " 'xcit_tiny_12_p16_224',\n",
+ " 'xcit_tiny_12_p16_224_dist',\n",
+ " 'xcit_tiny_12_p16_384_dist',\n",
+ " 'xcit_tiny_24_p8_224',\n",
+ " 'xcit_tiny_24_p8_224_dist',\n",
+ " 'xcit_tiny_24_p8_384_dist',\n",
+ " 'xcit_tiny_24_p16_224',\n",
+ " 'xcit_tiny_24_p16_224_dist',\n",
+ " 'xcit_tiny_24_p16_384_dist']\n"
+ ]
+ }
+ ],
+ "source": [
+ "import timm\n",
+ "from pprint import pprint\n",
+ "model_names = timm.list_models(pretrained=True)\n",
+ "pprint(model_names)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "id": "158eada7",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Files already downloaded and verified\n",
+ "--------------- 0 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (6.7878) | Acc: (17.19%) (22/128)\n",
+ "Batch_idx: 10 | Loss: (6.1913) | Acc: (15.70%) (221/1408)\n",
+ "Batch_idx: 20 | Loss: (4.9707) | Acc: (16.33%) (439/2688)\n",
+ "Batch_idx: 30 | Loss: (4.3882) | Acc: (15.90%) (631/3968)\n",
+ "Batch_idx: 40 | Loss: (3.9776) | Acc: (15.70%) (824/5248)\n",
+ "Batch_idx: 50 | Loss: (3.6887) | Acc: (15.43%) (1007/6528)\n",
+ "Batch_idx: 60 | Loss: (3.4578) | Acc: (15.97%) (1247/7808)\n",
+ "Batch_idx: 70 | Loss: (3.2882) | Acc: (16.52%) (1501/9088)\n",
+ "Batch_idx: 80 | Loss: (3.1544) | Acc: (17.28%) (1792/10368)\n",
+ "Batch_idx: 90 | Loss: (3.0407) | Acc: (18.11%) (2110/11648)\n",
+ "Batch_idx: 100 | Loss: (2.9361) | Acc: (19.10%) (2469/12928)\n",
+ "Batch_idx: 110 | Loss: (2.8507) | Acc: (19.82%) (2816/14208)\n",
+ "Batch_idx: 120 | Loss: (2.7830) | Acc: (20.52%) (3178/15488)\n",
+ "Batch_idx: 130 | Loss: (2.7165) | Acc: (21.17%) (3550/16768)\n",
+ "Batch_idx: 140 | Loss: (2.6645) | Acc: (21.78%) (3930/18048)\n",
+ "Batch_idx: 150 | Loss: (2.6163) | Acc: (22.43%) (4336/19328)\n",
+ "Batch_idx: 160 | Loss: (2.5705) | Acc: (23.13%) (4766/20608)\n",
+ "Batch_idx: 170 | Loss: (2.5279) | Acc: (23.72%) (5192/21888)\n",
+ "Batch_idx: 180 | Loss: (2.4909) | Acc: (24.38%) (5648/23168)\n",
+ "Batch_idx: 190 | Loss: (2.4536) | Acc: (25.01%) (6114/24448)\n",
+ "Batch_idx: 200 | Loss: (2.4250) | Acc: (25.47%) (6552/25728)\n",
+ "Batch_idx: 210 | Loss: (2.3927) | Acc: (26.07%) (7040/27008)\n",
+ "Batch_idx: 220 | Loss: (2.3636) | Acc: (26.58%) (7520/28288)\n",
+ "Batch_idx: 230 | Loss: (2.3348) | Acc: (27.21%) (8045/29568)\n",
+ "Batch_idx: 240 | Loss: (2.3102) | Acc: (27.69%) (8542/30848)\n",
+ "Batch_idx: 250 | Loss: (2.2851) | Acc: (28.17%) (9052/32128)\n",
+ "Batch_idx: 260 | Loss: (2.2638) | Acc: (28.60%) (9555/33408)\n",
+ "Batch_idx: 270 | Loss: (2.2435) | Acc: (28.97%) (10049/34688)\n",
+ "Batch_idx: 280 | Loss: (2.2257) | Acc: (29.32%) (10547/35968)\n",
+ "Batch_idx: 290 | Loss: (2.2090) | Acc: (29.68%) (11054/37248)\n",
+ "Batch_idx: 300 | Loss: (2.1929) | Acc: (30.05%) (11578/38528)\n",
+ "Batch_idx: 310 | Loss: (2.1745) | Acc: (30.39%) (12099/39808)\n",
+ "Batch_idx: 320 | Loss: (2.1581) | Acc: (30.82%) (12662/41088)\n",
+ "Batch_idx: 330 | Loss: (2.1425) | Acc: (31.19%) (13214/42368)\n",
+ "Batch_idx: 340 | Loss: (2.1275) | Acc: (31.53%) (13762/43648)\n",
+ "Batch_idx: 350 | Loss: (2.1124) | Acc: (31.78%) (14280/44928)\n",
+ "Batch_idx: 360 | Loss: (2.0975) | Acc: (32.13%) (14848/46208)\n",
+ "Batch_idx: 370 | Loss: (2.0842) | Acc: (32.44%) (15405/47488)\n",
+ "Batch_idx: 380 | Loss: (2.0726) | Acc: (32.69%) (15943/48768)\n",
+ "Batch_idx: 390 | Loss: (2.0591) | Acc: (32.98%) (16489/50000)\n",
+ "# TEST : Loss: (1.4994) | Acc: (45.73%) (4573/10000)\n",
+ "--------------- 1 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (1.5272) | Acc: (45.31%) (58/128)\n",
+ "Batch_idx: 10 | Loss: (1.4751) | Acc: (47.73%) (672/1408)\n",
+ "Batch_idx: 20 | Loss: (1.5015) | Acc: (46.73%) (1256/2688)\n",
+ "Batch_idx: 30 | Loss: (1.5056) | Acc: (46.17%) (1832/3968)\n",
+ "Batch_idx: 40 | Loss: (1.4999) | Acc: (46.32%) (2431/5248)\n",
+ "Batch_idx: 50 | Loss: (1.5005) | Acc: (45.97%) (3001/6528)\n",
+ "Batch_idx: 60 | Loss: (1.4874) | Acc: (46.66%) (3643/7808)\n",
+ "Batch_idx: 70 | Loss: (1.4836) | Acc: (46.72%) (4246/9088)\n",
+ "Batch_idx: 80 | Loss: (1.4738) | Acc: (46.98%) (4871/10368)\n",
+ "Batch_idx: 90 | Loss: (1.4734) | Acc: (46.94%) (5467/11648)\n",
+ "Batch_idx: 100 | Loss: (1.4654) | Acc: (47.32%) (6117/12928)\n",
+ "Batch_idx: 110 | Loss: (1.4646) | Acc: (47.55%) (6756/14208)\n",
+ "Batch_idx: 120 | Loss: (1.4705) | Acc: (47.55%) (7365/15488)\n",
+ "Batch_idx: 130 | Loss: (1.4717) | Acc: (47.42%) (7952/16768)\n",
+ "Batch_idx: 140 | Loss: (1.4763) | Acc: (47.40%) (8555/18048)\n",
+ "Batch_idx: 150 | Loss: (1.4818) | Acc: (47.42%) (9166/19328)\n",
+ "Batch_idx: 160 | Loss: (1.4842) | Acc: (47.37%) (9761/20608)\n",
+ "Batch_idx: 170 | Loss: (1.4848) | Acc: (47.36%) (10367/21888)\n",
+ "Batch_idx: 180 | Loss: (1.4856) | Acc: (47.50%) (11005/23168)\n",
+ "Batch_idx: 190 | Loss: (1.4825) | Acc: (47.65%) (11649/24448)\n",
+ "Batch_idx: 200 | Loss: (1.4802) | Acc: (47.82%) (12303/25728)\n",
+ "Batch_idx: 210 | Loss: (1.4753) | Acc: (47.88%) (12931/27008)\n",
+ "Batch_idx: 220 | Loss: (1.4746) | Acc: (47.84%) (13532/28288)\n",
+ "Batch_idx: 230 | Loss: (1.4733) | Acc: (47.93%) (14173/29568)\n",
+ "Batch_idx: 240 | Loss: (1.4714) | Acc: (47.93%) (14784/30848)\n",
+ "Batch_idx: 250 | Loss: (1.4769) | Acc: (47.91%) (15391/32128)\n",
+ "Batch_idx: 260 | Loss: (1.4770) | Acc: (47.88%) (15995/33408)\n",
+ "Batch_idx: 270 | Loss: (1.4766) | Acc: (47.87%) (16605/34688)\n",
+ "Batch_idx: 280 | Loss: (1.4767) | Acc: (47.93%) (17238/35968)\n",
+ "Batch_idx: 290 | Loss: (1.4746) | Acc: (48.02%) (17887/37248)\n",
+ "Batch_idx: 300 | Loss: (1.4707) | Acc: (48.24%) (18584/38528)\n",
+ "Batch_idx: 310 | Loss: (1.4682) | Acc: (48.34%) (19245/39808)\n",
+ "Batch_idx: 320 | Loss: (1.4666) | Acc: (48.40%) (19886/41088)\n",
+ "Batch_idx: 330 | Loss: (1.4640) | Acc: (48.47%) (20534/42368)\n",
+ "Batch_idx: 340 | Loss: (1.4608) | Acc: (48.55%) (21193/43648)\n",
+ "Batch_idx: 350 | Loss: (1.4584) | Acc: (48.61%) (21838/44928)\n",
+ "Batch_idx: 360 | Loss: (1.4554) | Acc: (48.68%) (22492/46208)\n",
+ "Batch_idx: 370 | Loss: (1.4548) | Acc: (48.77%) (23159/47488)\n",
+ "Batch_idx: 380 | Loss: (1.4526) | Acc: (48.88%) (23836/48768)\n",
+ "Batch_idx: 390 | Loss: (1.4490) | Acc: (48.98%) (24491/50000)\n",
+ "# TEST : Loss: (1.5942) | Acc: (53.26%) (5326/10000)\n",
+ "--------------- 2 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (1.3294) | Acc: (50.00%) (64/128)\n",
+ "Batch_idx: 10 | Loss: (1.3550) | Acc: (54.55%) (768/1408)\n",
+ "Batch_idx: 20 | Loss: (1.3358) | Acc: (54.80%) (1473/2688)\n",
+ "Batch_idx: 30 | Loss: (1.3282) | Acc: (55.04%) (2184/3968)\n",
+ "Batch_idx: 40 | Loss: (1.3079) | Acc: (55.32%) (2903/5248)\n",
+ "Batch_idx: 50 | Loss: (1.3033) | Acc: (55.10%) (3597/6528)\n",
+ "Batch_idx: 60 | Loss: (1.3054) | Acc: (55.16%) (4307/7808)\n",
+ "Batch_idx: 70 | Loss: (1.2969) | Acc: (55.15%) (5012/9088)\n",
+ "Batch_idx: 80 | Loss: (1.3009) | Acc: (55.50%) (5754/10368)\n",
+ "Batch_idx: 90 | Loss: (1.2981) | Acc: (55.68%) (6486/11648)\n",
+ "Batch_idx: 100 | Loss: (1.2953) | Acc: (55.89%) (7226/12928)\n",
+ "Batch_idx: 110 | Loss: (1.2898) | Acc: (55.93%) (7946/14208)\n",
+ "Batch_idx: 120 | Loss: (1.2842) | Acc: (55.95%) (8665/15488)\n",
+ "Batch_idx: 130 | Loss: (1.2855) | Acc: (55.94%) (9380/16768)\n",
+ "Batch_idx: 140 | Loss: (1.2811) | Acc: (56.07%) (10120/18048)\n",
+ "Batch_idx: 150 | Loss: (1.2890) | Acc: (55.89%) (10802/19328)\n",
+ "Batch_idx: 160 | Loss: (1.2899) | Acc: (55.63%) (11465/20608)\n",
+ "Batch_idx: 170 | Loss: (1.2932) | Acc: (55.66%) (12182/21888)\n",
+ "Batch_idx: 180 | Loss: (1.2922) | Acc: (55.69%) (12903/23168)\n",
