Model save path: ./New_Models/bn_False_dataset_MNIST_epochs_200_lr_0.01_model_type_vgg19_rand_seed_979323_test_samples_None_train_samples_None_weight_decay_0.005.pth.tar
Training Set:
Layer 0: Conv2d(1, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.09116754680871964
Inter Cos: 0.10967149585485458
Norm Quadratic Average: 23.567672729492188
Nearest Class Center Accuracy: 0.8079833333333334

Layer 1: Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.11452236771583557
Inter Cos: 0.1374807506799698
Norm Quadratic Average: 68.92739868164062
Nearest Class Center Accuracy: 0.8004

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.13600698113441467
Inter Cos: 0.17506898939609528
Norm Quadratic Average: 118.78282165527344
Nearest Class Center Accuracy: 0.7845833333333333

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.14282989501953125
Inter Cos: 0.1858660727739334
Norm Quadratic Average: 209.7696533203125
Nearest Class Center Accuracy: 0.7912833333333333

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.17462334036827087
Inter Cos: 0.19652627408504486
Norm Quadratic Average: 141.53770446777344
Nearest Class Center Accuracy: 0.83175

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.19965985417366028
Inter Cos: 0.2070666253566742
Norm Quadratic Average: 99.86284637451172
Nearest Class Center Accuracy: 0.8614666666666667

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2224782109260559
Inter Cos: 0.24274767935276031
Norm Quadratic Average: 86.70005798339844
Nearest Class Center Accuracy: 0.8814

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2668374180793762
Inter Cos: 0.269483745098114
Norm Quadratic Average: 78.72382354736328
Nearest Class Center Accuracy: 0.921

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.26190322637557983
Inter Cos: 0.2865966260433197
Norm Quadratic Average: 35.517093658447266
Nearest Class Center Accuracy: 0.9226

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.33800044655799866
Inter Cos: 0.36907273530960083
Norm Quadratic Average: 21.749784469604492
Nearest Class Center Accuracy: 0.9218666666666666

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.43185433745384216
Inter Cos: 0.3447472155094147
Norm Quadratic Average: 21.04486083984375
Nearest Class Center Accuracy: 0.9382

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5133815407752991
Inter Cos: 0.4051876366138458
Norm Quadratic Average: 24.034055709838867
Nearest Class Center Accuracy: 0.9534333333333334

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5691372156143188
Inter Cos: 0.41353797912597656
Norm Quadratic Average: 15.787888526916504
Nearest Class Center Accuracy: 0.926

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6649438142776489
Inter Cos: 0.5173119902610779
Norm Quadratic Average: 13.452836990356445
Nearest Class Center Accuracy: 0.9423666666666667

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7405070066452026
Inter Cos: 0.48078739643096924
Norm Quadratic Average: 14.621057510375977
Nearest Class Center Accuracy: 0.9641166666666666

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8041489124298096
Inter Cos: 0.4184056222438812
Norm Quadratic Average: 16.3846435546875
Nearest Class Center Accuracy: 0.9771

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.64596688747406
Linear Weight Rank: 44
Intra Cos: 0.8425329923629761
Inter Cos: 0.3195604979991913
Norm Quadratic Average: 70.33148956298828
Nearest Class Center Accuracy: 0.9856333333333334

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.6576188802719116
Linear Weight Rank: 2718
Intra Cos: 0.899458646774292
Inter Cos: 0.35105955600738525
Norm Quadratic Average: 46.710933685302734
Nearest Class Center Accuracy: 0.9944666666666667

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.6450575590133667
Linear Weight Rank: 9
Intra Cos: 0.9167450070381165
Inter Cos: 0.3063699007034302
Norm Quadratic Average: 28.894311904907227
Nearest Class Center Accuracy: 0.9971833333333333

Output Layer:
Intra Cos: 0.9415408968925476
Inter Cos: 0.37114861607551575
Norm Quadratic Average: 19.941667556762695
Nearest Class Center Accuracy: 0.9982666666666666

