Model save path: ./New_Models/bn_False_dataset_MNIST_epochs_200_lr_0.01_model_type_vgg19_rand_seed_979323_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: 0.09116754680871964
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: 0.11930275708436966
Inter Cos: 0.14346779882907867
Norm Quadratic Average: 66.06468200683594
Nearest Class Center Accuracy: 0.79905

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.14122262597084045
Inter Cos: 0.18424905836582184
Norm Quadratic Average: 127.71859741210938
Nearest Class Center Accuracy: 0.7735

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1471170037984848
Inter Cos: 0.19372916221618652
Norm Quadratic Average: 241.8737030029297
Nearest Class Center Accuracy: 0.7709333333333334

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.16703958809375763
Inter Cos: 0.20502957701683044
Norm Quadratic Average: 153.18150329589844
Nearest Class Center Accuracy: 0.7996833333333333

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.16876982152462006
Inter Cos: 0.2326435148715973
Norm Quadratic Average: 114.4035873413086
Nearest Class Center Accuracy: 0.7858

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.16925963759422302
Inter Cos: 0.24964340031147003
Norm Quadratic Average: 110.90860748291016
Nearest Class Center Accuracy: 0.8009833333333334

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.18075613677501678
Inter Cos: 0.265375018119812
Norm Quadratic Average: 135.84341430664062
Nearest Class Center Accuracy: 0.8536166666666667

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.16224047541618347
Inter Cos: 0.30529817938804626
Norm Quadratic Average: 90.52437591552734
Nearest Class Center Accuracy: 0.8666166666666667

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.16591128706932068
Inter Cos: 0.39918750524520874
Norm Quadratic Average: 70.92466735839844
Nearest Class Center Accuracy: 0.879

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.22048784792423248
Inter Cos: 0.4232071042060852
Norm Quadratic Average: 70.24883270263672
Nearest Class Center Accuracy: 0.9117

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3105446696281433
Inter Cos: 0.40637505054473877
Norm Quadratic Average: 63.654052734375
Nearest Class Center Accuracy: 0.92965

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4096710681915283
Inter Cos: 0.23347973823547363
Norm Quadratic Average: 26.194730758666992
Nearest Class Center Accuracy: 0.9128833333333334

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6231781244277954
Inter Cos: 0.35832250118255615
Norm Quadratic Average: 16.73606300354004
Nearest Class Center Accuracy: 0.9115333333333333

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7021219730377197
Inter Cos: 0.3638567626476288
Norm Quadratic Average: 16.480674743652344
Nearest Class Center Accuracy: 0.94555

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7600506544113159
Inter Cos: 0.3543558418750763
Norm Quadratic Average: 16.721065521240234
Nearest Class Center Accuracy: 0.9662166666666666

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.5005018711090088
Linear Weight Rank: 7
Intra Cos: 0.8105210661888123
Inter Cos: 0.3830973207950592
Norm Quadratic Average: 72.12088775634766
Nearest Class Center Accuracy: 0.9818

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.5095304250717163
Linear Weight Rank: 2749
Intra Cos: 0.873727560043335
Inter Cos: 0.3660367727279663
Norm Quadratic Average: 47.95663833618164
Nearest Class Center Accuracy: 0.9887

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.498878836631775
Linear Weight Rank: 9
Intra Cos: 0.8830412030220032
Inter Cos: 0.3178497552871704
Norm Quadratic Average: 27.303098678588867
Nearest Class Center Accuracy: 0.9909833333333333

Output Layer:
Intra Cos: 0.9253031015396118
Inter Cos: 0.36633288860321045
Norm Quadratic Average: 16.630762100219727
Nearest Class Center Accuracy: 0.9914833333333334

Test Set:
Average Loss: 0.040173198188841346
Accuracy: 0.9871
NC1 Within Class Collapse: 1.0472075939178467
NC2 Equinorm: Features: 0.06976301968097687, Weights: 0.036829955875873566
NC2 Equiangle: Features: 0.29133216010199653, Weights: 0.18723186916775172
NC3 Self-Duality: 0.08642935752868652
NC4 NCC Mismatch: 0.005099999999999993

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: 0.13100099563598633
Inter Cos: 0.1570388823747635
Norm Quadratic Average: 66.49056243896484
Nearest Class Center Accuracy: 0.8153

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.15564289689064026
Inter Cos: 0.20178160071372986
Norm Quadratic Average: 128.41014099121094
Nearest Class Center Accuracy: 0.7932

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1618691235780716
Inter Cos: 0.2127116173505783
Norm Quadratic Average: 243.23741149902344
Nearest Class Center Accuracy: 0.7921

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.17835521697998047
Inter Cos: 0.22629736363887787
Norm Quadratic Average: 153.67422485351562
Nearest Class Center Accuracy: 0.8176

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.18173499405384064
Inter Cos: 0.2576765716075897
Norm Quadratic Average: 114.65139770507812
Nearest Class Center Accuracy: 0.8089

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.18210242688655853
Inter Cos: 0.27696141600608826
Norm Quadratic Average: 111.11793518066406
Nearest Class Center Accuracy: 0.8277

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.19464465975761414
Inter Cos: 0.2924773693084717
Norm Quadratic Average: 136.36550903320312
Nearest Class Center Accuracy: 0.8703

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.17233163118362427
Inter Cos: 0.3071286082267761
Norm Quadratic Average: 90.92716217041016
Nearest Class Center Accuracy: 0.8802

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.1708284467458725
Inter Cos: 0.3991716504096985
Norm Quadratic Average: 71.3724365234375
Nearest Class Center Accuracy: 0.8854

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2233445793390274
Inter Cos: 0.41499024629592896
Norm Quadratic Average: 70.89205932617188
Nearest Class Center Accuracy: 0.9126

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.30910593271255493
Inter Cos: 0.3922964930534363
Norm Quadratic Average: 64.37467956542969
Nearest Class Center Accuracy: 0.9316

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.40789398550987244
Inter Cos: 0.22961106896400452
Norm Quadratic Average: 26.468040466308594
Nearest Class Center Accuracy: 0.9201

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6213482618331909
Inter Cos: 0.3866199553012848
Norm Quadratic Average: 16.927274703979492
Nearest Class Center Accuracy: 0.9188

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6994701623916626
Inter Cos: 0.3902052044868469
Norm Quadratic Average: 16.708131790161133
Nearest Class Center Accuracy: 0.9469

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7538507580757141
Inter Cos: 0.3786756098270416
Norm Quadratic Average: 16.97623062133789
Nearest Class Center Accuracy: 0.964

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 1.5005018711090088
Linear Weight Rank: 7
Intra Cos: 0.8139722347259521
Inter Cos: 0.399509072303772
Norm Quadratic Average: 73.31278228759766
Nearest Class Center Accuracy: 0.9762

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 1.5095304250717163
Linear Weight Rank: 2749
Intra Cos: 0.8734375238418579
Inter Cos: 0.3706660270690918
Norm Quadratic Average: 48.79044723510742
Nearest Class Center Accuracy: 0.983

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 1.498878836631775
Linear Weight Rank: 9
Intra Cos: 0.8781107068061829
Inter Cos: 0.3189999461174011
Norm Quadratic Average: 27.7851619720459
Nearest Class Center Accuracy: 0.9864

Output Layer:
Intra Cos: 0.9155706167221069
Inter Cos: 0.36607033014297485
Norm Quadratic Average: 16.93719482421875
Nearest Class Center Accuracy: 0.9871

