Model save path: ./New_Models/bn_False_dataset_MNIST_epochs_200_lr_0.01_model_type_vgg19_rand_seed_265358_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: 0.09116753190755844
Inter Cos: 0.10967151820659637
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: 0.09161445498466492
Inter Cos: 0.11046087741851807
Norm Quadratic Average: 63.40493392944336
Nearest Class Center Accuracy: 0.81125

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.12554888427257538
Inter Cos: 0.13599500060081482
Norm Quadratic Average: 67.73603820800781
Nearest Class Center Accuracy: 0.8408833333333333

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1331464797258377
Inter Cos: 0.14192257821559906
Norm Quadratic Average: 90.41719818115234
Nearest Class Center Accuracy: 0.8535833333333334

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2076662927865982
Inter Cos: 0.1701391637325287
Norm Quadratic Average: 61.006248474121094
Nearest Class Center Accuracy: 0.9144166666666667

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2421053647994995
Inter Cos: 0.18020552396774292
Norm Quadratic Average: 51.51734924316406
Nearest Class Center Accuracy: 0.94145

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2877484858036041
Inter Cos: 0.18977753818035126
Norm Quadratic Average: 38.26802444458008
Nearest Class Center Accuracy: 0.95775

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3304799199104309
Inter Cos: 0.1962970644235611
Norm Quadratic Average: 28.222393035888672
Nearest Class Center Accuracy: 0.9653166666666667

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3931024670600891
Inter Cos: 0.2013823240995407
Norm Quadratic Average: 13.030858039855957
Nearest Class Center Accuracy: 0.98125

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5417512655258179
Inter Cos: 0.2678733766078949
Norm Quadratic Average: 10.567998886108398
Nearest Class Center Accuracy: 0.9903666666666666

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6086857318878174
Inter Cos: 0.3077969551086426
Norm Quadratic Average: 11.241573333740234
Nearest Class Center Accuracy: 0.9933333333333333

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6325840353965759
Inter Cos: 0.31235450506210327
Norm Quadratic Average: 11.985800743103027
Nearest Class Center Accuracy: 0.9950833333333333

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6504300236701965
Inter Cos: 0.30042406916618347
Norm Quadratic Average: 7.780238628387451
Nearest Class Center Accuracy: 0.9944

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8620082139968872
Inter Cos: 0.3026496469974518
Norm Quadratic Average: 6.6331305503845215
Nearest Class Center Accuracy: 0.9962

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9164031744003296
Inter Cos: 0.30523401498794556
Norm Quadratic Average: 6.276575088500977
Nearest Class Center Accuracy: 0.9965666666666667

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9310163855552673
Inter Cos: 0.3386133909225464
Norm Quadratic Average: 5.852505207061768
Nearest Class Center Accuracy: 0.99715

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 31.705385208129883
Linear Weight Rank: 4031
Intra Cos: 0.9416166543960571
Inter Cos: 0.2981396019458771
Norm Quadratic Average: 33.56544494628906
Nearest Class Center Accuracy: 0.99795

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 13.037038803100586
Linear Weight Rank: 3669
Intra Cos: 0.9472290277481079
Inter Cos: 0.31302061676979065
Norm Quadratic Average: 29.000368118286133
Nearest Class Center Accuracy: 0.9984333333333333

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 2.6295318603515625
Linear Weight Rank: 10
Intra Cos: 0.9499105215072632
Inter Cos: 0.28778886795043945
Norm Quadratic Average: 25.564910888671875
Nearest Class Center Accuracy: 0.9989

Output Layer:
Intra Cos: 0.9798187613487244
Inter Cos: 0.3293706476688385
Norm Quadratic Average: 24.858154296875
Nearest Class Center Accuracy: 0.9998666666666667

