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_0.007.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: 0.11538634449243546
Inter Cos: 0.1392330378293991
Norm Quadratic Average: 63.890098571777344
Nearest Class Center Accuracy: 0.79975

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1358633190393448
Inter Cos: 0.17729368805885315
Norm Quadratic Average: 105.4555892944336
Nearest Class Center Accuracy: 0.7822

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.13661707937717438
Inter Cos: 0.1791660636663437
Norm Quadratic Average: 142.4394073486328
Nearest Class Center Accuracy: 0.8039666666666667

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.17587056756019592
Inter Cos: 0.18166427314281464
Norm Quadratic Average: 67.77233123779297
Nearest Class Center Accuracy: 0.8711833333333333

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.20882093906402588
Inter Cos: 0.1928490549325943
Norm Quadratic Average: 31.009620666503906
Nearest Class Center Accuracy: 0.9026833333333333

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.25463762879371643
Inter Cos: 0.23712216317653656
Norm Quadratic Average: 15.650615692138672
Nearest Class Center Accuracy: 0.8728833333333333

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2918100357055664
Inter Cos: 0.2885003089904785
Norm Quadratic Average: 14.813636779785156
Nearest Class Center Accuracy: 0.8649

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.294307142496109
Inter Cos: 0.33743947744369507
Norm Quadratic Average: 10.591995239257812
Nearest Class Center Accuracy: 0.8498833333333333

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3615557849407196
Inter Cos: 0.4252505600452423
Norm Quadratic Average: 11.061969757080078
Nearest Class Center Accuracy: 0.8712333333333333

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4417916536331177
Inter Cos: 0.4630765914916992
Norm Quadratic Average: 13.497542381286621
Nearest Class Center Accuracy: 0.9256166666666666

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5719638466835022
Inter Cos: 0.45469191670417786
Norm Quadratic Average: 14.913886070251465
Nearest Class Center Accuracy: 0.9606

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6073901057243347
Inter Cos: 0.4006552994251251
Norm Quadratic Average: 7.9780449867248535
Nearest Class Center Accuracy: 0.9454666666666667

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6874626874923706
Inter Cos: 0.5481556057929993
Norm Quadratic Average: 7.722341537475586
Nearest Class Center Accuracy: 0.9610166666666666

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7404220700263977
Inter Cos: 0.558974027633667
Norm Quadratic Average: 11.327495574951172
Nearest Class Center Accuracy: 0.97515

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7848972082138062
Inter Cos: 0.5651369690895081
Norm Quadratic Average: 16.793825149536133
Nearest Class Center Accuracy: 0.9833166666666666

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.497244954109192
Linear Weight Rank: 6
Intra Cos: 0.8418689370155334
Inter Cos: 0.5638803839683533
Norm Quadratic Average: 86.83753967285156
Nearest Class Center Accuracy: 0.9905333333333334

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.5123413801193237
Linear Weight Rank: 2500
Intra Cos: 0.8784389495849609
Inter Cos: 0.5504059791564941
Norm Quadratic Average: 61.3020133972168
Nearest Class Center Accuracy: 0.9924666666666667

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.4992398023605347
Linear Weight Rank: 9
Intra Cos: 0.9129459857940674
Inter Cos: 0.43875259160995483
Norm Quadratic Average: 34.98731994628906
Nearest Class Center Accuracy: 0.9948

Output Layer:
Intra Cos: 0.9587446451187134
Inter Cos: 0.4954133629798889
Norm Quadratic Average: 21.42237091064453
Nearest Class Center Accuracy: 0.9969166666666667

Test Set:
Average Loss: 0.03865387089326978
Accuracy: 0.9875
NC1 Within Class Collapse: 1.6702935695648193
NC2 Equinorm: Features: 0.16426095366477966, Weights: 0.0748601034283638
NC2 Equiangle: Features: 0.37368159823947483, Weights: 0.17357067532009549
NC3 Self-Duality: 0.21982713043689728
NC4 NCC Mismatch: 0.008600000000000052

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: 0.12691795825958252
Inter Cos: 0.15250863134860992
Norm Quadratic Average: 64.2757339477539
Nearest Class Center Accuracy: 0.8165

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.15013235807418823
Inter Cos: 0.19398094713687897
Norm Quadratic Average: 105.94450378417969
Nearest Class Center Accuracy: 0.8016

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1512470841407776
Inter Cos: 0.1956801414489746
Norm Quadratic Average: 143.0583953857422
Nearest Class Center Accuracy: 0.8206

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.187188059091568
Inter Cos: 0.19688719511032104
Norm Quadratic Average: 67.94527435302734
Nearest Class Center Accuracy: 0.8796

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.22243928909301758
Inter Cos: 0.20960305631160736
Norm Quadratic Average: 31.0905818939209
Nearest Class Center Accuracy: 0.913

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2716832756996155
Inter Cos: 0.24198246002197266
Norm Quadratic Average: 15.627888679504395
Nearest Class Center Accuracy: 0.8889

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3143572211265564
Inter Cos: 0.29907506704330444
Norm Quadratic Average: 14.790724754333496
Nearest Class Center Accuracy: 0.8836

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.30881595611572266
Inter Cos: 0.3480379283428192
Norm Quadratic Average: 10.584365844726562
Nearest Class Center Accuracy: 0.8685

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.38024047017097473
Inter Cos: 0.42505496740341187
Norm Quadratic Average: 11.090691566467285
Nearest Class Center Accuracy: 0.8873

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.46087828278541565
Inter Cos: 0.45366916060447693
Norm Quadratic Average: 13.595849990844727
Nearest Class Center Accuracy: 0.9316

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5885526537895203
Inter Cos: 0.4371061623096466
Norm Quadratic Average: 15.064453125
Nearest Class Center Accuracy: 0.9597

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6296142339706421
Inter Cos: 0.40856966376304626
Norm Quadratic Average: 8.053580284118652
Nearest Class Center Accuracy: 0.9471

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7043944597244263
Inter Cos: 0.5628643035888672
Norm Quadratic Average: 7.7852020263671875
Nearest Class Center Accuracy: 0.9587

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7536013722419739
Inter Cos: 0.5694351196289062
Norm Quadratic Average: 11.435277938842773
Nearest Class Center Accuracy: 0.9696

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7946305871009827
Inter Cos: 0.5723057389259338
Norm Quadratic Average: 16.96796989440918
Nearest Class Center Accuracy: 0.9748

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.497244954109192
Linear Weight Rank: 6
Intra Cos: 0.8473327159881592
Inter Cos: 0.5674344301223755
Norm Quadratic Average: 87.74970245361328
Nearest Class Center Accuracy: 0.9821

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.5123413801193237
Linear Weight Rank: 2500
Intra Cos: 0.8801807761192322
Inter Cos: 0.551681399345398
Norm Quadratic Average: 61.945674896240234
Nearest Class Center Accuracy: 0.9829

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.4992398023605347
Linear Weight Rank: 9
Intra Cos: 0.9084916114807129
Inter Cos: 0.4416849613189697
Norm Quadratic Average: 35.35752487182617
Nearest Class Center Accuracy: 0.9843

Output Layer:
Intra Cos: 0.9479585886001587
Inter Cos: 0.4908820390701294
Norm Quadratic Average: 21.66020393371582
Nearest Class Center Accuracy: 0.9862

