Model save path: ./New_Models/bn_True_dataset_MNIST_epochs_200_lr_0.01_model_type_vgg19_rand_seed_846264_test_samples_None_train_samples_None_weight_decay_0.0003.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: 0.10967151820659637
Norm Quadratic Average: 23.567686080932617
Nearest Class Center Accuracy: 0.8079833333333334

Layer 1: Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.06186744198203087
Inter Cos: 0.07460034638643265
Norm Quadratic Average: 44.32401657104492
Nearest Class Center Accuracy: 0.8273666666666667

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10247378796339035
Inter Cos: 0.1000305712223053
Norm Quadratic Average: 28.923221588134766
Nearest Class Center Accuracy: 0.8690666666666667

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10327673703432083
Inter Cos: 0.09504425525665283
Norm Quadratic Average: 29.726547241210938
Nearest Class Center Accuracy: 0.88225

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.17344999313354492
Inter Cos: 0.1120854839682579
Norm Quadratic Average: 17.483945846557617
Nearest Class Center Accuracy: 0.9339166666666666

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.20681357383728027
Inter Cos: 0.11680418252944946
Norm Quadratic Average: 19.432701110839844
Nearest Class Center Accuracy: 0.9536333333333333

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.23093608021736145
Inter Cos: 0.11477486789226532
Norm Quadratic Average: 19.764232635498047
Nearest Class Center Accuracy: 0.9664

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.25717389583587646
Inter Cos: 0.1128077283501625
Norm Quadratic Average: 20.290517807006836
Nearest Class Center Accuracy: 0.97425

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3305957019329071
Inter Cos: 0.09805628657341003
Norm Quadratic Average: 13.939114570617676
Nearest Class Center Accuracy: 0.99275

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4746752977371216
Inter Cos: 0.13014991581439972
Norm Quadratic Average: 14.818257331848145
Nearest Class Center Accuracy: 0.9976333333333334

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6047624349594116
Inter Cos: 0.119953453540802
Norm Quadratic Average: 15.666369438171387
Nearest Class Center Accuracy: 0.9992833333333333

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7120618224143982
Inter Cos: 0.12617993354797363
Norm Quadratic Average: 15.867838859558105
Nearest Class Center Accuracy: 0.9997333333333334

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.833945631980896
Inter Cos: 0.14771422743797302
Norm Quadratic Average: 13.137055397033691
Nearest Class Center Accuracy: 0.9997666666666667

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9354177713394165
Inter Cos: 0.11540155112743378
Norm Quadratic Average: 8.3323335647583
Nearest Class Center Accuracy: 0.9999833333333333

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.959912896156311
Inter Cos: 0.06629902869462967
Norm Quadratic Average: 8.61873722076416
Nearest Class Center Accuracy: 1.0

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9733694195747375
Inter Cos: 0.002578272018581629
Norm Quadratic Average: 8.878152847290039
Nearest Class Center Accuracy: 1.0

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 31.674518585205078
Linear Weight Rank: 4031
Intra Cos: 0.9886249899864197
Inter Cos: -0.006049581803381443
Norm Quadratic Average: 73.88092041015625
Nearest Class Center Accuracy: 1.0

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 12.964618682861328
Linear Weight Rank: 3670
Intra Cos: 0.9919309616088867
Inter Cos: 0.04572054371237755
Norm Quadratic Average: 41.48237228393555
Nearest Class Center Accuracy: 1.0

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.9386622905731201
Linear Weight Rank: 10
Intra Cos: 0.9908033013343811
Inter Cos: 0.06208197772502899
Norm Quadratic Average: 23.525856018066406
Nearest Class Center Accuracy: 1.0

Output Layer:
Intra Cos: 0.9990153312683105
Inter Cos: 0.17756995558738708
Norm Quadratic Average: 14.678060531616211
Nearest Class Center Accuracy: 1.0