+ "Batch_idx: 190 | Loss: (1.2943) | Acc: (55.66%) (13607/24448)\n",
+ "Batch_idx: 200 | Loss: (1.2925) | Acc: (55.67%) (14324/25728)\n",
+ "Batch_idx: 210 | Loss: (1.2944) | Acc: (55.66%) (15034/27008)\n",
+ "Batch_idx: 220 | Loss: (1.2930) | Acc: (55.58%) (15722/28288)\n",
+ "Batch_idx: 230 | Loss: (1.2884) | Acc: (55.69%) (16465/29568)\n",
+ "Batch_idx: 240 | Loss: (1.2877) | Acc: (55.68%) (17175/30848)\n",
+ "Batch_idx: 250 | Loss: (1.2837) | Acc: (55.82%) (17935/32128)\n",
+ "Batch_idx: 260 | Loss: (1.2802) | Acc: (55.88%) (18667/33408)\n",
+ "Batch_idx: 270 | Loss: (1.2794) | Acc: (55.88%) (19383/34688)\n",
+ "Batch_idx: 280 | Loss: (1.2759) | Acc: (56.01%) (20144/35968)\n",
+ "Batch_idx: 290 | Loss: (1.2734) | Acc: (56.11%) (20899/37248)\n",
+ "Batch_idx: 300 | Loss: (1.2711) | Acc: (56.23%) (21664/38528)\n",
+ "Batch_idx: 310 | Loss: (1.2695) | Acc: (56.31%) (22417/39808)\n",
+ "Batch_idx: 320 | Loss: (1.2667) | Acc: (56.36%) (23156/41088)\n",
+ "Batch_idx: 330 | Loss: (1.2655) | Acc: (56.39%) (23891/42368)\n",
+ "Batch_idx: 340 | Loss: (1.2642) | Acc: (56.39%) (24612/43648)\n",
+ "Batch_idx: 350 | Loss: (1.2612) | Acc: (56.43%) (25355/44928)\n",
+ "Batch_idx: 360 | Loss: (1.2591) | Acc: (56.53%) (26123/46208)\n",
+ "Batch_idx: 370 | Loss: (1.2562) | Acc: (56.61%) (26884/47488)\n",
+ "Batch_idx: 380 | Loss: (1.2551) | Acc: (56.68%) (27643/48768)\n",
+ "Batch_idx: 390 | Loss: (1.2519) | Acc: (56.71%) (28357/50000)\n",
+ "# TEST : Loss: (1.1601) | Acc: (59.29%) (5929/10000)\n",
+ "--------------- 3 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (1.2399) | Acc: (53.12%) (68/128)\n",
+ "Batch_idx: 10 | Loss: (1.1675) | Acc: (58.10%) (818/1408)\n",
+ "Batch_idx: 20 | Loss: (1.1488) | Acc: (58.67%) (1577/2688)\n",
+ "Batch_idx: 30 | Loss: (1.1262) | Acc: (59.30%) (2353/3968)\n",
+ "Batch_idx: 40 | Loss: (1.1289) | Acc: (59.87%) (3142/5248)\n",
+ "Batch_idx: 50 | Loss: (1.1367) | Acc: (59.96%) (3914/6528)\n",
+ "Batch_idx: 60 | Loss: (1.1360) | Acc: (60.21%) (4701/7808)\n",
+ "Batch_idx: 70 | Loss: (1.1407) | Acc: (60.22%) (5473/9088)\n",
+ "Batch_idx: 80 | Loss: (1.1535) | Acc: (59.81%) (6201/10368)\n",
+ "Batch_idx: 90 | Loss: (1.1610) | Acc: (59.65%) (6948/11648)\n",
+ "Batch_idx: 100 | Loss: (1.1648) | Acc: (59.65%) (7711/12928)\n",
+ "Batch_idx: 110 | Loss: (1.1641) | Acc: (59.70%) (8482/14208)\n",
+ "Batch_idx: 120 | Loss: (1.1606) | Acc: (59.84%) (9268/15488)\n",
+ "Batch_idx: 130 | Loss: (1.1597) | Acc: (59.77%) (10022/16768)\n",
+ "Batch_idx: 140 | Loss: (1.1565) | Acc: (60.01%) (10831/18048)\n",
+ "Batch_idx: 150 | Loss: (1.1517) | Acc: (60.16%) (11628/19328)\n",
+ "Batch_idx: 160 | Loss: (1.1536) | Acc: (60.11%) (12387/20608)\n",
+ "Batch_idx: 170 | Loss: (1.1480) | Acc: (60.27%) (13192/21888)\n",
+ "Batch_idx: 180 | Loss: (1.1497) | Acc: (60.29%) (13967/23168)\n",
+ "Batch_idx: 190 | Loss: (1.1476) | Acc: (60.34%) (14752/24448)\n",
+ "Batch_idx: 200 | Loss: (1.1495) | Acc: (60.44%) (15549/25728)\n",
+ "Batch_idx: 210 | Loss: (1.1495) | Acc: (60.49%) (16336/27008)\n",
+ "Batch_idx: 220 | Loss: (1.1458) | Acc: (60.57%) (17133/28288)\n",
+ "Batch_idx: 230 | Loss: (1.1428) | Acc: (60.65%) (17934/29568)\n",
+ "Batch_idx: 240 | Loss: (1.1392) | Acc: (60.75%) (18741/30848)\n",
+ "Batch_idx: 250 | Loss: (1.1397) | Acc: (60.74%) (19515/32128)\n",
+ "Batch_idx: 260 | Loss: (1.1406) | Acc: (60.74%) (20291/33408)\n",
+ "Batch_idx: 270 | Loss: (1.1387) | Acc: (60.76%) (21078/34688)\n",
+ "Batch_idx: 280 | Loss: (1.1376) | Acc: (60.86%) (21890/35968)\n",
+ "Batch_idx: 290 | Loss: (1.1354) | Acc: (60.92%) (22693/37248)\n",
+ "Batch_idx: 300 | Loss: (1.1356) | Acc: (60.97%) (23489/38528)\n",
+ "Batch_idx: 310 | Loss: (1.1338) | Acc: (60.98%) (24275/39808)\n",
+ "Batch_idx: 320 | Loss: (1.1342) | Acc: (61.00%) (25064/41088)\n",
+ "Batch_idx: 330 | Loss: (1.1313) | Acc: (61.05%) (25865/42368)\n",
+ "Batch_idx: 340 | Loss: (1.1334) | Acc: (60.99%) (26620/43648)\n",
+ "Batch_idx: 350 | Loss: (1.1423) | Acc: (60.73%) (27283/44928)\n",
+ "Batch_idx: 360 | Loss: (1.1467) | Acc: (60.62%) (28009/46208)\n",
+ "Batch_idx: 370 | Loss: (1.1514) | Acc: (60.53%) (28746/47488)\n",
+ "Batch_idx: 380 | Loss: (1.1563) | Acc: (60.42%) (29464/48768)\n",
+ "Batch_idx: 390 | Loss: (1.1605) | Acc: (60.32%) (30160/50000)\n",
+ "# TEST : Loss: (1.2462) | Acc: (54.88%) (5488/10000)\n",
+ "--------------- 4 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (1.4096) | Acc: (52.34%) (67/128)\n",
+ "Batch_idx: 10 | Loss: (1.2851) | Acc: (58.10%) (818/1408)\n",
+ "Batch_idx: 20 | Loss: (1.2482) | Acc: (58.85%) (1582/2688)\n",
+ "Batch_idx: 30 | Loss: (1.2693) | Acc: (58.22%) (2310/3968)\n",
+ "Batch_idx: 40 | Loss: (1.2667) | Acc: (57.91%) (3039/5248)\n",
+ "Batch_idx: 50 | Loss: (1.2635) | Acc: (57.75%) (3770/6528)\n",
+ "Batch_idx: 60 | Loss: (1.2533) | Acc: (58.34%) (4555/7808)\n",
+ "Batch_idx: 70 | Loss: (1.2463) | Acc: (58.63%) (5328/9088)\n",
+ "Batch_idx: 80 | Loss: (1.2344) | Acc: (58.87%) (6104/10368)\n",
+ "Batch_idx: 90 | Loss: (1.2354) | Acc: (58.81%) (6850/11648)\n",
+ "Batch_idx: 100 | Loss: (1.2477) | Acc: (58.20%) (7524/12928)\n",
+ "Batch_idx: 110 | Loss: (1.2766) | Acc: (57.25%) (8134/14208)\n",
+ "Batch_idx: 120 | Loss: (1.2961) | Acc: (56.51%) (8753/15488)\n",
+ "Batch_idx: 130 | Loss: (1.3124) | Acc: (55.95%) (9382/16768)\n",
+ "Batch_idx: 140 | Loss: (1.3238) | Acc: (55.49%) (10014/18048)\n",
+ "Batch_idx: 150 | Loss: (1.3335) | Acc: (55.13%) (10656/19328)\n",
+ "Batch_idx: 160 | Loss: (1.3325) | Acc: (55.12%) (11359/20608)\n",
+ "Batch_idx: 170 | Loss: (1.3322) | Acc: (55.06%) (12052/21888)\n",
+ "Batch_idx: 180 | Loss: (1.3321) | Acc: (54.96%) (12734/23168)\n",
+ "Batch_idx: 190 | Loss: (1.3295) | Acc: (55.06%) (13461/24448)\n",
+ "Batch_idx: 200 | Loss: (1.3268) | Acc: (55.13%) (14184/25728)\n",
+ "Batch_idx: 210 | Loss: (1.3294) | Acc: (55.07%) (14872/27008)\n",
+ "Batch_idx: 220 | Loss: (1.3288) | Acc: (55.17%) (15606/28288)\n",
+ "Batch_idx: 230 | Loss: (1.3253) | Acc: (55.26%) (16339/29568)\n",
+ "Batch_idx: 240 | Loss: (1.3231) | Acc: (55.35%) (17075/30848)\n",
+ "Batch_idx: 250 | Loss: (1.3204) | Acc: (55.50%) (17832/32128)\n",
+ "Batch_idx: 260 | Loss: (1.3153) | Acc: (55.67%) (18598/33408)\n",
+ "Batch_idx: 270 | Loss: (1.3089) | Acc: (55.88%) (19383/34688)\n",
+ "Batch_idx: 280 | Loss: (1.3089) | Acc: (55.86%) (20091/35968)\n",
+ "Batch_idx: 290 | Loss: (1.3029) | Acc: (56.04%) (20873/37248)\n",
+ "Batch_idx: 300 | Loss: (1.3010) | Acc: (56.16%) (21638/38528)\n",
+ "Batch_idx: 310 | Loss: (1.2972) | Acc: (56.28%) (22403/39808)\n",
+ "Batch_idx: 320 | Loss: (1.2938) | Acc: (56.43%) (23186/41088)\n",
+ "Batch_idx: 330 | Loss: (1.2904) | Acc: (56.55%) (23959/42368)\n",
+ "Batch_idx: 340 | Loss: (1.2862) | Acc: (56.70%) (24749/43648)\n",
+ "Batch_idx: 350 | Loss: (1.2814) | Acc: (56.89%) (25561/44928)\n",
+ "Batch_idx: 360 | Loss: (1.2785) | Acc: (56.99%) (26335/46208)\n",
+ "Batch_idx: 370 | Loss: (1.2763) | Acc: (57.07%) (27103/47488)\n",
+ "Batch_idx: 380 | Loss: (1.2731) | Acc: (57.14%) (27865/48768)\n",
+ "Batch_idx: 390 | Loss: (1.2688) | Acc: (57.25%) (28624/50000)\n",
+ "# TEST : Loss: (1.1549) | Acc: (60.04%) (6004/10000)\n",
+ "--------------- 5 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (1.0893) | Acc: (65.62%) (84/128)\n",
+ "Batch_idx: 10 | Loss: (1.0880) | Acc: (62.71%) (883/1408)\n",
+ "Batch_idx: 20 | Loss: (1.0549) | Acc: (64.40%) (1731/2688)\n",
+ "Batch_idx: 30 | Loss: (1.0474) | Acc: (64.99%) (2579/3968)\n",
+ "Batch_idx: 40 | Loss: (1.0604) | Acc: (64.04%) (3361/5248)\n",
