Model save path: ./New_Models/bn_True_dataset_MNIST_epochs_200_lr_0.01_model_type_vgg19_rand_seed_314159_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: 0.09116753935813904
Inter Cos: 0.10967151075601578
Norm Quadratic Average: 23.567678451538086
Nearest Class Center Accuracy: 0.8079833333333334

Layer 1: Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.0632854551076889
Inter Cos: 0.07754713296890259
Norm Quadratic Average: 41.62641906738281
Nearest Class Center Accuracy: 0.8280166666666666

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10072699189186096
Inter Cos: 0.0953802540898323
Norm Quadratic Average: 27.225830078125
Nearest Class Center Accuracy: 0.8731

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10712531954050064
Inter Cos: 0.09902296215295792
Norm Quadratic Average: 29.01584243774414
Nearest Class Center Accuracy: 0.8846833333333334

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.18321819603443146
Inter Cos: 0.12260052561759949
Norm Quadratic Average: 18.881139755249023
Nearest Class Center Accuracy: 0.9315833333333333

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2146683782339096
Inter Cos: 0.12978550791740417
Norm Quadratic Average: 19.32108497619629
Nearest Class Center Accuracy: 0.95245

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.24146273732185364
Inter Cos: 0.12433464825153351
Norm Quadratic Average: 19.70536994934082
Nearest Class Center Accuracy: 0.9656666666666667

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.26721706986427307
Inter Cos: 0.11580701917409897
Norm Quadratic Average: 20.510448455810547
Nearest Class Center Accuracy: 0.9740166666666666

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3386341631412506
Inter Cos: 0.1314675509929657
Norm Quadratic Average: 14.055482864379883
Nearest Class Center Accuracy: 0.9924833333333334

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.47715291380882263
Inter Cos: 0.17859899997711182
Norm Quadratic Average: 14.790782928466797
Nearest Class Center Accuracy: 0.99765

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5881643295288086
Inter Cos: 0.19594573974609375
Norm Quadratic Average: 15.758859634399414
Nearest Class Center Accuracy: 0.99925

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6892589926719666
Inter Cos: 0.19105012714862823
Norm Quadratic Average: 16.076396942138672
Nearest Class Center Accuracy: 0.9998166666666667

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8046326637268066
Inter Cos: 0.23511852324008942
Norm Quadratic Average: 13.16473388671875
Nearest Class Center Accuracy: 0.9997666666666667

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9196257591247559
Inter Cos: 0.13787637650966644
Norm Quadratic Average: 8.484085083007812
Nearest Class Center Accuracy: 0.9999833333333333

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9605665802955627
Inter Cos: 0.05657746270298958
Norm Quadratic Average: 8.657578468322754
Nearest Class Center Accuracy: 1.0

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9762500524520874
Inter Cos: -0.011273080483078957
Norm Quadratic Average: 8.917949676513672
Nearest Class Center Accuracy: 1.0

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 31.671566009521484
Linear Weight Rank: 4031
Intra Cos: 0.9865811467170715
Inter Cos: -0.004517120309174061
Norm Quadratic Average: 73.66104125976562
Nearest Class Center Accuracy: 1.0

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 12.96650218963623
Linear Weight Rank: 3669
Intra Cos: 0.99051433801651
Inter Cos: 0.06588969379663467
Norm Quadratic Average: 41.39972686767578
Nearest Class Center Accuracy: 1.0

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.945918083190918
Linear Weight Rank: 10
Intra Cos: 0.9898279905319214
Inter Cos: 0.0875822901725769
Norm Quadratic Average: 23.608875274658203
Nearest Class Center Accuracy: 1.0

Output Layer:
Intra Cos: 0.9988667964935303
Inter Cos: 0.16540135443210602
Norm Quadratic Average: 14.658767700195312
Nearest Class Center Accuracy: 1.0

