Model save path: ./New_Models/bn_False_dataset_MNIST_epochs_200_lr_0.01_model_type_vgg19_rand_seed_338327_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: 0.09116753190755844
Inter Cos: 0.10967152565717697
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: 0.11801565438508987
Inter Cos: 0.14141911268234253
Norm Quadratic Average: 67.29712677001953
Nearest Class Center Accuracy: 0.7994166666666667

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1397966593503952
Inter Cos: 0.18092410266399384
Norm Quadratic Average: 127.266845703125
Nearest Class Center Accuracy: 0.7766833333333333

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.14499448239803314
Inter Cos: 0.18739281594753265
Norm Quadratic Average: 224.27748107910156
Nearest Class Center Accuracy: 0.7817666666666667

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.16968387365341187
Inter Cos: 0.18258270621299744
Norm Quadratic Average: 124.6476058959961
Nearest Class Center Accuracy: 0.83065

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1999858170747757
Inter Cos: 0.19586428999900818
Norm Quadratic Average: 69.13837432861328
Nearest Class Center Accuracy: 0.8768

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.24204397201538086
Inter Cos: 0.23571360111236572
Norm Quadratic Average: 46.03945541381836
Nearest Class Center Accuracy: 0.9106

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2862150967121124
Inter Cos: 0.2816354036331177
Norm Quadratic Average: 35.77677536010742
Nearest Class Center Accuracy: 0.93775

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.31795254349708557
Inter Cos: 0.28196173906326294
Norm Quadratic Average: 13.88602352142334
Nearest Class Center Accuracy: 0.9549

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5031409859657288
Inter Cos: 0.32839885354042053
Norm Quadratic Average: 9.429010391235352
Nearest Class Center Accuracy: 0.9576

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.597621500492096
Inter Cos: 0.38450342416763306
Norm Quadratic Average: 11.966951370239258
Nearest Class Center Accuracy: 0.9681166666666666

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6579998135566711
Inter Cos: 0.4337700605392456
Norm Quadratic Average: 16.649009704589844
Nearest Class Center Accuracy: 0.9712

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6656505465507507
Inter Cos: 0.3820836544036865
Norm Quadratic Average: 12.399412155151367
Nearest Class Center Accuracy: 0.9446

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6719056963920593
Inter Cos: 0.4461168944835663
Norm Quadratic Average: 11.37445068359375
Nearest Class Center Accuracy: 0.9505

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7397735714912415
Inter Cos: 0.4644160270690918
Norm Quadratic Average: 13.772018432617188
Nearest Class Center Accuracy: 0.9669833333333333

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7956475019454956
Inter Cos: 0.48017528653144836
Norm Quadratic Average: 16.807275772094727
Nearest Class Center Accuracy: 0.97735

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.5490670204162598
Linear Weight Rank: 13
Intra Cos: 0.8315852880477905
Inter Cos: 0.4571353495121002
Norm Quadratic Average: 73.74539184570312
Nearest Class Center Accuracy: 0.9849666666666667

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.5588669776916504
Linear Weight Rank: 2577
Intra Cos: 0.8853868842124939
Inter Cos: 0.3887043595314026
Norm Quadratic Average: 49.07857131958008
Nearest Class Center Accuracy: 0.9912666666666666

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.5490468740463257
Linear Weight Rank: 9
Intra Cos: 0.9054705500602722
Inter Cos: 0.30884411931037903
Norm Quadratic Average: 29.34446144104004
Nearest Class Center Accuracy: 0.9941666666666666

Output Layer:
Intra Cos: 0.9528050422668457
Inter Cos: 0.3889358341693878
Norm Quadratic Average: 18.809226989746094
Nearest Class Center Accuracy: 0.9960666666666667

Test Set:
Average Loss: 0.03441643786374479
Accuracy: 0.9902
NC1 Within Class Collapse: 1.174760103225708
NC2 Equinorm: Features: 0.12484502792358398, Weights: 0.04407331719994545
NC2 Equiangle: Features: 0.30425804985894095, Weights: 0.19861585828993056
NC3 Self-Duality: 0.11564593017101288
NC4 NCC Mismatch: 0.008000000000000007

Layer 0: Conv2d(1, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10219582170248032
Inter Cos: 0.12048853188753128
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: 0.12965159118175507
Inter Cos: 0.15482661128044128
Norm Quadratic Average: 67.7187271118164
Nearest Class Center Accuracy: 0.8161

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.15407900512218475
Inter Cos: 0.19809582829475403
Norm Quadratic Average: 127.91936492919922
Nearest Class Center Accuracy: 0.7964

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1594369113445282
Inter Cos: 0.20564918220043182
Norm Quadratic Average: 225.45376586914062
Nearest Class Center Accuracy: 0.8019

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.17986296117305756
Inter Cos: 0.20108632743358612
Norm Quadratic Average: 125.02214813232422
Nearest Class Center Accuracy: 0.8433

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.21136727929115295
Inter Cos: 0.21563124656677246
Norm Quadratic Average: 69.3450927734375
Nearest Class Center Accuracy: 0.8885

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2542358338832855
Inter Cos: 0.2533297538757324
Norm Quadratic Average: 46.15303421020508
Nearest Class Center Accuracy: 0.9154

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.29889965057373047
Inter Cos: 0.30253440141677856
Norm Quadratic Average: 35.98968505859375
Nearest Class Center Accuracy: 0.9419

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3338456153869629
Inter Cos: 0.30121734738349915
Norm Quadratic Average: 13.98381233215332
Nearest Class Center Accuracy: 0.9583

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5170984864234924
Inter Cos: 0.35658442974090576
Norm Quadratic Average: 9.49687385559082
Nearest Class Center Accuracy: 0.9558

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6022635698318481
Inter Cos: 0.40325281023979187
Norm Quadratic Average: 12.091560363769531
Nearest Class Center Accuracy: 0.9658

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6568578481674194
Inter Cos: 0.45159125328063965
Norm Quadratic Average: 16.861770629882812
Nearest Class Center Accuracy: 0.9705

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6696818470954895
Inter Cos: 0.38403594493865967
Norm Quadratic Average: 12.520902633666992
Nearest Class Center Accuracy: 0.9437

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6944260001182556
Inter Cos: 0.46358051896095276
Norm Quadratic Average: 11.477263450622559
Nearest Class Center Accuracy: 0.9476

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.759040355682373
Inter Cos: 0.47826552391052246
Norm Quadratic Average: 13.912175178527832
Nearest Class Center Accuracy: 0.9632

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8117161393165588
Inter Cos: 0.49180683493614197
Norm Quadratic Average: 16.989328384399414
Nearest Class Center Accuracy: 0.972

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.5490670204162598
Linear Weight Rank: 13
Intra Cos: 0.8444714546203613
Inter Cos: 0.4696253538131714
Norm Quadratic Average: 74.55838775634766
Nearest Class Center Accuracy: 0.979

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.5588669776916504
Linear Weight Rank: 2577
Intra Cos: 0.8878211379051208
Inter Cos: 0.4002853035926819
Norm Quadratic Average: 49.652774810791016
Nearest Class Center Accuracy: 0.9849

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.5490468740463257
Linear Weight Rank: 9
Intra Cos: 0.9042763113975525
Inter Cos: 0.32748955488204956
Norm Quadratic Average: 29.700288772583008
Nearest Class Center Accuracy: 0.987

Output Layer:
Intra Cos: 0.9404551982879639
Inter Cos: 0.40771251916885376
Norm Quadratic Average: 19.05634880065918
Nearest Class Center Accuracy: 0.9885