+ "Batch_idx: 50 | Loss: (1.0637) | Acc: (63.73%) (4160/6528)\n",
+ "Batch_idx: 60 | Loss: (1.0567) | Acc: (63.90%) (4989/7808)\n",
+ "Batch_idx: 70 | Loss: (1.0540) | Acc: (64.21%) (5835/9088)\n",
+ "Batch_idx: 80 | Loss: (1.0562) | Acc: (64.12%) (6648/10368)\n",
+ "Batch_idx: 90 | Loss: (1.0583) | Acc: (64.05%) (7460/11648)\n",
+ "Batch_idx: 100 | Loss: (1.0550) | Acc: (64.16%) (8295/12928)\n",
+ "Batch_idx: 110 | Loss: (1.0587) | Acc: (64.20%) (9122/14208)\n",
+ "Batch_idx: 120 | Loss: (1.0600) | Acc: (64.20%) (9943/15488)\n",
+ "Batch_idx: 130 | Loss: (1.0598) | Acc: (64.27%) (10777/16768)\n",
+ "Batch_idx: 140 | Loss: (1.0589) | Acc: (64.43%) (11628/18048)\n",
+ "Batch_idx: 150 | Loss: (1.0591) | Acc: (64.50%) (12467/19328)\n",
+ "Batch_idx: 160 | Loss: (1.0595) | Acc: (64.43%) (13277/20608)\n",
+ "Batch_idx: 170 | Loss: (1.0575) | Acc: (64.45%) (14107/21888)\n",
+ "Batch_idx: 180 | Loss: (1.0548) | Acc: (64.49%) (14941/23168)\n",
+ "Batch_idx: 190 | Loss: (1.0522) | Acc: (64.54%) (15779/24448)\n",
+ "Batch_idx: 200 | Loss: (1.0517) | Acc: (64.60%) (16620/25728)\n",
+ "Batch_idx: 210 | Loss: (1.0491) | Acc: (64.66%) (17464/27008)\n",
+ "Batch_idx: 220 | Loss: (1.0473) | Acc: (64.70%) (18303/28288)\n",
+ "Batch_idx: 230 | Loss: (1.0450) | Acc: (64.74%) (19142/29568)\n",
+ "Batch_idx: 240 | Loss: (1.0424) | Acc: (64.73%) (19969/30848)\n",
+ "Batch_idx: 250 | Loss: (1.0453) | Acc: (64.78%) (20811/32128)\n",
+ "Batch_idx: 260 | Loss: (1.0437) | Acc: (64.78%) (21642/33408)\n",
+ "Batch_idx: 270 | Loss: (1.0424) | Acc: (64.86%) (22498/34688)\n",
+ "Batch_idx: 280 | Loss: (1.0420) | Acc: (64.82%) (23315/35968)\n",
+ "Batch_idx: 290 | Loss: (1.0401) | Acc: (64.89%) (24170/37248)\n",
+ "Batch_idx: 300 | Loss: (1.0385) | Acc: (64.93%) (25017/38528)\n",
+ "Batch_idx: 310 | Loss: (1.0395) | Acc: (64.92%) (25844/39808)\n",
+ "Batch_idx: 320 | Loss: (1.0369) | Acc: (64.98%) (26699/41088)\n",
+ "Batch_idx: 330 | Loss: (1.0361) | Acc: (65.00%) (27538/42368)\n",
+ "Batch_idx: 340 | Loss: (1.0344) | Acc: (65.08%) (28408/43648)\n",
+ "Batch_idx: 350 | Loss: (1.0322) | Acc: (65.12%) (29259/44928)\n",
+ "Batch_idx: 360 | Loss: (1.0315) | Acc: (65.14%) (30099/46208)\n",
+ "Batch_idx: 370 | Loss: (1.0347) | Acc: (65.08%) (30904/47488)\n",
+ "Batch_idx: 380 | Loss: (1.0358) | Acc: (65.04%) (31718/48768)\n",
+ "Batch_idx: 390 | Loss: (1.0389) | Acc: (64.96%) (32479/50000)\n",
+ "# TEST : Loss: (2.5007) | Acc: (56.74%) (5674/10000)\n",
+ "--------------- 6 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (1.1126) | Acc: (64.84%) (83/128)\n",
+ "Batch_idx: 10 | Loss: (1.2058) | Acc: (58.81%) (828/1408)\n",
+ "Batch_idx: 20 | Loss: (1.3262) | Acc: (56.10%) (1508/2688)\n",
+ "Batch_idx: 30 | Loss: (1.3574) | Acc: (54.81%) (2175/3968)\n",
+ "Batch_idx: 40 | Loss: (1.3783) | Acc: (54.74%) (2873/5248)\n",
+ "Batch_idx: 50 | Loss: (1.3752) | Acc: (54.83%) (3579/6528)\n",
+ "Batch_idx: 60 | Loss: (1.3751) | Acc: (54.59%) (4262/7808)\n",
+ "Batch_idx: 70 | Loss: (1.3766) | Acc: (54.08%) (4915/9088)\n",
+ "Batch_idx: 80 | Loss: (1.3763) | Acc: (53.96%) (5595/10368)\n",
+ "Batch_idx: 90 | Loss: (1.3611) | Acc: (54.44%) (6341/11648)\n",
+ "Batch_idx: 100 | Loss: (1.3621) | Acc: (54.45%) (7039/12928)\n",
+ "Batch_idx: 110 | Loss: (1.3569) | Acc: (54.61%) (7759/14208)\n",
+ "Batch_idx: 120 | Loss: (1.3511) | Acc: (54.74%) (8478/15488)\n",
+ "Batch_idx: 130 | Loss: (1.3421) | Acc: (55.10%) (9240/16768)\n",
+ "Batch_idx: 140 | Loss: (1.3363) | Acc: (55.23%) (9968/18048)\n",
+ "Batch_idx: 150 | Loss: (1.3300) | Acc: (55.45%) (10718/19328)\n",
+ "Batch_idx: 160 | Loss: (1.3207) | Acc: (55.75%) (11488/20608)\n",
+ "Batch_idx: 170 | Loss: (1.3124) | Acc: (55.97%) (12251/21888)\n",
+ "Batch_idx: 180 | Loss: (1.3009) | Acc: (56.30%) (13044/23168)\n",
+ "Batch_idx: 190 | Loss: (1.2947) | Acc: (56.43%) (13796/24448)\n",
+ "Batch_idx: 200 | Loss: (1.2854) | Acc: (56.61%) (14565/25728)\n",
+ "Batch_idx: 210 | Loss: (1.2782) | Acc: (56.81%) (15344/27008)\n",
+ "Batch_idx: 220 | Loss: (1.2706) | Acc: (57.14%) (16163/28288)\n",
+ "Batch_idx: 230 | Loss: (1.2663) | Acc: (57.33%) (16950/29568)\n",
+ "Batch_idx: 240 | Loss: (1.2649) | Acc: (57.46%) (17726/30848)\n",
+ "Batch_idx: 250 | Loss: (1.2602) | Acc: (57.62%) (18512/32128)\n",
+ "Batch_idx: 260 | Loss: (1.2548) | Acc: (57.83%) (19321/33408)\n",
+ "Batch_idx: 270 | Loss: (1.2510) | Acc: (57.95%) (20103/34688)\n",
+ "Batch_idx: 280 | Loss: (1.2449) | Acc: (58.14%) (20913/35968)\n",
+ "Batch_idx: 290 | Loss: (1.2414) | Acc: (58.26%) (21701/37248)\n",
+ "Batch_idx: 300 | Loss: (1.2367) | Acc: (58.45%) (22518/38528)\n",
+ "Batch_idx: 310 | Loss: (1.2355) | Acc: (58.52%) (23295/39808)\n",
+ "Batch_idx: 320 | Loss: (1.2325) | Acc: (58.62%) (24085/41088)\n",
+ "Batch_idx: 330 | Loss: (1.2295) | Acc: (58.72%) (24878/42368)\n",
+ "Batch_idx: 340 | Loss: (1.2256) | Acc: (58.83%) (25676/43648)\n",
+ "Batch_idx: 350 | Loss: (1.2218) | Acc: (58.96%) (26488/44928)\n",
+ "Batch_idx: 360 | Loss: (1.2203) | Acc: (59.03%) (27278/46208)\n",
+ "Batch_idx: 370 | Loss: (1.2208) | Acc: (58.99%) (28013/47488)\n",
+ "Batch_idx: 380 | Loss: (1.2206) | Acc: (59.05%) (28799/48768)\n",
+ "Batch_idx: 390 | Loss: (1.2204) | Acc: (59.10%) (29548/50000)\n",
+ "# TEST : Loss: (1.9621) | Acc: (57.81%) (5781/10000)\n",
+ "--------------- 7 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (1.2813) | Acc: (59.38%) (76/128)\n",
+ "Batch_idx: 10 | Loss: (1.2505) | Acc: (57.60%) (811/1408)\n",
+ "Batch_idx: 20 | Loss: (1.2306) | Acc: (58.11%) (1562/2688)\n",
+ "Batch_idx: 30 | Loss: (1.2324) | Acc: (58.04%) (2303/3968)\n",
+ "Batch_idx: 40 | Loss: (1.2152) | Acc: (58.90%) (3091/5248)\n",
+ "Batch_idx: 50 | Loss: (1.2218) | Acc: (58.76%) (3836/6528)\n",
+ "Batch_idx: 60 | Loss: (1.2881) | Acc: (56.79%) (4434/7808)\n",
+ "Batch_idx: 70 | Loss: (1.3193) | Acc: (55.50%) (5044/9088)\n",
+ "Batch_idx: 80 | Loss: (1.3446) | Acc: (54.60%) (5661/10368)\n",
+ "Batch_idx: 90 | Loss: (1.3764) | Acc: (53.76%) (6262/11648)\n",
+ "Batch_idx: 100 | Loss: (1.4015) | Acc: (52.80%) (6826/12928)\n",
+ "Batch_idx: 110 | Loss: (1.4086) | Acc: (52.50%) (7459/14208)\n",
+ "Batch_idx: 120 | Loss: (1.4221) | Acc: (51.96%) (8047/15488)\n",
+ "Batch_idx: 130 | Loss: (1.4285) | Acc: (51.80%) (8685/16768)\n",
+ "Batch_idx: 140 | Loss: (1.4230) | Acc: (51.88%) (9364/18048)\n",
+ "Batch_idx: 150 | Loss: (1.4227) | Acc: (51.91%) (10033/19328)\n",
+ "Batch_idx: 160 | Loss: (1.4207) | Acc: (51.89%) (10693/20608)\n",
+ "Batch_idx: 170 | Loss: (1.4150) | Acc: (52.06%) (11394/21888)\n",
+ "Batch_idx: 180 | Loss: (1.4105) | Acc: (52.12%) (12075/23168)\n",
+ "Batch_idx: 190 | Loss: (1.4060) | Acc: (52.25%) (12775/24448)\n",
+ "Batch_idx: 200 | Loss: (1.4054) | Acc: (52.23%) (13438/25728)\n",
+ "Batch_idx: 210 | Loss: (1.4055) | Acc: (52.19%) (14096/27008)\n",
+ "Batch_idx: 220 | Loss: (1.4047) | Acc: (52.23%) (14775/28288)\n",
+ "Batch_idx: 230 | Loss: (1.4010) | Acc: (52.31%) (15468/29568)\n",
+ "Batch_idx: 240 | Loss: (1.4007) | Acc: (52.27%) (16124/30848)\n",
+ "Batch_idx: 250 | Loss: (1.4000) | Acc: (52.40%) (16835/32128)\n",
+ "Batch_idx: 260 | Loss: (1.3941) | Acc: (52.67%) (17595/33408)\n",
+ "Batch_idx: 270 | Loss: (1.3903) | Acc: (52.81%) (18317/34688)\n",
+ "Batch_idx: 280 | Loss: (1.3861) | Acc: (53.01%) (19066/35968)\n",
+ "Batch_idx: 290 | Loss: (1.3831) | Acc: (53.20%) (19815/37248)\n",
+ "Batch_idx: 300 | Loss: (1.3803) | Acc: (53.30%) (20535/38528)\n",
+ "Batch_idx: 310 | Loss: (1.3770) | Acc: (53.38%) (21251/39808)\n",
+ "Batch_idx: 320 | Loss: (1.3726) | Acc: (53.47%) (21971/41088)\n",
