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_0.003.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: 0.11001027375459671
Inter Cos: 0.1328403800725937
Norm Quadratic Average: 65.33634948730469
Nearest Class Center Accuracy: 0.8021166666666667

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.13202393054962158
Inter Cos: 0.1676303744316101
Norm Quadratic Average: 95.09664916992188
Nearest Class Center Accuracy: 0.7944166666666667

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.13937252759933472
Inter Cos: 0.18170620501041412
Norm Quadratic Average: 153.27545166015625
Nearest Class Center Accuracy: 0.7997833333333333

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.17750456929206848
Inter Cos: 0.19418665766716003
Norm Quadratic Average: 105.63626861572266
Nearest Class Center Accuracy: 0.8363333333333334

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.19949300587177277
Inter Cos: 0.20344345271587372
Norm Quadratic Average: 82.92282104492188
Nearest Class Center Accuracy: 0.8652833333333333

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.21986344456672668
Inter Cos: 0.22946684062480927
Norm Quadratic Average: 71.29410552978516
Nearest Class Center Accuracy: 0.8950166666666667

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2637428939342499
Inter Cos: 0.23042866587638855
Norm Quadratic Average: 50.20399475097656
Nearest Class Center Accuracy: 0.9338166666666666

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2632456421852112
Inter Cos: 0.19899849593639374
Norm Quadratic Average: 16.120908737182617
Nearest Class Center Accuracy: 0.9391

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.34988877177238464
Inter Cos: 0.2481480985879898
Norm Quadratic Average: 8.45002555847168
Nearest Class Center Accuracy: 0.8983333333333333

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.43218812346458435
Inter Cos: 0.30229651927948
Norm Quadratic Average: 8.806402206420898
Nearest Class Center Accuracy: 0.9108833333333334

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5124363899230957
Inter Cos: 0.410064160823822
Norm Quadratic Average: 10.807271957397461
Nearest Class Center Accuracy: 0.9373

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6205024719238281
Inter Cos: 0.4190843999385834
Norm Quadratic Average: 8.862316131591797
Nearest Class Center Accuracy: 0.93685

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7466654181480408
Inter Cos: 0.5102819204330444
Norm Quadratic Average: 8.956610679626465
Nearest Class Center Accuracy: 0.9613333333333334

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8181751370429993
Inter Cos: 0.5291890501976013
Norm Quadratic Average: 11.223808288574219
Nearest Class Center Accuracy: 0.9824833333333334

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8625521063804626
Inter Cos: 0.5235934257507324
Norm Quadratic Average: 14.103157997131348
Nearest Class Center Accuracy: 0.99025

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.6942038536071777
Linear Weight Rank: 892
Intra Cos: 0.8672812581062317
Inter Cos: 0.41707006096839905
Norm Quadratic Average: 63.11037826538086
Nearest Class Center Accuracy: 0.9923166666666666

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.7265595197677612
Linear Weight Rank: 2644
Intra Cos: 0.9127787351608276
Inter Cos: 0.404760479927063
Norm Quadratic Average: 46.184635162353516
Nearest Class Center Accuracy: 0.9974166666666666

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.7016754150390625
Linear Weight Rank: 9
Intra Cos: 0.9305506944656372
Inter Cos: 0.32520920038223267
Norm Quadratic Average: 30.030370712280273
Nearest Class Center Accuracy: 0.9990666666666667

Output Layer:
Intra Cos: 0.9694498777389526
Inter Cos: 0.3444157838821411
Norm Quadratic Average: 21.12765121459961
Nearest Class Center Accuracy: 0.9996333333333334

Test Set:
Average Loss: 0.03246361355157569
Accuracy: 0.9909
NC1 Within Class Collapse: 0.7834844589233398
NC2 Equinorm: Features: 0.08601994067430496, Weights: 0.059917598962783813
NC2 Equiangle: Features: 0.30110535091824003, Weights: 0.20323571099175347
NC3 Self-Duality: 0.0998072475194931
NC4 NCC Mismatch: 0.0047000000000000375

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: 0.12160128355026245
Inter Cos: 0.1457478553056717
Norm Quadratic Average: 65.67005920410156
Nearest Class Center Accuracy: 0.8188

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.14635275304317474
Inter Cos: 0.18340995907783508
Norm Quadratic Average: 95.41896057128906
Nearest Class Center Accuracy: 0.812

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.15404769778251648
Inter Cos: 0.19880180060863495
Norm Quadratic Average: 153.81991577148438
Nearest Class Center Accuracy: 0.8171

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1898326575756073
Inter Cos: 0.21274563670158386
Norm Quadratic Average: 105.77262878417969
Nearest Class Center Accuracy: 0.853

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2124381959438324
Inter Cos: 0.2230874001979828
Norm Quadratic Average: 83.02350616455078
Nearest Class Center Accuracy: 0.88

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.231246680021286
Inter Cos: 0.2233111560344696
Norm Quadratic Average: 71.44783782958984
Nearest Class Center Accuracy: 0.9093

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.273460328578949
Inter Cos: 0.2227657586336136
Norm Quadratic Average: 50.442203521728516
Nearest Class Center Accuracy: 0.9425

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.27383044362068176
Inter Cos: 0.19230204820632935
Norm Quadratic Average: 16.201900482177734
Nearest Class Center Accuracy: 0.9477

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.36335688829421997
Inter Cos: 0.27135029435157776
Norm Quadratic Average: 8.494630813598633
Nearest Class Center Accuracy: 0.9074

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.44312340021133423
Inter Cos: 0.3272738754749298
Norm Quadratic Average: 8.878936767578125
Nearest Class Center Accuracy: 0.9149

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5180118083953857
Inter Cos: 0.42921075224876404
Norm Quadratic Average: 10.935622215270996
Nearest Class Center Accuracy: 0.938

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6310994029045105
Inter Cos: 0.4322304129600525
Norm Quadratic Average: 8.95494270324707
Nearest Class Center Accuracy: 0.9369

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7584661245346069
Inter Cos: 0.5003796219825745
Norm Quadratic Average: 9.05261516571045
Nearest Class Center Accuracy: 0.9552

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8259984254837036
Inter Cos: 0.5163184404373169
Norm Quadratic Average: 11.354369163513184
Nearest Class Center Accuracy: 0.9725

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8661263585090637
Inter Cos: 0.5088589787483215
Norm Quadratic Average: 14.26424503326416
Nearest Class Center Accuracy: 0.9804

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.6942038536071777
Linear Weight Rank: 892
Intra Cos: 0.867902934551239
Inter Cos: 0.4335159957408905
Norm Quadratic Average: 63.8123664855957
Nearest Class Center Accuracy: 0.9833

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.7265595197677612
Linear Weight Rank: 2644
Intra Cos: 0.9102892875671387
Inter Cos: 0.4208594560623169
Norm Quadratic Average: 46.731605529785156
Nearest Class Center Accuracy: 0.9885

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.7016754150390625
Linear Weight Rank: 9
Intra Cos: 0.923681914806366
Inter Cos: 0.34002289175987244
Norm Quadratic Average: 30.397804260253906
Nearest Class Center Accuracy: 0.9894

Output Layer:
Intra Cos: 0.9554939866065979
Inter Cos: 0.3601536750793457
Norm Quadratic Average: 21.394615173339844
Nearest Class Center Accuracy: 0.9907