Test Set:
Average Loss: 0.03072947488948703
Accuracy: 0.9908
NC1 Within Class Collapse: 1.3057327270507812
NC2 Equinorm: Features: 0.08869028091430664, Weights: 0.04479397088289261
NC2 Equiangle: Features: 0.30440673828125, Weights: 0.21209693484836153
NC3 Self-Duality: 0.08377907425165176
NC4 NCC Mismatch: 0.0041999999999999815

Layer 0: Conv2d(1, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10219583660364151
Inter Cos: 0.12048852443695068
Norm Quadratic Average: 23.595197677612305
Nearest Class Center Accuracy: 0.8229

Layer 1: Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.12608806788921356
Inter Cos: 0.150632843375206
Norm Quadratic Average: 69.32929229736328
Nearest Class Center Accuracy: 0.8174

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1502884179353714
Inter Cos: 0.191548153758049
Norm Quadratic Average: 119.31155395507812
Nearest Class Center Accuracy: 0.8043

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1573646068572998
Inter Cos: 0.20346088707447052
Norm Quadratic Average: 210.70822143554688
Nearest Class Center Accuracy: 0.8099

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1865379959344864
Inter Cos: 0.21568907797336578
Norm Quadratic Average: 141.8251495361328
Nearest Class Center Accuracy: 0.847

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.21267347037792206
Inter Cos: 0.22693243622779846
Norm Quadratic Average: 100.04663848876953
Nearest Class Center Accuracy: 0.8749

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.23470020294189453
Inter Cos: 0.2530616819858551
Norm Quadratic Average: 86.82418060302734
Nearest Class Center Accuracy: 0.8966

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.28057077527046204
Inter Cos: 0.2642967402935028
Norm Quadratic Average: 79.03205108642578
Nearest Class Center Accuracy: 0.9287

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.27667978405952454
Inter Cos: 0.3014664649963379
Norm Quadratic Average: 35.706756591796875
Nearest Class Center Accuracy: 0.9287

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3515099883079529
Inter Cos: 0.3988305628299713
Norm Quadratic Average: 21.919546127319336
Nearest Class Center Accuracy: 0.9264

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.44366517663002014
Inter Cos: 0.36628496646881104
Norm Quadratic Average: 21.258243560791016
Nearest Class Center Accuracy: 0.9413

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5252279043197632
Inter Cos: 0.4225367307662964
Norm Quadratic Average: 24.309490203857422
Nearest Class Center Accuracy: 0.9562

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5862675905227661
Inter Cos: 0.42326784133911133
Norm Quadratic Average: 15.920454025268555
Nearest Class Center Accuracy: 0.9323

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6863853335380554
Inter Cos: 0.5239025354385376
Norm Quadratic Average: 13.570671081542969
Nearest Class Center Accuracy: 0.9444

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7599428296089172
Inter Cos: 0.4838774800300598
Norm Quadratic Average: 14.771618843078613
Nearest Class Center Accuracy: 0.9613

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8087093830108643
Inter Cos: 0.41870731115341187
Norm Quadratic Average: 16.575109481811523
Nearest Class Center Accuracy: 0.9713

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.64596688747406
Linear Weight Rank: 44
Intra Cos: 0.846228301525116
Inter Cos: 0.32450738549232483
Norm Quadratic Average: 71.16831970214844
Nearest Class Center Accuracy: 0.9784

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.6576188802719116
Linear Weight Rank: 2718
Intra Cos: 0.9004340767860413
Inter Cos: 0.3581888973712921
Norm Quadratic Average: 47.3029899597168
Nearest Class Center Accuracy: 0.986

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.6450575590133667
Linear Weight Rank: 9
Intra Cos: 0.9129122495651245
Inter Cos: 0.3137280344963074
Norm Quadratic Average: 29.26346778869629
Nearest Class Center Accuracy: 0.9897

Output Layer:
Intra Cos: 0.9355270862579346
Inter Cos: 0.3861400783061981
Norm Quadratic Average: 20.201875686645508
Nearest Class Center Accuracy: 0.9906