+ "Batch_idx: 330 | Loss: (1.3687) | Acc: (53.63%) (22722/42368)\n",
+ "Batch_idx: 340 | Loss: (1.3662) | Acc: (53.68%) (23431/43648)\n",
+ "Batch_idx: 350 | Loss: (1.3627) | Acc: (53.76%) (24153/44928)\n",
+ "Batch_idx: 360 | Loss: (1.3608) | Acc: (53.80%) (24862/46208)\n",
+ "Batch_idx: 370 | Loss: (1.3575) | Acc: (53.94%) (25616/47488)\n",
+ "Batch_idx: 380 | Loss: (1.3587) | Acc: (53.94%) (26305/48768)\n",
+ "Batch_idx: 390 | Loss: (1.3558) | Acc: (54.00%) (27000/50000)\n",
+ "# TEST : Loss: (1.3473) | Acc: (57.70%) (5770/10000)\n",
+ "--------------- 8 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (1.2674) | Acc: (58.59%) (75/128)\n",
+ "Batch_idx: 10 | Loss: (1.2259) | Acc: (57.53%) (810/1408)\n",
+ "Batch_idx: 20 | Loss: (1.2498) | Acc: (56.88%) (1529/2688)\n",
+ "Batch_idx: 30 | Loss: (1.2511) | Acc: (56.85%) (2256/3968)\n",
+ "Batch_idx: 40 | Loss: (1.2453) | Acc: (57.41%) (3013/5248)\n",
+ "Batch_idx: 50 | Loss: (1.2420) | Acc: (57.44%) (3750/6528)\n",
+ "Batch_idx: 60 | Loss: (1.2384) | Acc: (57.54%) (4493/7808)\n",
+ "Batch_idx: 70 | Loss: (1.2304) | Acc: (57.80%) (5253/9088)\n",
+ "Batch_idx: 80 | Loss: (1.2236) | Acc: (58.22%) (6036/10368)\n",
+ "Batch_idx: 90 | Loss: (1.2180) | Acc: (58.34%) (6796/11648)\n",
+ "Batch_idx: 100 | Loss: (1.2152) | Acc: (58.38%) (7548/12928)\n",
+ "Batch_idx: 110 | Loss: (1.2082) | Acc: (58.77%) (8350/14208)\n",
+ "Batch_idx: 120 | Loss: (1.2059) | Acc: (59.14%) (9160/15488)\n",
+ "Batch_idx: 130 | Loss: (1.2022) | Acc: (59.33%) (9949/16768)\n",
+ "Batch_idx: 140 | Loss: (1.1976) | Acc: (59.61%) (10758/18048)\n",
+ "Batch_idx: 150 | Loss: (1.1944) | Acc: (59.78%) (11554/19328)\n",
+ "Batch_idx: 160 | Loss: (1.1877) | Acc: (59.98%) (12361/20608)\n",
+ "Batch_idx: 170 | Loss: (1.1845) | Acc: (60.17%) (13170/21888)\n",
+ "Batch_idx: 180 | Loss: (1.1799) | Acc: (60.28%) (13965/23168)\n",
+ "Batch_idx: 190 | Loss: (1.1749) | Acc: (60.43%) (14774/24448)\n",
+ "Batch_idx: 200 | Loss: (1.1710) | Acc: (60.55%) (15579/25728)\n",
+ "Batch_idx: 210 | Loss: (1.1662) | Acc: (60.67%) (16386/27008)\n",
+ "Batch_idx: 220 | Loss: (1.1652) | Acc: (60.78%) (17193/28288)\n",
+ "Batch_idx: 230 | Loss: (1.1612) | Acc: (60.89%) (18003/29568)\n",
+ "Batch_idx: 240 | Loss: (1.1657) | Acc: (60.77%) (18746/30848)\n",
+ "Batch_idx: 250 | Loss: (1.1711) | Acc: (60.59%) (19467/32128)\n",
+ "Batch_idx: 260 | Loss: (1.1732) | Acc: (60.52%) (20218/33408)\n",
+ "Batch_idx: 270 | Loss: (1.1754) | Acc: (60.43%) (20962/34688)\n",
+ "Batch_idx: 280 | Loss: (1.1773) | Acc: (60.41%) (21729/35968)\n",
+ "Batch_idx: 290 | Loss: (1.1798) | Acc: (60.35%) (22479/37248)\n",
+ "Batch_idx: 300 | Loss: (1.1799) | Acc: (60.32%) (23242/38528)\n",
+ "Batch_idx: 310 | Loss: (1.1785) | Acc: (60.36%) (24030/39808)\n",
+ "Batch_idx: 320 | Loss: (1.1782) | Acc: (60.29%) (24772/41088)\n",
+ "Batch_idx: 330 | Loss: (1.1771) | Acc: (60.33%) (25560/42368)\n",
+ "Batch_idx: 340 | Loss: (1.1758) | Acc: (60.39%) (26361/43648)\n",
+ "Batch_idx: 350 | Loss: (1.1726) | Acc: (60.49%) (27178/44928)\n",
+ "Batch_idx: 360 | Loss: (1.1710) | Acc: (60.49%) (27949/46208)\n",
+ "Batch_idx: 370 | Loss: (1.1709) | Acc: (60.48%) (28721/47488)\n",
+ "Batch_idx: 380 | Loss: (1.1685) | Acc: (60.57%) (29540/48768)\n",
+ "Batch_idx: 390 | Loss: (1.1672) | Acc: (60.61%) (30304/50000)\n",
+ "# TEST : Loss: (1.4486) | Acc: (51.91%) (5191/10000)\n",
+ "--------------- 9 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (1.0208) | Acc: (62.50%) (80/128)\n",
+ "Batch_idx: 10 | Loss: (1.1209) | Acc: (60.80%) (856/1408)\n",
+ "Batch_idx: 20 | Loss: (1.1610) | Acc: (60.04%) (1614/2688)\n",
+ "Batch_idx: 30 | Loss: (1.1439) | Acc: (60.74%) (2410/3968)\n",
+ "Batch_idx: 40 | Loss: (1.1365) | Acc: (60.67%) (3184/5248)\n",
+ "Batch_idx: 50 | Loss: (1.1332) | Acc: (60.62%) (3957/6528)\n",
+ "Batch_idx: 60 | Loss: (1.1275) | Acc: (61.10%) (4771/7808)\n",
+ "Batch_idx: 70 | Loss: (1.1239) | Acc: (61.19%) (5561/9088)\n",
+ "Batch_idx: 80 | Loss: (1.1221) | Acc: (61.47%) (6373/10368)\n",
+ "Batch_idx: 90 | Loss: (1.1167) | Acc: (61.50%) (7163/11648)\n",
+ "Batch_idx: 100 | Loss: (1.1090) | Acc: (61.53%) (7955/12928)\n",
+ "Batch_idx: 110 | Loss: (1.1034) | Acc: (61.66%) (8761/14208)\n",
+ "Batch_idx: 120 | Loss: (1.1024) | Acc: (61.63%) (9545/15488)\n",
+ "Batch_idx: 130 | Loss: (1.0980) | Acc: (61.78%) (10359/16768)\n",
+ "Batch_idx: 140 | Loss: (1.0921) | Acc: (61.99%) (11188/18048)\n",
+ "Batch_idx: 150 | Loss: (1.0929) | Acc: (61.87%) (11959/19328)\n",
+ "Batch_idx: 160 | Loss: (1.0909) | Acc: (61.86%) (12749/20608)\n",
+ "Batch_idx: 170 | Loss: (1.0907) | Acc: (61.82%) (13531/21888)\n",
+ "Batch_idx: 180 | Loss: (1.0893) | Acc: (61.99%) (14363/23168)\n",
+ "Batch_idx: 190 | Loss: (1.0861) | Acc: (62.11%) (15185/24448)\n",
+ "Batch_idx: 200 | Loss: (1.0839) | Acc: (62.12%) (15982/25728)\n",
+ "Batch_idx: 210 | Loss: (1.0815) | Acc: (62.29%) (16822/27008)\n",
+ "Batch_idx: 220 | Loss: (1.0810) | Acc: (62.37%) (17642/28288)\n",
+ "Batch_idx: 230 | Loss: (1.0783) | Acc: (62.39%) (18447/29568)\n",
+ "Batch_idx: 240 | Loss: (1.0769) | Acc: (62.46%) (19268/30848)\n",
+ "Batch_idx: 250 | Loss: (1.0754) | Acc: (62.48%) (20072/32128)\n",
+ "Batch_idx: 260 | Loss: (1.0730) | Acc: (62.53%) (20889/33408)\n",
+ "Batch_idx: 270 | Loss: (1.0705) | Acc: (62.61%) (21719/34688)\n",
+ "Batch_idx: 280 | Loss: (1.0684) | Acc: (62.70%) (22552/35968)\n",
+ "Batch_idx: 290 | Loss: (1.0678) | Acc: (62.70%) (23356/37248)\n",
+ "Batch_idx: 300 | Loss: (1.0645) | Acc: (62.86%) (24219/38528)\n",
+ "Batch_idx: 310 | Loss: (1.0635) | Acc: (62.88%) (25032/39808)\n",
+ "Batch_idx: 320 | Loss: (1.0617) | Acc: (62.95%) (25863/41088)\n",
+ "Batch_idx: 330 | Loss: (1.0587) | Acc: (63.07%) (26721/42368)\n",
+ "Batch_idx: 340 | Loss: (1.0568) | Acc: (63.13%) (27557/43648)\n",
+ "Batch_idx: 350 | Loss: (1.0552) | Acc: (63.19%) (28389/44928)\n",
+ "Batch_idx: 360 | Loss: (1.0539) | Acc: (63.19%) (29201/46208)\n",
+ "Batch_idx: 370 | Loss: (1.0521) | Acc: (63.28%) (30050/47488)\n",
+ "Batch_idx: 380 | Loss: (1.0498) | Acc: (63.37%) (30904/48768)\n",
+ "Batch_idx: 390 | Loss: (1.0480) | Acc: (63.42%) (31709/50000)\n",
+ "# TEST : Loss: (1.0547) | Acc: (64.93%) (6493/10000)\n",
+ "--------------- 10 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (0.9723) | Acc: (68.75%) (88/128)\n",
+ "Batch_idx: 10 | Loss: (0.9503) | Acc: (66.97%) (943/1408)\n",
+ "Batch_idx: 20 | Loss: (0.9230) | Acc: (67.19%) (1806/2688)\n",
+ "Batch_idx: 30 | Loss: (0.9340) | Acc: (67.49%) (2678/3968)\n",
+ "Batch_idx: 40 | Loss: (0.9455) | Acc: (67.24%) (3529/5248)\n",
+ "Batch_idx: 50 | Loss: (0.9366) | Acc: (67.52%) (4408/6528)\n",
+ "Batch_idx: 60 | Loss: (0.9393) | Acc: (67.52%) (5272/7808)\n",
+ "Batch_idx: 70 | Loss: (0.9348) | Acc: (67.67%) (6150/9088)\n",
+ "Batch_idx: 80 | Loss: (0.9292) | Acc: (67.87%) (7037/10368)\n",
+ "Batch_idx: 90 | Loss: (0.9312) | Acc: (67.72%) (7888/11648)\n",
+ "Batch_idx: 100 | Loss: (0.9327) | Acc: (67.67%) (8749/12928)\n",
+ "Batch_idx: 110 | Loss: (0.9316) | Acc: (67.66%) (9613/14208)\n",
+ "Batch_idx: 120 | Loss: (0.9273) | Acc: (67.73%) (10490/15488)\n",
+ "Batch_idx: 130 | Loss: (0.9229) | Acc: (67.76%) (11362/16768)\n",
+ "Batch_idx: 140 | Loss: (0.9281) | Acc: (67.58%) (12196/18048)\n",
+ "Batch_idx: 150 | Loss: (0.9277) | Acc: (67.58%) (13062/19328)\n",
+ "Batch_idx: 160 | Loss: (0.9319) | Acc: (67.40%) (13889/20608)\n",
+ "Batch_idx: 170 | Loss: (0.9294) | Acc: (67.46%) (14766/21888)\n",
+ "Batch_idx: 180 | Loss: (0.9308) | Acc: (67.50%) (15638/23168)\n",