Test Set:
Average Loss: 0.024337236122341347
Accuracy: 0.9946
NC1 Within Class Collapse: 0.6852509379386902
NC2 Equinorm: Features: 0.11993460357189178, Weights: 0.04721539467573166
NC2 Equiangle: Features: 0.24089266459147135, Weights: 0.1382918251885308
NC3 Self-Duality: 0.19652995467185974
NC4 NCC Mismatch: 0.0043999999999999595

Layer 0: Conv2d(1, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10219582915306091
Inter Cos: 0.12048853933811188
Norm Quadratic Average: 23.595195770263672
Nearest Class Center Accuracy: 0.8229

Layer 1: Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.10252322256565094
Inter Cos: 0.1215609461069107
Norm Quadratic Average: 63.526512145996094
Nearest Class Center Accuracy: 0.8236

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1386662721633911
Inter Cos: 0.1462666392326355
Norm Quadratic Average: 67.64350891113281
Nearest Class Center Accuracy: 0.8554

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.14615920186042786
Inter Cos: 0.1533687859773636
Norm Quadratic Average: 90.45113372802734
Nearest Class Center Accuracy: 0.8645

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.22478608787059784
Inter Cos: 0.18458744883537292
Norm Quadratic Average: 61.01241683959961
Nearest Class Center Accuracy: 0.9238

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.25742384791374207
Inter Cos: 0.19490858912467957
Norm Quadratic Average: 51.5419807434082
Nearest Class Center Accuracy: 0.9482

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.30241212248802185
Inter Cos: 0.197758749127388
Norm Quadratic Average: 38.283477783203125
Nearest Class Center Accuracy: 0.9617

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3464224338531494
Inter Cos: 0.20987823605537415
Norm Quadratic Average: 28.237518310546875
Nearest Class Center Accuracy: 0.9691

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.40847325325012207
Inter Cos: 0.20555666089057922
Norm Quadratic Average: 13.053051948547363
Nearest Class Center Accuracy: 0.9797

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5499891042709351
Inter Cos: 0.2865524888038635
Norm Quadratic Average: 10.614986419677734
Nearest Class Center Accuracy: 0.9865

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6147161722183228
Inter Cos: 0.32784295082092285
Norm Quadratic Average: 11.31187915802002
Nearest Class Center Accuracy: 0.9877

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6379754543304443
Inter Cos: 0.3314914107322693
Norm Quadratic Average: 12.071744918823242
Nearest Class Center Accuracy: 0.9889

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6513522863388062
Inter Cos: 0.31135210394859314
Norm Quadratic Average: 7.836592197418213
Nearest Class Center Accuracy: 0.9874

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8609826564788818
Inter Cos: 0.31703487038612366
Norm Quadratic Average: 6.691440105438232
Nearest Class Center Accuracy: 0.9884

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9155340194702148
Inter Cos: 0.30579280853271484
Norm Quadratic Average: 6.332675457000732
Nearest Class Center Accuracy: 0.9889

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9289605021476746
Inter Cos: 0.33904343843460083
Norm Quadratic Average: 5.902068614959717
Nearest Class Center Accuracy: 0.9895

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 31.705385208129883
Linear Weight Rank: 4031
Intra Cos: 0.9385325908660889
Inter Cos: 0.2976340353488922
Norm Quadratic Average: 33.84196853637695
Nearest Class Center Accuracy: 0.9899

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 13.037038803100586
Linear Weight Rank: 3669
Intra Cos: 0.9435173273086548
Inter Cos: 0.3117022216320038
Norm Quadratic Average: 29.236587524414062
Nearest Class Center Accuracy: 0.9909

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 2.6295318603515625
Linear Weight Rank: 10
Intra Cos: 0.9452921152114868
Inter Cos: 0.286397248506546
Norm Quadratic Average: 25.771728515625
Nearest Class Center Accuracy: 0.9916

Output Layer:
Intra Cos: 0.9691815972328186
Inter Cos: 0.33543041348457336
Norm Quadratic Average: 25.045368194580078
Nearest Class Center Accuracy: 0.9932

