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_0.0003.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: 0.0641440749168396
Inter Cos: 0.07694800198078156
Norm Quadratic Average: 42.573631286621094
Nearest Class Center Accuracy: 0.8269

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10333320498466492
Inter Cos: 0.09962642937898636
Norm Quadratic Average: 24.955860137939453
Nearest Class Center Accuracy: 0.87235

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10868865251541138
Inter Cos: 0.10137289762496948
Norm Quadratic Average: 27.93535614013672
Nearest Class Center Accuracy: 0.886

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.17991028726100922
Inter Cos: 0.12365145981311798
Norm Quadratic Average: 17.787662506103516
Nearest Class Center Accuracy: 0.9334833333333333

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.21233753859996796
Inter Cos: 0.133197620511055
Norm Quadratic Average: 18.868223190307617
Nearest Class Center Accuracy: 0.9542666666666667

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.23792845010757446
Inter Cos: 0.13371172547340393
Norm Quadratic Average: 19.970272064208984
Nearest Class Center Accuracy: 0.9669666666666666

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.259351909160614
Inter Cos: 0.12155140936374664
Norm Quadratic Average: 20.417465209960938
Nearest Class Center Accuracy: 0.9752833333333333

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.32574349641799927
Inter Cos: 0.12781037390232086
Norm Quadratic Average: 13.878554344177246
Nearest Class Center Accuracy: 0.9928333333333333

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4659402370452881
Inter Cos: 0.15549442172050476
Norm Quadratic Average: 14.89299488067627
Nearest Class Center Accuracy: 0.9974666666666666

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.593543529510498
Inter Cos: 0.1507416069507599
Norm Quadratic Average: 15.742015838623047
Nearest Class Center Accuracy: 0.9989

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6971710920333862
Inter Cos: 0.10772301256656647
Norm Quadratic Average: 16.11397361755371
Nearest Class Center Accuracy: 0.99975

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8135793209075928
Inter Cos: 0.21383197605609894
Norm Quadratic Average: 13.321678161621094
Nearest Class Center Accuracy: 0.9999166666666667

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9268523454666138
Inter Cos: 0.1485985964536667
Norm Quadratic Average: 8.452545166015625
Nearest Class Center Accuracy: 0.9999833333333333

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9615908861160278
Inter Cos: 0.07180096209049225
Norm Quadratic Average: 8.635381698608398
Nearest Class Center Accuracy: 1.0

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9776264429092407
Inter Cos: 0.014364298433065414
Norm Quadratic Average: 8.845962524414062
Nearest Class Center Accuracy: 1.0

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 31.672698974609375
Linear Weight Rank: 4031
Intra Cos: 0.9906399846076965
Inter Cos: -0.014121351763606071
Norm Quadratic Average: 74.03799438476562
Nearest Class Center Accuracy: 1.0

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 12.96849250793457
Linear Weight Rank: 3669
Intra Cos: 0.9930715560913086
Inter Cos: 0.0064629758708179
Norm Quadratic Average: 41.5059928894043
Nearest Class Center Accuracy: 1.0

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.9583890438079834
Linear Weight Rank: 10
Intra Cos: 0.9914631247520447
Inter Cos: 0.05075903981924057
Norm Quadratic Average: 23.447925567626953
Nearest Class Center Accuracy: 1.0

Output Layer:
Intra Cos: 0.9991286396980286
Inter Cos: 0.17852230370044708
Norm Quadratic Average: 14.517058372497559
Nearest Class Center Accuracy: 1.0

Test Set:
Average Loss: 0.022598454718184075
Accuracy: 0.9956
NC1 Within Class Collapse: 0.13391044735908508
NC2 Equinorm: Features: 0.02896158955991268, Weights: 0.02577289752662182
NC2 Equiangle: Features: 0.08823214636908637, Weights: 0.06505191591050889
NC3 Self-Duality: 0.17882221937179565
NC4 NCC Mismatch: 9.999999999998899e-05

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: 0.07246848940849304
Inter Cos: 0.07834595441818237
Norm Quadratic Average: 42.432193756103516
Nearest Class Center Accuracy: 0.8396

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.11377718299627304
Inter Cos: 0.10080903768539429
Norm Quadratic Average: 24.766313552856445
Nearest Class Center Accuracy: 0.8848

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.11951816082000732
Inter Cos: 0.10245634615421295
Norm Quadratic Average: 27.74676513671875
Nearest Class Center Accuracy: 0.8972

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.19497621059417725
Inter Cos: 0.13536518812179565
Norm Quadratic Average: 17.665708541870117
Nearest Class Center Accuracy: 0.9401

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.22577986121177673
Inter Cos: 0.14112147688865662
Norm Quadratic Average: 18.754623413085938
Nearest Class Center Accuracy: 0.9569

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.25118330121040344
Inter Cos: 0.1374112069606781
Norm Quadratic Average: 19.859130859375
Nearest Class Center Accuracy: 0.9673

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.27124738693237305
Inter Cos: 0.13434137403964996
Norm Quadratic Average: 20.31505584716797
Nearest Class Center Accuracy: 0.9741

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.33567875623703003
Inter Cos: 0.12790550291538239
Norm Quadratic Average: 13.823928833007812
Nearest Class Center Accuracy: 0.9886

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4740334153175354
Inter Cos: 0.15380319952964783
Norm Quadratic Average: 14.8526029586792
Nearest Class Center Accuracy: 0.9916

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.59816575050354
Inter Cos: 0.14746378362178802
Norm Quadratic Average: 15.710604667663574
Nearest Class Center Accuracy: 0.9933

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6982762217521667
Inter Cos: 0.10240438580513
Norm Quadratic Average: 16.085311889648438
Nearest Class Center Accuracy: 0.9939

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8077905774116516
Inter Cos: 0.20674005150794983
Norm Quadratic Average: 13.30036449432373
Nearest Class Center Accuracy: 0.9937

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9162014722824097
Inter Cos: 0.14133688807487488
Norm Quadratic Average: 8.44007396697998
Nearest Class Center Accuracy: 0.9941

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9508500099182129
Inter Cos: 0.0725063607096672
Norm Quadratic Average: 8.62005615234375
Nearest Class Center Accuracy: 0.995

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9621992707252502
Inter Cos: 0.01434885710477829
Norm Quadratic Average: 8.82988166809082
Nearest Class Center Accuracy: 0.9954

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 31.672698974609375
Linear Weight Rank: 4031
Intra Cos: 0.9701372981071472
Inter Cos: -0.013626215979456902
Norm Quadratic Average: 73.88780975341797
Nearest Class Center Accuracy: 0.9954

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 12.96849250793457
Linear Weight Rank: 3669
Intra Cos: 0.9719966053962708
Inter Cos: 0.006505673285573721
Norm Quadratic Average: 41.41516876220703
Nearest Class Center Accuracy: 0.9956

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.9583890438079834
Linear Weight Rank: 10
Intra Cos: 0.9704716205596924
Inter Cos: 0.05120063200592995
Norm Quadratic Average: 23.40462303161621
Nearest Class Center Accuracy: 0.9957

Output Layer:
Intra Cos: 0.983802318572998
Inter Cos: 0.18855935335159302
Norm Quadratic Average: 14.47927188873291
Nearest Class Center Accuracy: 0.9956