+ "Batch_idx: 190 | Loss: (0.9289) | Acc: (67.56%) (16517/24448)\n",
+ "Batch_idx: 200 | Loss: (0.9271) | Acc: (67.63%) (17400/25728)\n",
+ "Batch_idx: 210 | Loss: (0.9291) | Acc: (67.51%) (18234/27008)\n",
+ "Batch_idx: 220 | Loss: (0.9287) | Acc: (67.44%) (19077/28288)\n",
+ "Batch_idx: 230 | Loss: (0.9274) | Acc: (67.43%) (19937/29568)\n",
+ "Batch_idx: 240 | Loss: (0.9253) | Acc: (67.58%) (20847/30848)\n",
+ "Batch_idx: 250 | Loss: (0.9240) | Acc: (67.61%) (21721/32128)\n",
+ "Batch_idx: 260 | Loss: (0.9239) | Acc: (67.61%) (22588/33408)\n",
+ "Batch_idx: 270 | Loss: (0.9243) | Acc: (67.60%) (23448/34688)\n",
+ "Batch_idx: 280 | Loss: (0.9212) | Acc: (67.69%) (24347/35968)\n",
+ "Batch_idx: 290 | Loss: (0.9203) | Acc: (67.74%) (25230/37248)\n",
+ "Batch_idx: 300 | Loss: (0.9193) | Acc: (67.76%) (26106/38528)\n",
+ "Batch_idx: 310 | Loss: (0.9180) | Acc: (67.85%) (27008/39808)\n",
+ "Batch_idx: 320 | Loss: (0.9162) | Acc: (67.92%) (27906/41088)\n",
+ "Batch_idx: 330 | Loss: (0.9161) | Acc: (67.91%) (28774/42368)\n",
+ "Batch_idx: 340 | Loss: (0.9136) | Acc: (67.99%) (29676/43648)\n",
+ "Batch_idx: 350 | Loss: (0.9138) | Acc: (67.99%) (30545/44928)\n",
+ "Batch_idx: 360 | Loss: (0.9129) | Acc: (68.03%) (31435/46208)\n",
+ "Batch_idx: 370 | Loss: (0.9135) | Acc: (67.98%) (32284/47488)\n",
+ "Batch_idx: 380 | Loss: (0.9131) | Acc: (68.02%) (33171/48768)\n",
+ "Batch_idx: 390 | Loss: (0.9131) | Acc: (68.04%) (34022/50000)\n",
+ "# TEST : Loss: (1.2420) | Acc: (64.36%) (6436/10000)\n",
+ "--------------- 11 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (0.9254) | Acc: (74.22%) (95/128)\n",
+ "Batch_idx: 10 | Loss: (0.9103) | Acc: (69.53%) (979/1408)\n",
+ "Batch_idx: 20 | Loss: (0.8921) | Acc: (69.35%) (1864/2688)\n",
+ "Batch_idx: 30 | Loss: (0.8826) | Acc: (69.96%) (2776/3968)\n",
+ "Batch_idx: 40 | Loss: (0.8781) | Acc: (70.08%) (3678/5248)\n",
+ "Batch_idx: 50 | Loss: (0.8747) | Acc: (69.91%) (4564/6528)\n",
+ "Batch_idx: 60 | Loss: (0.8738) | Acc: (69.90%) (5458/7808)\n",
+ "Batch_idx: 70 | Loss: (0.8776) | Acc: (69.43%) (6310/9088)\n",
+ "Batch_idx: 80 | Loss: (0.8798) | Acc: (69.19%) (7174/10368)\n",
+ "Batch_idx: 90 | Loss: (0.8799) | Acc: (69.22%) (8063/11648)\n",
+ "Batch_idx: 100 | Loss: (0.8740) | Acc: (69.47%) (8981/12928)\n",
+ "Batch_idx: 110 | Loss: (0.8698) | Acc: (69.64%) (9894/14208)\n",
+ "Batch_idx: 120 | Loss: (0.8666) | Acc: (69.74%) (10802/15488)\n",
+ "Batch_idx: 130 | Loss: (0.8635) | Acc: (69.81%) (11706/16768)\n",
+ "Batch_idx: 140 | Loss: (0.8603) | Acc: (69.88%) (12612/18048)\n",
+ "Batch_idx: 150 | Loss: (0.8592) | Acc: (69.98%) (13525/19328)\n",
+ "Batch_idx: 160 | Loss: (0.8606) | Acc: (69.90%) (14404/20608)\n",
+ "Batch_idx: 170 | Loss: (0.8625) | Acc: (69.97%) (15315/21888)\n",
+ "Batch_idx: 180 | Loss: (0.8628) | Acc: (69.97%) (16211/23168)\n",
+ "Batch_idx: 190 | Loss: (0.8623) | Acc: (69.99%) (17111/24448)\n",
+ "Batch_idx: 200 | Loss: (0.8609) | Acc: (70.04%) (18021/25728)\n",
+ "Batch_idx: 210 | Loss: (0.8615) | Acc: (70.08%) (18927/27008)\n",
+ "Batch_idx: 220 | Loss: (0.8586) | Acc: (70.19%) (19855/28288)\n",
+ "Batch_idx: 230 | Loss: (0.8556) | Acc: (70.31%) (20789/29568)\n",
+ "Batch_idx: 240 | Loss: (0.8558) | Acc: (70.28%) (21681/30848)\n",
+ "Batch_idx: 250 | Loss: (0.8563) | Acc: (70.26%) (22574/32128)\n",
+ "Batch_idx: 260 | Loss: (0.8559) | Acc: (70.32%) (23494/33408)\n",
+ "Batch_idx: 270 | Loss: (0.8561) | Acc: (70.31%) (24388/34688)\n",
+ "Batch_idx: 280 | Loss: (0.8549) | Acc: (70.33%) (25296/35968)\n",
+ "Batch_idx: 290 | Loss: (0.8550) | Acc: (70.33%) (26196/37248)\n",
+ "Batch_idx: 300 | Loss: (0.8547) | Acc: (70.30%) (27086/38528)\n",
+ "Batch_idx: 310 | Loss: (0.8532) | Acc: (70.34%) (27999/39808)\n",
+ "Batch_idx: 320 | Loss: (0.8523) | Acc: (70.33%) (28898/41088)\n",
+ "Batch_idx: 330 | Loss: (0.8514) | Acc: (70.33%) (29797/42368)\n",
+ "Batch_idx: 340 | Loss: (0.8503) | Acc: (70.36%) (30710/43648)\n",
+ "Batch_idx: 350 | Loss: (0.8509) | Acc: (70.32%) (31593/44928)\n",
+ "Batch_idx: 360 | Loss: (0.8484) | Acc: (70.39%) (32528/46208)\n",
+ "Batch_idx: 370 | Loss: (0.8489) | Acc: (70.36%) (33414/47488)\n",
+ "Batch_idx: 380 | Loss: (0.8478) | Acc: (70.39%) (34327/48768)\n",
+ "Batch_idx: 390 | Loss: (0.8482) | Acc: (70.38%) (35192/50000)\n",
+ "# TEST : Loss: (1.0747) | Acc: (67.93%) (6793/10000)\n",
+ "--------------- 12 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (0.8594) | Acc: (69.53%) (89/128)\n",
+ "Batch_idx: 10 | Loss: (0.7688) | Acc: (71.80%) (1011/1408)\n",
+ "Batch_idx: 20 | Loss: (0.8054) | Acc: (71.06%) (1910/2688)\n",
+ "Batch_idx: 30 | Loss: (0.7964) | Acc: (71.70%) (2845/3968)\n",
+ "Batch_idx: 40 | Loss: (0.7894) | Acc: (72.07%) (3782/5248)\n",
+ "Batch_idx: 50 | Loss: (0.7785) | Acc: (72.41%) (4727/6528)\n",
+ "Batch_idx: 60 | Loss: (0.7763) | Acc: (72.61%) (5669/7808)\n",
+ "Batch_idx: 70 | Loss: (0.7735) | Acc: (72.79%) (6615/9088)\n",
+ "Batch_idx: 80 | Loss: (0.7671) | Acc: (73.21%) (7590/10368)\n",
+ "Batch_idx: 90 | Loss: (0.7646) | Acc: (73.31%) (8539/11648)\n",
+ "Batch_idx: 100 | Loss: (0.7618) | Acc: (73.43%) (9493/12928)\n",
+ "Batch_idx: 110 | Loss: (0.7643) | Acc: (73.28%) (10411/14208)\n",
+ "Batch_idx: 120 | Loss: (0.7674) | Acc: (73.27%) (11348/15488)\n",
+ "Batch_idx: 130 | Loss: (0.7704) | Acc: (73.21%) (12276/16768)\n",
+ "Batch_idx: 140 | Loss: (0.7686) | Acc: (73.28%) (13225/18048)\n",
+ "Batch_idx: 150 | Loss: (0.7659) | Acc: (73.31%) (14169/19328)\n",
+ "Batch_idx: 160 | Loss: (0.7676) | Acc: (73.23%) (15091/20608)\n",
+ "Batch_idx: 170 | Loss: (0.7668) | Acc: (73.21%) (16025/21888)\n",
+ "Batch_idx: 180 | Loss: (0.7635) | Acc: (73.31%) (16984/23168)\n",
+ "Batch_idx: 190 | Loss: (0.7622) | Acc: (73.36%) (17935/24448)\n",
+ "Batch_idx: 200 | Loss: (0.7623) | Acc: (73.33%) (18866/25728)\n",
+ "Batch_idx: 210 | Loss: (0.7611) | Acc: (73.42%) (19830/27008)\n",
+ "Batch_idx: 220 | Loss: (0.7609) | Acc: (73.44%) (20775/28288)\n",
+ "Batch_idx: 230 | Loss: (0.7603) | Acc: (73.40%) (21704/29568)\n",
+ "Batch_idx: 240 | Loss: (0.7603) | Acc: (73.45%) (22657/30848)\n",
+ "Batch_idx: 250 | Loss: (0.7603) | Acc: (73.43%) (23591/32128)\n",
+ "Batch_idx: 260 | Loss: (0.7609) | Acc: (73.41%) (24525/33408)\n",
+ "Batch_idx: 270 | Loss: (0.7615) | Acc: (73.40%) (25461/34688)\n",
+ "Batch_idx: 280 | Loss: (0.7619) | Acc: (73.39%) (26397/35968)\n",
+ "Batch_idx: 290 | Loss: (0.7623) | Acc: (73.37%) (27330/37248)\n",
+ "Batch_idx: 300 | Loss: (0.7626) | Acc: (73.35%) (28262/38528)\n",
+ "Batch_idx: 310 | Loss: (0.7629) | Acc: (73.36%) (29202/39808)\n",
+ "Batch_idx: 320 | Loss: (0.7633) | Acc: (73.34%) (30132/41088)\n",
+ "Batch_idx: 330 | Loss: (0.7624) | Acc: (73.36%) (31080/42368)\n",
+ "Batch_idx: 340 | Loss: (0.7615) | Acc: (73.37%) (32024/43648)\n",
+ "Batch_idx: 350 | Loss: (0.7607) | Acc: (73.42%) (32986/44928)\n",
+ "Batch_idx: 360 | Loss: (0.7598) | Acc: (73.43%) (33932/46208)\n",
+ "Batch_idx: 370 | Loss: (0.7585) | Acc: (73.50%) (34904/47488)\n",
+ "Batch_idx: 380 | Loss: (0.7582) | Acc: (73.53%) (35857/48768)\n",
+ "Batch_idx: 390 | Loss: (0.7582) | Acc: (73.50%) (36750/50000)\n",
+ "# TEST : Loss: (0.8912) | Acc: (69.07%) (6907/10000)\n",
+ "--------------- 13 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (0.6083) | Acc: (82.03%) (105/128)\n",
+ "Batch_idx: 10 | Loss: (0.6617) | Acc: (75.99%) (1070/1408)\n",
+ "Batch_idx: 20 | Loss: (0.6558) | Acc: (76.38%) (2053/2688)\n",
+ "Batch_idx: 30 | Loss: (0.6586) | Acc: (76.08%) (3019/3968)\n",
+ "Batch_idx: 40 | Loss: (0.6576) | Acc: (76.37%) (4008/5248)\n",
+ "Batch_idx: 50 | Loss: (0.6695) | Acc: (75.81%) (4949/6528)\n",
