Model save path: ./New_Models/bn_True_dataset_MNIST_epochs_200_lr_0.01_model_type_vgg19_rand_seed_979323_test_samples_None_train_samples_None_weight_decay_0.0005.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: 0.06346476078033447
Inter Cos: 0.07750663161277771
Norm Quadratic Average: 19.339248657226562
Nearest Class Center Accuracy: 0.8279

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.09898336976766586
Inter Cos: 0.0937482938170433
Norm Quadratic Average: 10.391762733459473
Nearest Class Center Accuracy: 0.8790333333333333

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10100998729467392
Inter Cos: 0.09252419322729111
Norm Quadratic Average: 12.306222915649414
Nearest Class Center Accuracy: 0.8915333333333333

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.17247672379016876
Inter Cos: 0.11159331351518631
Norm Quadratic Average: 7.980443477630615
Nearest Class Center Accuracy: 0.9361333333333334

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.20322762429714203
Inter Cos: 0.11582775413990021
Norm Quadratic Average: 8.730169296264648
Nearest Class Center Accuracy: 0.9552166666666667

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.22963137924671173
Inter Cos: 0.11815135180950165
Norm Quadratic Average: 9.112960815429688
Nearest Class Center Accuracy: 0.9680666666666666

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2604619264602661
Inter Cos: 0.12072137743234634
Norm Quadratic Average: 9.220658302307129
Nearest Class Center Accuracy: 0.9755666666666667

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3396739959716797
Inter Cos: 0.1187887191772461
Norm Quadratic Average: 6.563582897186279
Nearest Class Center Accuracy: 0.99335

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5318118333816528
Inter Cos: 0.14148321747779846
Norm Quadratic Average: 7.152146339416504
Nearest Class Center Accuracy: 0.9982833333333333

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6669269800186157
Inter Cos: 0.11105336248874664
Norm Quadratic Average: 7.691113471984863
Nearest Class Center Accuracy: 0.9994

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7674251198768616
Inter Cos: 0.08342020958662033
Norm Quadratic Average: 7.8133225440979
Nearest Class Center Accuracy: 0.9999166666666667

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8590877652168274
Inter Cos: 0.13281948864459991
Norm Quadratic Average: 6.809864521026611
Nearest Class Center Accuracy: 0.9999

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9560009241104126
Inter Cos: 0.048686299473047256
Norm Quadratic Average: 4.196434497833252
Nearest Class Center Accuracy: 1.0

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9785290956497192
Inter Cos: -0.02020343765616417
Norm Quadratic Average: 4.259670734405518
Nearest Class Center Accuracy: 1.0

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9906661510467529
Inter Cos: -0.03329667076468468
Norm Quadratic Average: 4.307551860809326
Nearest Class Center Accuracy: 1.0

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 14.676819801330566
Linear Weight Rank: 4031
Intra Cos: 0.9975354075431824
Inter Cos: -0.0302019864320755
Norm Quadratic Average: 43.46059036254883
Nearest Class Center Accuracy: 1.0

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 6.267735004425049
Linear Weight Rank: 3667
Intra Cos: 0.9977073073387146
Inter Cos: 0.012646925635635853
Norm Quadratic Average: 27.12813949584961
Nearest Class Center Accuracy: 1.0

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 2.071561336517334
Linear Weight Rank: 10
Intra Cos: 0.9972224235534668
Inter Cos: 0.053776808083057404
Norm Quadratic Average: 17.679447174072266
Nearest Class Center Accuracy: 1.0

Output Layer:
Intra Cos: 0.9997223019599915
Inter Cos: 0.11660086363554001
Norm Quadratic Average: 12.47611141204834
Nearest Class Center Accuracy: 1.0

Test Set:
Average Loss: 0.02475362075183075
Accuracy: 0.9949
NC1 Within Class Collapse: 0.09548592567443848
NC2 Equinorm: Features: 0.023251738399267197, Weights: 0.0187420342117548
NC2 Equiangle: Features: 0.09555142720540365, Weights: 0.06590613259209527
NC3 Self-Duality: 0.04983362555503845
NC4 NCC Mismatch: 0.00019999999999997797

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: 0.07141819596290588
Inter Cos: 0.07889973372220993
Norm Quadratic Average: 19.275211334228516
Nearest Class Center Accuracy: 0.8389

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1085781678557396
Inter Cos: 0.09582117199897766
Norm Quadratic Average: 10.311700820922852
Nearest Class Center Accuracy: 0.8897

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.11090515553951263
Inter Cos: 0.09377700835466385
Norm Quadratic Average: 12.235115051269531
Nearest Class Center Accuracy: 0.9006

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1838071197271347
Inter Cos: 0.11728832870721817
Norm Quadratic Average: 7.942008018493652
Nearest Class Center Accuracy: 0.9413

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.21592874825000763
Inter Cos: 0.12749573588371277
Norm Quadratic Average: 8.693523406982422
Nearest Class Center Accuracy: 0.9592

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.24196864664554596
Inter Cos: 0.11685187369585037
Norm Quadratic Average: 9.072460174560547
Nearest Class Center Accuracy: 0.9687

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.271731436252594
Inter Cos: 0.11775850504636765
Norm Quadratic Average: 9.181654930114746
Nearest Class Center Accuracy: 0.9757

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.35037317872047424
Inter Cos: 0.11926685273647308
Norm Quadratic Average: 6.536683559417725
Nearest Class Center Accuracy: 0.9889

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.539564311504364
Inter Cos: 0.14249256253242493
Norm Quadratic Average: 7.12730598449707
Nearest Class Center Accuracy: 0.993

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.671250581741333
Inter Cos: 0.11656900495290756
Norm Quadratic Average: 7.668416500091553
Nearest Class Center Accuracy: 0.9935

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7679942846298218
Inter Cos: 0.09443823248147964
Norm Quadratic Average: 7.793128967285156
Nearest Class Center Accuracy: 0.9941

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8525232076644897
Inter Cos: 0.13110646605491638
Norm Quadratic Average: 6.791627407073975
Nearest Class Center Accuracy: 0.9933

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9448473453521729
Inter Cos: 0.045701172202825546
Norm Quadratic Average: 4.185060977935791
Nearest Class Center Accuracy: 0.9947

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9600086808204651
Inter Cos: -0.01589830033481121
Norm Quadratic Average: 4.248765468597412
Nearest Class Center Accuracy: 0.9947

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9692458510398865
Inter Cos: -0.03719629347324371
Norm Quadratic Average: 4.296881675720215
Nearest Class Center Accuracy: 0.9949

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 14.676819801330566
Linear Weight Rank: 4031
Intra Cos: 0.9770758748054504
Inter Cos: -0.03148572891950607
Norm Quadratic Average: 43.33992004394531
Nearest Class Center Accuracy: 0.995

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 6.267735004425049
Linear Weight Rank: 3667
Intra Cos: 0.9778222441673279
Inter Cos: 0.017889026552438736
Norm Quadratic Average: 27.050975799560547
Nearest Class Center Accuracy: 0.995

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 2.071561336517334
Linear Weight Rank: 10
Intra Cos: 0.9779037237167358
Inter Cos: 0.059692613780498505
Norm Quadratic Average: 17.63192367553711
Nearest Class Center Accuracy: 0.9951

Output Layer:
Intra Cos: 0.9838964939117432
Inter Cos: 0.12545165419578552
Norm Quadratic Average: 12.437864303588867
Nearest Class Center Accuracy: 0.9951