Test Set:
Average Loss: 0.022333072821988025
Accuracy: 0.9954
NC1 Within Class Collapse: 0.13800206780433655
NC2 Equinorm: Features: 0.04132886976003647, Weights: 0.019046809524297714
NC2 Equiangle: Features: 0.08901801639133029, Weights: 0.07139759063720703
NC3 Self-Duality: 0.17146098613739014
NC4 NCC Mismatch: 0.00019999999999997797

Layer 0: Conv2d(1, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10219582915306091
Inter Cos: 0.12048851698637009
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: 0.07103476673364639
Inter Cos: 0.07867950946092606
Norm Quadratic Average: 41.46731185913086
Nearest Class Center Accuracy: 0.8409

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.11093814671039581
Inter Cos: 0.09586858004331589
Norm Quadratic Average: 27.032737731933594
Nearest Class Center Accuracy: 0.884

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.11701381951570511
Inter Cos: 0.09913589060306549
Norm Quadratic Average: 28.83857536315918
Nearest Class Center Accuracy: 0.8958

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.19657787680625916
Inter Cos: 0.13091276586055756
Norm Quadratic Average: 18.75506591796875
Nearest Class Center Accuracy: 0.9405

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.22783805429935455
Inter Cos: 0.1338893324136734
Norm Quadratic Average: 19.204544067382812
Nearest Class Center Accuracy: 0.9564

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.25520285964012146
Inter Cos: 0.12024228274822235
Norm Quadratic Average: 19.602813720703125
Nearest Class Center Accuracy: 0.9675

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2798706293106079
Inter Cos: 0.1217268854379654
Norm Quadratic Average: 20.420612335205078
Nearest Class Center Accuracy: 0.974

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.34986773133277893
Inter Cos: 0.13286887109279633
Norm Quadratic Average: 14.003189086914062
Nearest Class Center Accuracy: 0.99

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4864947199821472
Inter Cos: 0.17802666127681732
Norm Quadratic Average: 14.758699417114258
Nearest Class Center Accuracy: 0.9923

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5936002731323242
Inter Cos: 0.19306865334510803
Norm Quadratic Average: 15.73582649230957
Nearest Class Center Accuracy: 0.9931

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6907368898391724
Inter Cos: 0.18576379120349884
Norm Quadratic Average: 16.061187744140625
Nearest Class Center Accuracy: 0.9936

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7978460192680359
Inter Cos: 0.22628605365753174
Norm Quadratic Average: 13.154091835021973
Nearest Class Center Accuracy: 0.9929

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.907523512840271
Inter Cos: 0.13521498441696167
Norm Quadratic Average: 8.474447250366211
Nearest Class Center Accuracy: 0.9944

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9354525208473206
Inter Cos: 0.05570368468761444
Norm Quadratic Average: 8.644742012023926
Nearest Class Center Accuracy: 0.9952

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9489467740058899
Inter Cos: -0.009103460237383842
Norm Quadratic Average: 8.901177406311035
Nearest Class Center Accuracy: 0.9955

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 31.671566009521484
Linear Weight Rank: 4031
Intra Cos: 0.9615972638130188
Inter Cos: -0.0024524200707674026
Norm Quadratic Average: 73.4934310913086
Nearest Class Center Accuracy: 0.9956

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 12.96650218963623
Linear Weight Rank: 3669
Intra Cos: 0.9661233425140381
Inter Cos: 0.0660240575671196
Norm Quadratic Average: 41.302589416503906
Nearest Class Center Accuracy: 0.9956

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.945918083190918
Linear Weight Rank: 10
Intra Cos: 0.9645214676856995
Inter Cos: 0.08671644330024719
Norm Quadratic Average: 23.562259674072266
Nearest Class Center Accuracy: 0.9956

Output Layer:
Intra Cos: 0.9851359724998474
Inter Cos: 0.1754179745912552
Norm Quadratic Average: 14.618298530578613
Nearest Class Center Accuracy: 0.9955