+ "Batch_idx: 60 | Loss: (0.6710) | Acc: (75.83%) (5921/7808)\n",
+ "Batch_idx: 70 | Loss: (0.6722) | Acc: (75.89%) (6897/9088)\n",
+ "Batch_idx: 80 | Loss: (0.6732) | Acc: (75.91%) (7870/10368)\n",
+ "Batch_idx: 90 | Loss: (0.6759) | Acc: (75.94%) (8846/11648)\n",
+ "Batch_idx: 100 | Loss: (0.6776) | Acc: (75.78%) (9797/12928)\n",
+ "Batch_idx: 110 | Loss: (0.6784) | Acc: (75.69%) (10754/14208)\n",
+ "Batch_idx: 120 | Loss: (0.6776) | Acc: (75.86%) (11749/15488)\n",
+ "Batch_idx: 130 | Loss: (0.6762) | Acc: (76.00%) (12743/16768)\n",
+ "Batch_idx: 140 | Loss: (0.6750) | Acc: (76.00%) (13716/18048)\n",
+ "Batch_idx: 150 | Loss: (0.6757) | Acc: (75.95%) (14680/19328)\n",
+ "Batch_idx: 160 | Loss: (0.6729) | Acc: (76.07%) (15677/20608)\n",
+ "Batch_idx: 170 | Loss: (0.6701) | Acc: (76.23%) (16686/21888)\n",
+ "Batch_idx: 180 | Loss: (0.6719) | Acc: (76.15%) (17643/23168)\n",
+ "Batch_idx: 190 | Loss: (0.6737) | Acc: (76.07%) (18597/24448)\n",
+ "Batch_idx: 200 | Loss: (0.6753) | Acc: (76.00%) (19552/25728)\n",
+ "Batch_idx: 210 | Loss: (0.6732) | Acc: (76.03%) (20533/27008)\n",
+ "Batch_idx: 220 | Loss: (0.6740) | Acc: (76.08%) (21522/28288)\n",
+ "Batch_idx: 230 | Loss: (0.6744) | Acc: (76.05%) (22487/29568)\n",
+ "Batch_idx: 240 | Loss: (0.6760) | Acc: (75.98%) (23439/30848)\n",
+ "Batch_idx: 250 | Loss: (0.6746) | Acc: (76.01%) (24419/32128)\n",
+ "Batch_idx: 260 | Loss: (0.6766) | Acc: (75.93%) (25368/33408)\n",
+ "Batch_idx: 270 | Loss: (0.6762) | Acc: (75.99%) (26359/34688)\n",
+ "Batch_idx: 280 | Loss: (0.6772) | Acc: (75.96%) (27323/35968)\n",
+ "Batch_idx: 290 | Loss: (0.6779) | Acc: (75.99%) (28306/37248)\n",
+ "Batch_idx: 300 | Loss: (0.6787) | Acc: (75.96%) (29264/38528)\n",
+ "Batch_idx: 310 | Loss: (0.6784) | Acc: (75.95%) (30235/39808)\n",
+ "Batch_idx: 320 | Loss: (0.6776) | Acc: (75.98%) (31219/41088)\n",
+ "Batch_idx: 330 | Loss: (0.6777) | Acc: (75.99%) (32197/42368)\n",
+ "Batch_idx: 340 | Loss: (0.6782) | Acc: (76.01%) (33175/43648)\n",
+ "Batch_idx: 350 | Loss: (0.6772) | Acc: (76.03%) (34159/44928)\n",
+ "Batch_idx: 360 | Loss: (0.6774) | Acc: (76.01%) (35125/46208)\n",
+ "Batch_idx: 370 | Loss: (0.6767) | Acc: (76.03%) (36106/47488)\n",
+ "Batch_idx: 380 | Loss: (0.6767) | Acc: (76.05%) (37090/48768)\n",
+ "Batch_idx: 390 | Loss: (0.6763) | Acc: (76.09%) (38045/50000)\n",
+ "# TEST : Loss: (0.8689) | Acc: (70.06%) (7006/10000)\n",
+ "--------------- 14 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (0.5542) | Acc: (81.25%) (104/128)\n",
+ "Batch_idx: 10 | Loss: (0.5555) | Acc: (81.39%) (1146/1408)\n",
+ "Batch_idx: 20 | Loss: (0.5758) | Acc: (80.28%) (2158/2688)\n",
+ "Batch_idx: 30 | Loss: (0.5820) | Acc: (79.89%) (3170/3968)\n",
+ "Batch_idx: 40 | Loss: (0.5875) | Acc: (79.78%) (4187/5248)\n",
+ "Batch_idx: 50 | Loss: (0.5797) | Acc: (79.78%) (5208/6528)\n",
+ "Batch_idx: 60 | Loss: (0.5729) | Acc: (79.85%) (6235/7808)\n",
+ "Batch_idx: 70 | Loss: (0.5798) | Acc: (79.62%) (7236/9088)\n",
+ "Batch_idx: 80 | Loss: (0.5787) | Acc: (79.62%) (8255/10368)\n",
+ "Batch_idx: 90 | Loss: (0.5787) | Acc: (79.66%) (9279/11648)\n",
+ "Batch_idx: 100 | Loss: (0.5784) | Acc: (79.63%) (10295/12928)\n",
+ "Batch_idx: 110 | Loss: (0.5790) | Acc: (79.50%) (11295/14208)\n",
+ "Batch_idx: 120 | Loss: (0.5816) | Acc: (79.38%) (12295/15488)\n",
+ "Batch_idx: 130 | Loss: (0.5879) | Acc: (79.17%) (13275/16768)\n",
+ "Batch_idx: 140 | Loss: (0.5887) | Acc: (79.12%) (14280/18048)\n",
+ "Batch_idx: 150 | Loss: (0.5899) | Acc: (79.15%) (15299/19328)\n",
+ "Batch_idx: 160 | Loss: (0.5924) | Acc: (79.10%) (16300/20608)\n",
+ "Batch_idx: 170 | Loss: (0.5920) | Acc: (79.13%) (17319/21888)\n",
+ "Batch_idx: 180 | Loss: (0.5929) | Acc: (79.09%) (18323/23168)\n",
+ "Batch_idx: 190 | Loss: (0.5957) | Acc: (78.93%) (19297/24448)\n",
+ "Batch_idx: 200 | Loss: (0.5984) | Acc: (78.87%) (20291/25728)\n",
+ "Batch_idx: 210 | Loss: (0.6007) | Acc: (78.82%) (21288/27008)\n",
+ "Batch_idx: 220 | Loss: (0.6007) | Acc: (78.80%) (22291/28288)\n",
+ "Batch_idx: 230 | Loss: (0.5990) | Acc: (78.84%) (23311/29568)\n",
+ "Batch_idx: 240 | Loss: (0.5988) | Acc: (78.81%) (24310/30848)\n",
+ "Batch_idx: 250 | Loss: (0.5987) | Acc: (78.84%) (25330/32128)\n",
+ "Batch_idx: 260 | Loss: (0.5996) | Acc: (78.81%) (26329/33408)\n",
+ "Batch_idx: 270 | Loss: (0.6003) | Acc: (78.79%) (27330/34688)\n",
+ "Batch_idx: 280 | Loss: (0.6034) | Acc: (78.71%) (28309/35968)\n",
+ "Batch_idx: 290 | Loss: (0.6038) | Acc: (78.67%) (29304/37248)\n",
+ "Batch_idx: 300 | Loss: (0.6057) | Acc: (78.59%) (30278/38528)\n",
+ "Batch_idx: 310 | Loss: (0.6085) | Acc: (78.50%) (31249/39808)\n",
+ "Batch_idx: 320 | Loss: (0.6085) | Acc: (78.48%) (32244/41088)\n",
+ "Batch_idx: 330 | Loss: (0.6100) | Acc: (78.41%) (33222/42368)\n",
+ "Batch_idx: 340 | Loss: (0.6120) | Acc: (78.33%) (34189/43648)\n",
+ "Batch_idx: 350 | Loss: (0.6132) | Acc: (78.30%) (35179/44928)\n",
+ "Batch_idx: 360 | Loss: (0.6137) | Acc: (78.28%) (36173/46208)\n",
+ "Batch_idx: 370 | Loss: (0.6134) | Acc: (78.27%) (37170/47488)\n",
+ "Batch_idx: 380 | Loss: (0.6147) | Acc: (78.25%) (38159/48768)\n",
+ "Batch_idx: 390 | Loss: (0.6157) | Acc: (78.20%) (39101/50000)\n",
+ "# TEST : Loss: (0.9834) | Acc: (69.63%) (6963/10000)\n",
+ "--------------- 15 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (0.5775) | Acc: (81.25%) (104/128)\n",
+ "Batch_idx: 10 | Loss: (0.5257) | Acc: (82.10%) (1156/1408)\n",
+ "Batch_idx: 20 | Loss: (0.5135) | Acc: (81.92%) (2202/2688)\n",
+ "Batch_idx: 30 | Loss: (0.5216) | Acc: (81.68%) (3241/3968)\n",
+ "Batch_idx: 40 | Loss: (0.5251) | Acc: (81.71%) (4288/5248)\n",
+ "Batch_idx: 50 | Loss: (0.5188) | Acc: (81.71%) (5334/6528)\n",
+ "Batch_idx: 60 | Loss: (0.5295) | Acc: (81.22%) (6342/7808)\n",
+ "Batch_idx: 70 | Loss: (0.5247) | Acc: (81.43%) (7400/9088)\n",
+ "Batch_idx: 80 | Loss: (0.5220) | Acc: (81.53%) (8453/10368)\n",
+ "Batch_idx: 90 | Loss: (0.5202) | Acc: (81.52%) (9496/11648)\n",
+ "Batch_idx: 100 | Loss: (0.5187) | Acc: (81.44%) (10529/12928)\n",
+ "Batch_idx: 110 | Loss: (0.5223) | Acc: (81.30%) (11551/14208)\n",
+ "Batch_idx: 120 | Loss: (0.5244) | Acc: (81.28%) (12589/15488)\n",
+ "Batch_idx: 130 | Loss: (0.5240) | Acc: (81.36%) (13643/16768)\n",
+ "Batch_idx: 140 | Loss: (0.5260) | Acc: (81.34%) (14680/18048)\n",
+ "Batch_idx: 150 | Loss: (0.5293) | Acc: (81.24%) (15703/19328)\n",
+ "Batch_idx: 160 | Loss: (0.5310) | Acc: (81.14%) (16721/20608)\n",
+ "Batch_idx: 170 | Loss: (0.5307) | Acc: (81.18%) (17768/21888)\n",
+ "Batch_idx: 180 | Loss: (0.5319) | Acc: (81.06%) (18779/23168)\n",
+ "Batch_idx: 190 | Loss: (0.5328) | Acc: (81.00%) (19804/24448)\n",
+ "Batch_idx: 200 | Loss: (0.5351) | Acc: (80.93%) (20822/25728)\n",
+ "Batch_idx: 210 | Loss: (0.5360) | Acc: (80.86%) (21839/27008)\n",
+ "Batch_idx: 220 | Loss: (0.5390) | Acc: (80.78%) (22851/28288)\n",
+ "Batch_idx: 230 | Loss: (0.5400) | Acc: (80.75%) (23876/29568)\n",
+ "Batch_idx: 240 | Loss: (0.5397) | Acc: (80.72%) (24901/30848)\n",
+ "Batch_idx: 250 | Loss: (0.5430) | Acc: (80.62%) (25903/32128)\n",
+ "Batch_idx: 260 | Loss: (0.5429) | Acc: (80.64%) (26939/33408)\n",
+ "Batch_idx: 270 | Loss: (0.5446) | Acc: (80.60%) (27958/34688)\n",
+ "Batch_idx: 280 | Loss: (0.5441) | Acc: (80.58%) (28983/35968)\n",
+ "Batch_idx: 290 | Loss: (0.5460) | Acc: (80.51%) (29990/37248)\n",
+ "Batch_idx: 300 | Loss: (0.5481) | Acc: (80.45%) (30996/38528)\n",
+ "Batch_idx: 310 | Loss: (0.5490) | Acc: (80.47%) (32032/39808)\n",
+ "Batch_idx: 320 | Loss: (0.5483) | Acc: (80.50%) (33077/41088)\n",