Test Set:
Average Loss: 0.018356131823474425
Accuracy: 0.9968
NC1 Within Class Collapse: 0.12613171339035034
NC2 Equinorm: Features: 0.022232865914702415, Weights: 0.014507015235722065
NC2 Equiangle: Features: 0.10577919218275282, Weights: 0.08028753598531087
NC3 Self-Duality: 0.17018680274486542
NC4 NCC Mismatch: 9.999999999998899e-05

Layer 0: Conv2d(1, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10219582170248032
Inter Cos: 0.12048851698637009
Norm Quadratic Average: 23.59519386291504
Nearest Class Center Accuracy: 0.8229

Layer 1: Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.06969090551137924
Inter Cos: 0.07544398307800293
Norm Quadratic Average: 44.18630599975586
Nearest Class Center Accuracy: 0.8395

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.11308155208826065
Inter Cos: 0.10020111501216888
Norm Quadratic Average: 28.711334228515625
Nearest Class Center Accuracy: 0.8803

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.11375047266483307
Inter Cos: 0.09656473994255066
Norm Quadratic Average: 29.53076934814453
Nearest Class Center Accuracy: 0.8931

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.18752379715442657
Inter Cos: 0.12230151146650314
Norm Quadratic Average: 17.364456176757812
Nearest Class Center Accuracy: 0.9414

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.21946527063846588
Inter Cos: 0.1259821504354477
Norm Quadratic Average: 19.304101943969727
Nearest Class Center Accuracy: 0.958

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.24313856661319733
Inter Cos: 0.11462830752134323
Norm Quadratic Average: 19.65576171875
Nearest Class Center Accuracy: 0.9679

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.26886770129203796
Inter Cos: 0.11431118100881577
Norm Quadratic Average: 20.209178924560547
Nearest Class Center Accuracy: 0.9749

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3398469090461731
Inter Cos: 0.09937920421361923
Norm Quadratic Average: 13.89841365814209
Nearest Class Center Accuracy: 0.9886

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4809083938598633
Inter Cos: 0.12993746995925903
Norm Quadratic Average: 14.796292304992676
Nearest Class Center Accuracy: 0.9918

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6084247827529907
Inter Cos: 0.11829368025064468
Norm Quadratic Average: 15.66262435913086
Nearest Class Center Accuracy: 0.9929

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7137691974639893
Inter Cos: 0.12401670962572098
Norm Quadratic Average: 15.871278762817383
Nearest Class Center Accuracy: 0.9944

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8302354216575623
Inter Cos: 0.1436779499053955
Norm Quadratic Average: 13.144017219543457
Nearest Class Center Accuracy: 0.9939

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9279208183288574
Inter Cos: 0.1103513091802597
Norm Quadratic Average: 8.332324981689453
Nearest Class Center Accuracy: 0.995

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9461871385574341
Inter Cos: 0.06152765825390816
Norm Quadratic Average: 8.612401008605957
Nearest Class Center Accuracy: 0.9959

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9579176306724548
Inter Cos: -0.0018768827430903912
Norm Quadratic Average: 8.867366790771484
Nearest Class Center Accuracy: 0.9962

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 31.674518585205078
Linear Weight Rank: 4031
Intra Cos: 0.9692865610122681
Inter Cos: -0.004638017155230045
Norm Quadratic Average: 73.75677490234375
Nearest Class Center Accuracy: 0.9968

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 12.964618682861328
Linear Weight Rank: 3670
Intra Cos: 0.9734829068183899
Inter Cos: 0.046054571866989136
Norm Quadratic Average: 41.410945892333984
Nearest Class Center Accuracy: 0.9969

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.9386622905731201
Linear Weight Rank: 10
Intra Cos: 0.9718793034553528
Inter Cos: 0.0619417205452919
Norm Quadratic Average: 23.493610382080078
Nearest Class Center Accuracy: 0.9969

Output Layer:
Intra Cos: 0.9874765872955322
Inter Cos: 0.18834161758422852
Norm Quadratic Average: 14.645659446716309
Nearest Class Center Accuracy: 0.9969

