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_0.0005.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: 0.0949498787522316
Inter Cos: 0.11457142233848572
Norm Quadratic Average: 60.817039489746094
Nearest Class Center Accuracy: 0.8100333333333334

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.12330310046672821
Inter Cos: 0.14304092526435852
Norm Quadratic Average: 75.80675506591797
Nearest Class Center Accuracy: 0.8231833333333334

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.12867577373981476
Inter Cos: 0.15277916193008423
Norm Quadratic Average: 118.9881362915039
Nearest Class Center Accuracy: 0.8327

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.20374436676502228
Inter Cos: 0.17474576830863953
Norm Quadratic Average: 91.52490997314453
Nearest Class Center Accuracy: 0.8899166666666667

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.23512974381446838
Inter Cos: 0.17395983636379242
Norm Quadratic Average: 82.14038848876953
Nearest Class Center Accuracy: 0.9221

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2647215723991394
Inter Cos: 0.17576923966407776
Norm Quadratic Average: 66.26206970214844
Nearest Class Center Accuracy: 0.9431166666666667

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.29340800642967224
Inter Cos: 0.1634538173675537
Norm Quadratic Average: 48.31075668334961
Nearest Class Center Accuracy: 0.9546833333333333

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3227543532848358
Inter Cos: 0.17252030968666077
Norm Quadratic Average: 20.34459686279297
Nearest Class Center Accuracy: 0.9739

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5065683722496033
Inter Cos: 0.30278685688972473
Norm Quadratic Average: 12.822885513305664
Nearest Class Center Accuracy: 0.98515

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5911872982978821
Inter Cos: 0.394174188375473
Norm Quadratic Average: 11.688183784484863
Nearest Class Center Accuracy: 0.9888166666666667

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6337262392044067
Inter Cos: 0.4164656400680542
Norm Quadratic Average: 11.926386833190918
Nearest Class Center Accuracy: 0.9914166666666666

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6895937919616699
Inter Cos: 0.3425731658935547
Norm Quadratic Average: 7.992324352264404
Nearest Class Center Accuracy: 0.9923333333333333

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8392667770385742
Inter Cos: 0.2952883541584015
Norm Quadratic Average: 7.080578804016113
Nearest Class Center Accuracy: 0.9966166666666667

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.881744384765625
Inter Cos: 0.3437685966491699
Norm Quadratic Average: 7.3279924392700195
Nearest Class Center Accuracy: 0.9973666666666666

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9033036231994629
Inter Cos: 0.38217025995254517
Norm Quadratic Average: 7.164085388183594
Nearest Class Center Accuracy: 0.9977833333333334

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 14.706066131591797
Linear Weight Rank: 4031
Intra Cos: 0.9160704016685486
Inter Cos: 0.35165464878082275
Norm Quadratic Average: 38.66277313232422
Nearest Class Center Accuracy: 0.9986

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 6.341848850250244
Linear Weight Rank: 3667
Intra Cos: 0.9336229562759399
Inter Cos: 0.3532003164291382
Norm Quadratic Average: 32.49604034423828
Nearest Class Center Accuracy: 0.9992

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 2.3163464069366455
Linear Weight Rank: 10
Intra Cos: 0.9393674731254578
Inter Cos: 0.3179698586463928
Norm Quadratic Average: 26.65814971923828
Nearest Class Center Accuracy: 0.99955

Output Layer:
Intra Cos: 0.9716046452522278
Inter Cos: 0.3468373417854309
Norm Quadratic Average: 24.29624366760254
Nearest Class Center Accuracy: 0.9999666666666667

Test Set:
Average Loss: 0.022494620684997062
Accuracy: 0.994
NC1 Within Class Collapse: 0.8892353773117065
NC2 Equinorm: Features: 0.09363528341054916, Weights: 0.02110479399561882
NC2 Equiangle: Features: 0.25746919843885635, Weights: 0.15773642857869466
NC3 Self-Duality: 0.13593904674053192
NC4 NCC Mismatch: 0.0042999999999999705

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: 0.10620030760765076
Inter Cos: 0.12621361017227173
Norm Quadratic Average: 60.9571533203125
Nearest Class Center Accuracy: 0.8229

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.13733938336372375
Inter Cos: 0.15684062242507935
Norm Quadratic Average: 75.73567199707031
Nearest Class Center Accuracy: 0.8365

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1427295207977295
Inter Cos: 0.16737669706344604
Norm Quadratic Average: 118.96577453613281
Nearest Class Center Accuracy: 0.8464

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.21964053809642792
Inter Cos: 0.19060173630714417
Norm Quadratic Average: 91.43585968017578
Nearest Class Center Accuracy: 0.9009

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2509627342224121
Inter Cos: 0.18422795832157135
Norm Quadratic Average: 82.08318328857422
Nearest Class Center Accuracy: 0.9317

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.27930936217308044
Inter Cos: 0.16940733790397644
Norm Quadratic Average: 66.25818634033203
Nearest Class Center Accuracy: 0.9506

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.30618003010749817
Inter Cos: 0.16216328740119934
Norm Quadratic Average: 48.346858978271484
Nearest Class Center Accuracy: 0.9582

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.33622559905052185
Inter Cos: 0.18678441643714905
Norm Quadratic Average: 20.388410568237305
Nearest Class Center Accuracy: 0.9744

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5279446244239807
Inter Cos: 0.32114747166633606
Norm Quadratic Average: 12.875993728637695
Nearest Class Center Accuracy: 0.9837

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6130521893501282
Inter Cos: 0.4156070947647095
Norm Quadratic Average: 11.752185821533203
Nearest Class Center Accuracy: 0.9864

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6539101600646973
Inter Cos: 0.43879076838493347
Norm Quadratic Average: 11.999726295471191
Nearest Class Center Accuracy: 0.9874

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7096837162971497
Inter Cos: 0.35954010486602783
Norm Quadratic Average: 8.053434371948242
Nearest Class Center Accuracy: 0.9867

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8506491780281067
Inter Cos: 0.2990996539592743
Norm Quadratic Average: 7.147833347320557
Nearest Class Center Accuracy: 0.9903

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8873154520988464
Inter Cos: 0.34216195344924927
Norm Quadratic Average: 7.396860599517822
Nearest Class Center Accuracy: 0.9908

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9062697291374207
Inter Cos: 0.38028648495674133
Norm Quadratic Average: 7.225881576538086
Nearest Class Center Accuracy: 0.9914

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 14.706066131591797
Linear Weight Rank: 4031
Intra Cos: 0.918082594871521
Inter Cos: 0.35016557574272156
Norm Quadratic Average: 38.9546012878418
Nearest Class Center Accuracy: 0.9923

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 6.341848850250244
Linear Weight Rank: 3667
Intra Cos: 0.9317988753318787
Inter Cos: 0.35204988718032837
Norm Quadratic Average: 32.73427963256836
Nearest Class Center Accuracy: 0.9924

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 2.3163464069366455
Linear Weight Rank: 10
Intra Cos: 0.9384002685546875
Inter Cos: 0.31757405400276184
Norm Quadratic Average: 26.85260772705078
Nearest Class Center Accuracy: 0.993

Output Layer:
Intra Cos: 0.9665785431861877
Inter Cos: 0.34739950299263
Norm Quadratic Average: 24.46841049194336
Nearest Class Center Accuracy: 0.9943