+ "Batch_idx: 330 | Loss: (0.5494) | Acc: (80.45%) (34086/42368)\n",
+ "Batch_idx: 340 | Loss: (0.5500) | Acc: (80.45%) (35116/43648)\n",
+ "Batch_idx: 350 | Loss: (0.5511) | Acc: (80.40%) (36121/44928)\n",
+ "Batch_idx: 360 | Loss: (0.5520) | Acc: (80.35%) (37128/46208)\n",
+ "Batch_idx: 370 | Loss: (0.5544) | Acc: (80.30%) (38135/47488)\n",
+ "Batch_idx: 380 | Loss: (0.5549) | Acc: (80.31%) (39167/48768)\n",
+ "Batch_idx: 390 | Loss: (0.5558) | Acc: (80.29%) (40145/50000)\n",
+ "# TEST : Loss: (1.1882) | Acc: (69.23%) (6923/10000)\n",
+ "--------------- 16 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (0.5062) | Acc: (82.81%) (106/128)\n",
+ "Batch_idx: 10 | Loss: (0.4446) | Acc: (83.66%) (1178/1408)\n",
+ "Batch_idx: 20 | Loss: (0.4543) | Acc: (83.37%) (2241/2688)\n",
+ "Batch_idx: 30 | Loss: (0.4461) | Acc: (83.69%) (3321/3968)\n",
+ "Batch_idx: 40 | Loss: (0.4422) | Acc: (83.77%) (4396/5248)\n",
+ "Batch_idx: 50 | Loss: (0.4381) | Acc: (84.05%) (5487/6528)\n",
+ "Batch_idx: 60 | Loss: (0.4436) | Acc: (83.89%) (6550/7808)\n",
+ "Batch_idx: 70 | Loss: (0.4495) | Acc: (83.76%) (7612/9088)\n",
+ "Batch_idx: 80 | Loss: (0.4483) | Acc: (83.80%) (8688/10368)\n",
+ "Batch_idx: 90 | Loss: (0.4494) | Acc: (83.81%) (9762/11648)\n",
+ "Batch_idx: 100 | Loss: (0.4580) | Acc: (83.56%) (10803/12928)\n",
+ "Batch_idx: 110 | Loss: (0.4598) | Acc: (83.49%) (11862/14208)\n",
+ "Batch_idx: 120 | Loss: (0.4606) | Acc: (83.47%) (12928/15488)\n",
+ "Batch_idx: 130 | Loss: (0.4613) | Acc: (83.43%) (13989/16768)\n",
+ "Batch_idx: 140 | Loss: (0.4620) | Acc: (83.42%) (15055/18048)\n",
+ "Batch_idx: 150 | Loss: (0.4661) | Acc: (83.21%) (16083/19328)\n",
+ "Batch_idx: 160 | Loss: (0.4679) | Acc: (83.14%) (17134/20608)\n",
+ "Batch_idx: 170 | Loss: (0.4695) | Acc: (83.06%) (18180/21888)\n",
+ "Batch_idx: 180 | Loss: (0.4676) | Acc: (83.14%) (19261/23168)\n",
+ "Batch_idx: 190 | Loss: (0.4691) | Acc: (83.09%) (20314/24448)\n",
+ "Batch_idx: 200 | Loss: (0.4709) | Acc: (83.05%) (21368/25728)\n",
+ "Batch_idx: 210 | Loss: (0.4731) | Acc: (83.01%) (22419/27008)\n",
+ "Batch_idx: 220 | Loss: (0.4742) | Acc: (82.95%) (23466/28288)\n",
+ "Batch_idx: 230 | Loss: (0.4762) | Acc: (82.90%) (24512/29568)\n",
+ "Batch_idx: 240 | Loss: (0.4792) | Acc: (82.81%) (25546/30848)\n",
+ "Batch_idx: 250 | Loss: (0.4795) | Acc: (82.79%) (26598/32128)\n",
+ "Batch_idx: 260 | Loss: (0.4792) | Acc: (82.82%) (27668/33408)\n",
+ "Batch_idx: 270 | Loss: (0.4806) | Acc: (82.77%) (28710/34688)\n",
+ "Batch_idx: 280 | Loss: (0.4818) | Acc: (82.78%) (29776/35968)\n",
+ "Batch_idx: 290 | Loss: (0.4829) | Acc: (82.73%) (30817/37248)\n",
+ "Batch_idx: 300 | Loss: (0.4833) | Acc: (82.74%) (31879/38528)\n",
+ "Batch_idx: 310 | Loss: (0.4835) | Acc: (82.72%) (32931/39808)\n",
+ "Batch_idx: 320 | Loss: (0.4834) | Acc: (82.75%) (34001/41088)\n",
+ "Batch_idx: 330 | Loss: (0.4840) | Acc: (82.77%) (35069/42368)\n",
+ "Batch_idx: 340 | Loss: (0.4857) | Acc: (82.70%) (36098/43648)\n",
+ "Batch_idx: 350 | Loss: (0.4869) | Acc: (82.68%) (37146/44928)\n",
+ "Batch_idx: 360 | Loss: (0.4871) | Acc: (82.69%) (38209/46208)\n",
+ "Batch_idx: 370 | Loss: (0.4876) | Acc: (82.68%) (39264/47488)\n",
+ "Batch_idx: 380 | Loss: (0.4893) | Acc: (82.62%) (40291/48768)\n",
+ "Batch_idx: 390 | Loss: (0.4909) | Acc: (82.56%) (41282/50000)\n",
+ "# TEST : Loss: (1.0262) | Acc: (69.02%) (6902/10000)\n",
+ "--------------- 17 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (0.3814) | Acc: (85.16%) (109/128)\n",
+ "Batch_idx: 10 | Loss: (0.4067) | Acc: (86.15%) (1213/1408)\n",
+ "Batch_idx: 20 | Loss: (0.4004) | Acc: (86.31%) (2320/2688)\n",
+ "Batch_idx: 30 | Loss: (0.3917) | Acc: (86.62%) (3437/3968)\n",
+ "Batch_idx: 40 | Loss: (0.4006) | Acc: (86.22%) (4525/5248)\n",
+ "Batch_idx: 50 | Loss: (0.3992) | Acc: (86.32%) (5635/6528)\n",
+ "Batch_idx: 60 | Loss: (0.3954) | Acc: (86.42%) (6748/7808)\n",
+ "Batch_idx: 70 | Loss: (0.3929) | Acc: (86.41%) (7853/9088)\n",
+ "Batch_idx: 80 | Loss: (0.3966) | Acc: (86.15%) (8932/10368)\n",
+ "Batch_idx: 90 | Loss: (0.4032) | Acc: (85.79%) (9993/11648)\n",
+ "Batch_idx: 100 | Loss: (0.4082) | Acc: (85.59%) (11065/12928)\n",
+ "Batch_idx: 110 | Loss: (0.4119) | Acc: (85.43%) (12138/14208)\n",
+ "Batch_idx: 120 | Loss: (0.4222) | Acc: (85.10%) (13181/15488)\n",
+ "Batch_idx: 130 | Loss: (0.4229) | Acc: (85.13%) (14274/16768)\n",
+ "Batch_idx: 140 | Loss: (0.4257) | Acc: (85.00%) (15341/18048)\n",
+ "Batch_idx: 150 | Loss: (0.4291) | Acc: (84.84%) (16397/19328)\n",
+ "Batch_idx: 160 | Loss: (0.4286) | Acc: (84.82%) (17480/20608)\n",
+ "Batch_idx: 170 | Loss: (0.4291) | Acc: (84.81%) (18563/21888)\n",
+ "Batch_idx: 180 | Loss: (0.4316) | Acc: (84.74%) (19632/23168)\n",
+ "Batch_idx: 190 | Loss: (0.4345) | Acc: (84.67%) (20700/24448)\n",
+ "Batch_idx: 200 | Loss: (0.4365) | Acc: (84.54%) (21750/25728)\n",
+ "Batch_idx: 210 | Loss: (0.4369) | Acc: (84.57%) (22840/27008)\n",
+ "Batch_idx: 220 | Loss: (0.4379) | Acc: (84.55%) (23918/28288)\n",
+ "Batch_idx: 230 | Loss: (0.4376) | Acc: (84.53%) (24994/29568)\n",
+ "Batch_idx: 240 | Loss: (0.4402) | Acc: (84.47%) (26057/30848)\n",
+ "Batch_idx: 250 | Loss: (0.4417) | Acc: (84.41%) (27118/32128)\n",
+ "Batch_idx: 260 | Loss: (0.4429) | Acc: (84.34%) (28177/33408)\n",
+ "Batch_idx: 270 | Loss: (0.4422) | Acc: (84.38%) (29269/34688)\n",
+ "Batch_idx: 280 | Loss: (0.4421) | Acc: (84.41%) (30361/35968)\n",
+ "Batch_idx: 290 | Loss: (0.4436) | Acc: (84.37%) (31426/37248)\n",
+ "Batch_idx: 300 | Loss: (0.4464) | Acc: (84.23%) (32453/38528)\n",
+ "Batch_idx: 310 | Loss: (0.4477) | Acc: (84.18%) (33512/39808)\n",
+ "Batch_idx: 320 | Loss: (0.4482) | Acc: (84.18%) (34586/41088)\n",
+ "Batch_idx: 330 | Loss: (0.4489) | Acc: (84.14%) (35647/42368)\n",
+ "Batch_idx: 340 | Loss: (0.4493) | Acc: (84.13%) (36722/43648)\n",
+ "Batch_idx: 350 | Loss: (0.4495) | Acc: (84.09%) (37778/44928)\n",
+ "Batch_idx: 360 | Loss: (0.4506) | Acc: (84.05%) (38839/46208)\n",
+ "Batch_idx: 370 | Loss: (0.4520) | Acc: (84.02%) (39900/47488)\n",
+ "Batch_idx: 380 | Loss: (0.4547) | Acc: (83.91%) (40922/48768)\n",
+ "Batch_idx: 390 | Loss: (0.4555) | Acc: (83.89%) (41945/50000)\n",
+ "# TEST : Loss: (0.9655) | Acc: (69.83%) (6983/10000)\n",
+ "--------------- 18 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (0.2352) | Acc: (94.53%) (121/128)\n",
+ "Batch_idx: 10 | Loss: (0.3402) | Acc: (88.42%) (1245/1408)\n",
+ "Batch_idx: 20 | Loss: (0.3429) | Acc: (88.47%) (2378/2688)\n",
+ "Batch_idx: 30 | Loss: (0.3496) | Acc: (88.03%) (3493/3968)\n",
+ "Batch_idx: 40 | Loss: (0.3527) | Acc: (87.98%) (4617/5248)\n",
+ "Batch_idx: 50 | Loss: (0.3468) | Acc: (87.81%) (5732/6528)\n",
+ "Batch_idx: 60 | Loss: (0.3481) | Acc: (87.78%) (6854/7808)\n",
+ "Batch_idx: 70 | Loss: (0.3489) | Acc: (87.74%) (7974/9088)\n",
+ "Batch_idx: 80 | Loss: (0.3476) | Acc: (87.96%) (9120/10368)\n",
+ "Batch_idx: 90 | Loss: (0.3473) | Acc: (87.92%) (10241/11648)\n",
+ "Batch_idx: 100 | Loss: (0.3509) | Acc: (87.69%) (11337/12928)\n",
+ "Batch_idx: 110 | Loss: (0.3535) | Acc: (87.52%) (12435/14208)\n",
+ "Batch_idx: 120 | Loss: (0.3546) | Acc: (87.56%) (13562/15488)\n",
+ "Batch_idx: 130 | Loss: (0.3567) | Acc: (87.49%) (14670/16768)\n",
+ "Batch_idx: 140 | Loss: (0.3595) | Acc: (87.41%) (15776/18048)\n",
+ "Batch_idx: 150 | Loss: (0.3628) | Acc: (87.23%) (16860/19328)\n",
+ "Batch_idx: 160 | Loss: (0.3635) | Acc: (87.20%) (17970/20608)\n",
+ "Batch_idx: 170 | Loss: (0.3636) | Acc: (87.17%) (19079/21888)\n",
+ "Batch_idx: 180 | Loss: (0.3636) | Acc: (87.18%) (20197/23168)\n",
