Model save path: ./New_Models/bn_False_dataset_MNIST_epochs_200_lr_0.01_model_type_vgg19_rand_seed_846264_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: 0.10967151820659637
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: 0.09949398040771484
Inter Cos: 0.11917014420032501
Norm Quadratic Average: 51.73712158203125
Nearest Class Center Accuracy: 0.8071166666666667

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.13314205408096313
Inter Cos: 0.14396536350250244
Norm Quadratic Average: 50.38465118408203
Nearest Class Center Accuracy: 0.8170833333333334

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.17862902581691742
Inter Cos: 0.16632898151874542
Norm Quadratic Average: 66.12805938720703
Nearest Class Center Accuracy: 0.8322833333333334

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.21194805204868317
Inter Cos: 0.1914830207824707
Norm Quadratic Average: 44.21391677856445
Nearest Class Center Accuracy: 0.8854166666666666

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.24096514284610748
Inter Cos: 0.19279946386814117
Norm Quadratic Average: 26.987342834472656
Nearest Class Center Accuracy: 0.9136166666666666

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3075704276561737
Inter Cos: 0.28271007537841797
Norm Quadratic Average: 18.571308135986328
Nearest Class Center Accuracy: 0.9380666666666667

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3441861867904663
Inter Cos: 0.33365777134895325
Norm Quadratic Average: 24.230022430419922
Nearest Class Center Accuracy: 0.9496

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3472748398780823
Inter Cos: 0.3731676936149597
Norm Quadratic Average: 18.257978439331055
Nearest Class Center Accuracy: 0.9544166666666667

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4212828576564789
Inter Cos: 0.40456050634384155
Norm Quadratic Average: 16.67158317565918
Nearest Class Center Accuracy: 0.9571166666666666

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5164400935173035
Inter Cos: 0.384872168302536
Norm Quadratic Average: 18.494199752807617
Nearest Class Center Accuracy: 0.96625

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6126006245613098
Inter Cos: 0.37539124488830566
Norm Quadratic Average: 19.982357025146484
Nearest Class Center Accuracy: 0.97615

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5114525556564331
Inter Cos: 0.4147820770740509
Norm Quadratic Average: 11.944073677062988
Nearest Class Center Accuracy: 0.97345

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6576453447341919
Inter Cos: 0.5276916027069092
Norm Quadratic Average: 10.472637176513672
Nearest Class Center Accuracy: 0.9758

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7303619980812073
Inter Cos: 0.5630083680152893
Norm Quadratic Average: 12.63581371307373
Nearest Class Center Accuracy: 0.9831833333333333

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7900395393371582
Inter Cos: 0.5891008973121643
Norm Quadratic Average: 15.381097793579102
Nearest Class Center Accuracy: 0.9876666666666667

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.6510881185531616
Linear Weight Rank: 41
Intra Cos: 0.8517902493476868
Inter Cos: 0.5592913627624512
Norm Quadratic Average: 67.67184448242188
Nearest Class Center Accuracy: 0.9916666666666667

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.6651370525360107
Linear Weight Rank: 2591
Intra Cos: 0.9005075097084045
Inter Cos: 0.4904423952102661
Norm Quadratic Average: 44.166324615478516
Nearest Class Center Accuracy: 0.9932666666666666

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.6498799324035645
Linear Weight Rank: 9
Intra Cos: 0.912286639213562
Inter Cos: 0.3916529417037964
Norm Quadratic Average: 27.166152954101562
Nearest Class Center Accuracy: 0.99365

Output Layer:
Intra Cos: 0.9200577735900879
Inter Cos: 0.3898982107639313
Norm Quadratic Average: 18.70899772644043
Nearest Class Center Accuracy: 0.9946

Test Set:
Average Loss: 0.02421673403829336
Accuracy: 0.9925
NC1 Within Class Collapse: 1.0737755298614502
NC2 Equinorm: Features: 0.163526251912117, Weights: 0.04361478611826897
NC2 Equiangle: Features: 0.2528548770480686, Weights: 0.1998760011461046
NC3 Self-Duality: 0.10604078322649002
NC4 NCC Mismatch: 0.006199999999999983

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: 0.11069287359714508
Inter Cos: 0.13099275529384613
Norm Quadratic Average: 51.91512680053711
Nearest Class Center Accuracy: 0.8211

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.14627525210380554
Inter Cos: 0.1573854386806488
Norm Quadratic Average: 50.42565155029297
Nearest Class Center Accuracy: 0.829

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1929481029510498
Inter Cos: 0.17963188886642456
Norm Quadratic Average: 66.25141143798828
Nearest Class Center Accuracy: 0.845

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2261839210987091
Inter Cos: 0.18645907938480377
Norm Quadratic Average: 44.247467041015625
Nearest Class Center Accuracy: 0.8973

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2584022581577301
Inter Cos: 0.18500882387161255
Norm Quadratic Average: 27.028827667236328
Nearest Class Center Accuracy: 0.9216

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.33121418952941895
Inter Cos: 0.2889341413974762
Norm Quadratic Average: 18.597875595092773
Nearest Class Center Accuracy: 0.9483

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3669469952583313
Inter Cos: 0.34738069772720337
Norm Quadratic Average: 24.32113265991211
Nearest Class Center Accuracy: 0.9544

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.36771881580352783
Inter Cos: 0.39696213603019714
Norm Quadratic Average: 18.34952163696289
Nearest Class Center Accuracy: 0.9558

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.434457004070282
Inter Cos: 0.4280039072036743
Norm Quadratic Average: 16.789180755615234
Nearest Class Center Accuracy: 0.9583

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5236614346504211
Inter Cos: 0.37260520458221436
Norm Quadratic Average: 18.64381980895996
Nearest Class Center Accuracy: 0.9677

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6189783215522766
Inter Cos: 0.4111658036708832
Norm Quadratic Average: 20.162260055541992
Nearest Class Center Accuracy: 0.9764

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5213960409164429
Inter Cos: 0.4365449547767639
Norm Quadratic Average: 12.064840316772461
Nearest Class Center Accuracy: 0.9742

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6725895404815674
Inter Cos: 0.5509817600250244
Norm Quadratic Average: 10.588701248168945
Nearest Class Center Accuracy: 0.9743

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7464320063591003
Inter Cos: 0.5853669047355652
Norm Quadratic Average: 12.788725852966309
Nearest Class Center Accuracy: 0.9814

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8054461479187012
Inter Cos: 0.6104295253753662
Norm Quadratic Average: 15.58069133758545
Nearest Class Center Accuracy: 0.9838

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.6510881185531616
Linear Weight Rank: 41
Intra Cos: 0.8626710176467896
Inter Cos: 0.5794404745101929
Norm Quadratic Average: 68.54431915283203
Nearest Class Center Accuracy: 0.9881

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.6651370525360107
Linear Weight Rank: 2591
Intra Cos: 0.9026992917060852
Inter Cos: 0.5106121897697449
Norm Quadratic Average: 44.74884033203125
Nearest Class Center Accuracy: 0.9887

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.6498799324035645
Linear Weight Rank: 9
Intra Cos: 0.9141719937324524
Inter Cos: 0.4120636582374573
Norm Quadratic Average: 27.52760124206543
Nearest Class Center Accuracy: 0.9884

Output Layer:
Intra Cos: 0.9227573871612549
Inter Cos: 0.41096094250679016
Norm Quadratic Average: 18.96272087097168
Nearest Class Center Accuracy: 0.9892

