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_0.01.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: 0.11910362541675568
Inter Cos: 0.1427103877067566
Norm Quadratic Average: 64.89385223388672
Nearest Class Center Accuracy: 0.79875

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1417364776134491
Inter Cos: 0.1834084689617157
Norm Quadratic Average: 122.3777847290039
Nearest Class Center Accuracy: 0.7737333333333334

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.14488649368286133
Inter Cos: 0.189635768532753
Norm Quadratic Average: 209.4392852783203
Nearest Class Center Accuracy: 0.7819333333333334

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1716318279504776
Inter Cos: 0.19422562420368195
Norm Quadratic Average: 112.38860321044922
Nearest Class Center Accuracy: 0.837

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.19578056037425995
Inter Cos: 0.19926230609416962
Norm Quadratic Average: 58.856998443603516
Nearest Class Center Accuracy: 0.85675

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.22289597988128662
Inter Cos: 0.195609450340271
Norm Quadratic Average: 39.38698959350586
Nearest Class Center Accuracy: 0.8655333333333334

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2517332136631012
Inter Cos: 0.2539815306663513
Norm Quadratic Average: 41.445438385009766
Nearest Class Center Accuracy: 0.9109833333333334

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2595212161540985
Inter Cos: 0.34010928869247437
Norm Quadratic Average: 27.64316749572754
Nearest Class Center Accuracy: 0.90155

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.33026114106178284
Inter Cos: 0.3313048779964447
Norm Quadratic Average: 25.02851104736328
Nearest Class Center Accuracy: 0.8904833333333333

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4154949486255646
Inter Cos: 0.3378141522407532
Norm Quadratic Average: 28.471942901611328
Nearest Class Center Accuracy: 0.9212166666666667

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.524786651134491
Inter Cos: 0.40024060010910034
Norm Quadratic Average: 32.87713623046875
Nearest Class Center Accuracy: 0.9467166666666667

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4707413613796234
Inter Cos: 0.31121012568473816
Norm Quadratic Average: 19.267282485961914
Nearest Class Center Accuracy: 0.9380833333333334

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6701255440711975
Inter Cos: 0.4865618944168091
Norm Quadratic Average: 15.67501449584961
Nearest Class Center Accuracy: 0.9500666666666666

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6853611469268799
Inter Cos: 0.4920848309993744
Norm Quadratic Average: 17.040462493896484
Nearest Class Center Accuracy: 0.96525

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6872366666793823
Inter Cos: 0.5497612357139587
Norm Quadratic Average: 18.98491096496582
Nearest Class Center Accuracy: 0.9767666666666667

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.4710718393325806
Linear Weight Rank: 9
Intra Cos: 0.7846218943595886
Inter Cos: 0.524939775466919
Norm Quadratic Average: 82.22821807861328
Nearest Class Center Accuracy: 0.9868833333333333

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.4833353757858276
Linear Weight Rank: 2769
Intra Cos: 0.839618980884552
Inter Cos: 0.490244060754776
Norm Quadratic Average: 52.37178039550781
Nearest Class Center Accuracy: 0.9901166666666666

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.4700899124145508
Linear Weight Rank: 9
Intra Cos: 0.8683260083198547
Inter Cos: 0.4154845178127289
Norm Quadratic Average: 29.14435577392578
Nearest Class Center Accuracy: 0.9915666666666667

Output Layer:
Intra Cos: 0.9145928025245667
Inter Cos: 0.4719019830226898
Norm Quadratic Average: 17.591459274291992
Nearest Class Center Accuracy: 0.9925166666666667

Test Set:
Average Loss: 0.03942081452384591
Accuracy: 0.9874
NC1 Within Class Collapse: 1.29392409324646
NC2 Equinorm: Features: 0.10332590341567993, Weights: 0.048705294728279114
NC2 Equiangle: Features: 0.30398625267876517, Weights: 0.1938870324028863
NC3 Self-Duality: 0.11693862825632095
NC4 NCC Mismatch: 0.005700000000000038

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: 0.1307733952999115
Inter Cos: 0.1562071144580841
Norm Quadratic Average: 65.3096923828125
Nearest Class Center Accuracy: 0.8151

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.15611839294433594
Inter Cos: 0.20076772570610046
Norm Quadratic Average: 123.03691864013672
Nearest Class Center Accuracy: 0.794

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.15973111987113953
Inter Cos: 0.2080150991678238
Norm Quadratic Average: 210.5480499267578
Nearest Class Center Accuracy: 0.8038

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1822652816772461
Inter Cos: 0.21355657279491425
Norm Quadratic Average: 112.73381042480469
Nearest Class Center Accuracy: 0.8531

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.20816543698310852
Inter Cos: 0.22173605859279633
Norm Quadratic Average: 58.971187591552734
Nearest Class Center Accuracy: 0.872

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.23870772123336792
Inter Cos: 0.22201372683048248
Norm Quadratic Average: 39.34560775756836
Nearest Class Center Accuracy: 0.8851

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2676154375076294
Inter Cos: 0.2572799026966095
Norm Quadratic Average: 41.46616744995117
Nearest Class Center Accuracy: 0.9213

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2744925916194916
Inter Cos: 0.36655670404434204
Norm Quadratic Average: 27.713401794433594
Nearest Class Center Accuracy: 0.911

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3447152376174927
Inter Cos: 0.33211973309516907
Norm Quadratic Average: 25.125442504882812
Nearest Class Center Accuracy: 0.8985

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4290342628955841
Inter Cos: 0.36492919921875
Norm Quadratic Average: 28.672582626342773
Nearest Class Center Accuracy: 0.9235

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5351879000663757
Inter Cos: 0.4250030517578125
Norm Quadratic Average: 33.221397399902344
Nearest Class Center Accuracy: 0.9478

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4942914545536041
Inter Cos: 0.3308236002922058
Norm Quadratic Average: 19.47272491455078
Nearest Class Center Accuracy: 0.9379

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6777350306510925
Inter Cos: 0.5051680207252502
Norm Quadratic Average: 15.873238563537598
Nearest Class Center Accuracy: 0.9492

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.688772439956665
Inter Cos: 0.5042073726654053
Norm Quadratic Average: 17.26919937133789
Nearest Class Center Accuracy: 0.9609

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6889109015464783
Inter Cos: 0.5529895424842834
Norm Quadratic Average: 19.247159957885742
Nearest Class Center Accuracy: 0.9716

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.4710718393325806
Linear Weight Rank: 9
Intra Cos: 0.784162700176239
Inter Cos: 0.5467274188995361
Norm Quadratic Average: 83.49008178710938
Nearest Class Center Accuracy: 0.9818

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.4833353757858276
Linear Weight Rank: 2769
Intra Cos: 0.8391408920288086
Inter Cos: 0.507502019405365
Norm Quadratic Average: 53.190162658691406
Nearest Class Center Accuracy: 0.9847

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.4700899124145508
Linear Weight Rank: 9
Intra Cos: 0.8664683699607849
Inter Cos: 0.43053770065307617
Norm Quadratic Average: 29.599090576171875
Nearest Class Center Accuracy: 0.986

Output Layer:
Intra Cos: 0.905149519443512
Inter Cos: 0.4869433641433716
Norm Quadratic Average: 17.879276275634766
Nearest Class Center Accuracy: 0.987