+ "Batch_idx: 190 | Loss: (0.3677) | Acc: (87.00%) (21269/24448)\n",
+ "Batch_idx: 200 | Loss: (0.3696) | Acc: (86.91%) (22361/25728)\n",
+ "Batch_idx: 210 | Loss: (0.3704) | Acc: (86.91%) (23473/27008)\n",
+ "Batch_idx: 220 | Loss: (0.3728) | Acc: (86.82%) (24561/28288)\n",
+ "Batch_idx: 230 | Loss: (0.3736) | Acc: (86.75%) (25649/29568)\n",
+ "Batch_idx: 240 | Loss: (0.3743) | Acc: (86.71%) (26747/30848)\n",
+ "Batch_idx: 250 | Loss: (0.3759) | Acc: (86.63%) (27832/32128)\n",
+ "Batch_idx: 260 | Loss: (0.3785) | Acc: (86.54%) (28911/33408)\n",
+ "Batch_idx: 270 | Loss: (0.3791) | Acc: (86.52%) (30013/34688)\n",
+ "Batch_idx: 280 | Loss: (0.3784) | Acc: (86.52%) (31118/35968)\n",
+ "Batch_idx: 290 | Loss: (0.3778) | Acc: (86.53%) (32231/37248)\n",
+ "Batch_idx: 300 | Loss: (0.3804) | Acc: (86.40%) (33290/38528)\n",
+ "Batch_idx: 310 | Loss: (0.3796) | Acc: (86.44%) (34411/39808)\n",
+ "Batch_idx: 320 | Loss: (0.3811) | Acc: (86.40%) (35499/41088)\n",
+ "Batch_idx: 330 | Loss: (0.3823) | Acc: (86.34%) (36581/42368)\n",
+ "Batch_idx: 340 | Loss: (0.3839) | Acc: (86.29%) (37665/43648)\n",
+ "Batch_idx: 350 | Loss: (0.3846) | Acc: (86.28%) (38764/44928)\n",
+ "Batch_idx: 360 | Loss: (0.3857) | Acc: (86.25%) (39855/46208)\n",
+ "Batch_idx: 370 | Loss: (0.3868) | Acc: (86.20%) (40936/47488)\n",
+ "Batch_idx: 380 | Loss: (0.3871) | Acc: (86.18%) (42029/48768)\n",
+ "Batch_idx: 390 | Loss: (0.3882) | Acc: (86.14%) (43069/50000)\n",
+ "# TEST : Loss: (1.0966) | Acc: (69.37%) (6937/10000)\n",
+ "--------------- 19 Epoch ---------------\n",
+ "Batch_idx: 0 | Loss: (0.2576) | Acc: (92.97%) (119/128)\n",
+ "Batch_idx: 10 | Loss: (0.3043) | Acc: (89.84%) (1265/1408)\n",
+ "Batch_idx: 20 | Loss: (0.2996) | Acc: (89.73%) (2412/2688)\n",
+ "Batch_idx: 30 | Loss: (0.2934) | Acc: (89.97%) (3570/3968)\n",
+ "Batch_idx: 40 | Loss: (0.2968) | Acc: (89.79%) (4712/5248)\n",
+ "Batch_idx: 50 | Loss: (0.2978) | Acc: (89.75%) (5859/6528)\n",
+ "Batch_idx: 60 | Loss: (0.2975) | Acc: (89.74%) (7007/7808)\n",
+ "Batch_idx: 70 | Loss: (0.3019) | Acc: (89.56%) (8139/9088)\n",
+ "Batch_idx: 80 | Loss: (0.2970) | Acc: (89.65%) (9295/10368)\n",
+ "Batch_idx: 90 | Loss: (0.2952) | Acc: (89.54%) (10430/11648)\n",
+ "Batch_idx: 100 | Loss: (0.2972) | Acc: (89.31%) (11546/12928)\n",
+ "Batch_idx: 110 | Loss: (0.2971) | Acc: (89.28%) (12685/14208)\n",
+ "Batch_idx: 120 | Loss: (0.2993) | Acc: (89.17%) (13810/15488)\n",
+ "Batch_idx: 130 | Loss: (0.3027) | Acc: (89.09%) (14939/16768)\n",
+ "Batch_idx: 140 | Loss: (0.3043) | Acc: (89.03%) (16068/18048)\n",
+ "Batch_idx: 150 | Loss: (0.3074) | Acc: (88.93%) (17189/19328)\n",
+ "Batch_idx: 160 | Loss: (0.3112) | Acc: (88.80%) (18299/20608)\n",
+ "Batch_idx: 170 | Loss: (0.3134) | Acc: (88.69%) (19413/21888)\n",
+ "Batch_idx: 180 | Loss: (0.3144) | Acc: (88.64%) (20535/23168)\n",
+ "Batch_idx: 190 | Loss: (0.3161) | Acc: (88.61%) (21663/24448)\n",
+ "Batch_idx: 200 | Loss: (0.3181) | Acc: (88.55%) (22781/25728)\n",
+ "Batch_idx: 210 | Loss: (0.3198) | Acc: (88.48%) (23897/27008)\n",
+ "Batch_idx: 220 | Loss: (0.3213) | Acc: (88.45%) (25022/28288)\n",
+ "Batch_idx: 230 | Loss: (0.3250) | Acc: (88.33%) (26118/29568)\n",
+ "Batch_idx: 240 | Loss: (0.3262) | Acc: (88.28%) (27232/30848)\n",
+ "Batch_idx: 250 | Loss: (0.3258) | Acc: (88.28%) (28363/32128)\n",
+ "Batch_idx: 260 | Loss: (0.3277) | Acc: (88.21%) (29469/33408)\n",
+ "Batch_idx: 270 | Loss: (0.3279) | Acc: (88.17%) (30583/34688)\n",
+ "Batch_idx: 280 | Loss: (0.3296) | Acc: (88.13%) (31698/35968)\n",
+ "Batch_idx: 290 | Loss: (0.3322) | Acc: (88.03%) (32789/37248)\n",
+ "Batch_idx: 300 | Loss: (0.3341) | Acc: (88.00%) (33903/38528)\n",
+ "Batch_idx: 310 | Loss: (0.3341) | Acc: (88.01%) (35036/39808)\n",
+ "Batch_idx: 320 | Loss: (0.3354) | Acc: (87.99%) (36153/41088)\n",
+ "Batch_idx: 330 | Loss: (0.3367) | Acc: (87.95%) (37264/42368)\n",
+ "Batch_idx: 340 | Loss: (0.3379) | Acc: (87.92%) (38375/43648)\n",
+ "Batch_idx: 350 | Loss: (0.3389) | Acc: (87.88%) (39482/44928)\n",
+ "Batch_idx: 360 | Loss: (0.3384) | Acc: (87.92%) (40627/46208)\n",
+ "Batch_idx: 370 | Loss: (0.3399) | Acc: (87.86%) (41723/47488)\n",
+ "Batch_idx: 380 | Loss: (0.3407) | Acc: (87.84%) (42836/48768)\n",
+ "Batch_idx: 390 | Loss: (0.3418) | Acc: (87.79%) (43893/50000)\n",
+ "# TEST : Loss: (1.2919) | Acc: (69.13%) (6913/10000)\n"
+ ]
+ }
+ ],
+ "source": [
+ "import timm\n",
+ "import timm.optim\n",
+ "import torch\n",
+ "import torch.nn as nn\n",
+ "from torchvision import datasets, transforms\n",
+ "\n",
+ "batch_size = 128\n",
+ "learning_rate = 0.01\n",
+ "num_classes = 10\n",
+ "\n",
+ "transform_train = transforms.Compose([\n",
+ " transforms.ToTensor(),\n",
+ " transforms.Normalize(mean=(0.4914, 0.4824, 0.4467),\n",
+ " std=(0.2471, 0.2436, 0.2616))\n",
+ "])\n",
+ "\n",
+ "transform_test = transforms.Compose([\n",
+ " transforms.ToTensor(),\n",
+ " transforms.Normalize(mean=(0.4914, 0.4824, 0.4467),\n",
+ " std=(0.2471, 0.2436, 0.2616))\n",
+ "])\n",
+ "\n",
+ "\n",
+ "train_dataset = datasets.CIFAR10(root='./content/pytorch/data/cifar10/',\n",
+ " train=True,\n",
+ " transform=transform_train,\n",
+ " download=True)\n",
+ "\n",
+ "test_dataset = datasets.CIFAR10(root='./content/pytorch/data/cifar10/',\n",
+ " train=False,\n",
+ " transform=transform_test)\n",
+ "\n",
+ "train_loader = torch.utils.data.DataLoader(dataset=train_dataset,\n",
+ " batch_size=batch_size,\n",
+ " shuffle=True,\n",
+ " num_workers=2)\n",
+ "\n",
+ "test_loader = torch.utils.data.DataLoader(dataset=test_dataset,\n",
+ " batch_size=batch_size,\n",
+ " shuffle=False,\n",
+ " num_workers=2)\n",
+ "\n",
+ "model = timm.create_model(\n",
+ " 'efficientnet_b0', pretrained=True, num_classes=num_classes\n",
+ ").cuda()\n",
+ "optimizer = timm.optim.create_optimizer_v2(model, lr=learning_rate)\n",
+ "criterion = nn.CrossEntropyLoss().cuda()\n",
+ "\n",
+ "\n",
+ "def train():\n",
+ " model.train()\n",
+ " train_loss = 0\n",
+ " total = 0\n",
+ " correct = 0\n",
+ " for batch_idx, (data, target) in enumerate(train_loader):\n",
+ " \n",
+ " data, target = data.cuda(), target.cuda()\n",
+ " optimizer.zero_grad()\n",
+ " output = model(data)\n",
+ " loss = criterion(output, target)\n",
+ " loss.backward()\n",
+ " optimizer.step()\n",
+ "\n",
+ " train_loss += loss.item()\n",
+ " _, predicted = torch.max(output.data, 1)\n",
+ "\n",
+ " total += target.size(0)\n",
+ " correct += predicted.eq(target.data).cpu().sum()\n",
+ " if batch_idx % 10 == 0: \n",
+ " print('Batch_idx: {} | Loss: ({:.4f}) | Acc: ({:.2f}%) ({}/{})'\n",
+ " .format(batch_idx, train_loss/(batch_idx+1), 100.*correct/total, correct, total))\n",
+ "\n",
+ "def test():\n",
+ " model.eval()\n",
+ " test_loss = 0\n",
+ " correct = 0\n",
+ " total = 0\n",
+ " for batch_idx, (data, target) in enumerate(test_loader):\n",
+ " data, target = data.cuda(), target.cuda()\n",
+ "\n",
+ " outputs = model(data)\n",
+ " loss = criterion(outputs, target)\n",
+ "\n",
+ " test_loss += loss.item()\n",
+ " _, predicted = torch.max(outputs.data, 1)\n",
+ " total += target.size(0)\n",
+ " correct += predicted.eq(target.data).cpu().sum()\n",
+ " print('# TEST : Loss: ({:.4f}) | Acc: ({:.2f}%) ({}/{})'\n",
+ " .format(test_loss/(batch_idx+1), 100.*correct/total, correct, total))\n",
+ "\n",
+ "\n",
+ "for epoch in range(0, 20):\n",
+ " print('--------------- {} Epoch ---------------'.format(epoch))\n",
+ " train()\n",
+ " test()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "27d2b228",
+ "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": "3.9.12"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}