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_0.0001.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: 0.0794934630393982
Inter Cos: 0.096687912940979
Norm Quadratic Average: 54.966033935546875
Nearest Class Center Accuracy: 0.8214666666666667

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.12620307505130768
Inter Cos: 0.12492929399013519
Norm Quadratic Average: 47.73604202270508
Nearest Class Center Accuracy: 0.8649166666666667

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.14564622938632965
Inter Cos: 0.1418170928955078
Norm Quadratic Average: 57.580162048339844
Nearest Class Center Accuracy: 0.8811

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.2283453345298767
Inter Cos: 0.16077886521816254
Norm Quadratic Average: 37.84239196777344
Nearest Class Center Accuracy: 0.9315333333333333

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.26325222849845886
Inter Cos: 0.18660227954387665
Norm Quadratic Average: 37.36112976074219
Nearest Class Center Accuracy: 0.9496

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3021242916584015
Inter Cos: 0.2074335664510727
Norm Quadratic Average: 34.450767517089844
Nearest Class Center Accuracy: 0.9608

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.33741527795791626
Inter Cos: 0.20672312378883362
Norm Quadratic Average: 31.514301300048828
Nearest Class Center Accuracy: 0.9660833333333333

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.39189600944519043
Inter Cos: 0.21982735395431519
Norm Quadratic Average: 16.283222198486328
Nearest Class Center Accuracy: 0.9842

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5119377970695496
Inter Cos: 0.2259676158428192
Norm Quadratic Average: 13.345317840576172
Nearest Class Center Accuracy: 0.9909

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6129299402236938
Inter Cos: 0.24904902279376984
Norm Quadratic Average: 12.125404357910156
Nearest Class Center Accuracy: 0.9933666666666666

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6934292912483215
Inter Cos: 0.2613312900066376
Norm Quadratic Average: 11.38861083984375
Nearest Class Center Accuracy: 0.9937166666666667

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7489929795265198
Inter Cos: 0.19112539291381836
Norm Quadratic Average: 7.439281463623047
Nearest Class Center Accuracy: 0.9916833333333334

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.8796684145927429
Inter Cos: 0.24196548759937286
Norm Quadratic Average: 6.623340606689453
Nearest Class Center Accuracy: 0.9924333333333333

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9272857308387756
Inter Cos: 0.26121610403060913
Norm Quadratic Average: 6.1869354248046875
Nearest Class Center Accuracy: 0.99285

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.941560685634613
Inter Cos: 0.33144426345825195
Norm Quadratic Average: 5.840835094451904
Nearest Class Center Accuracy: 0.9935833333333334

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 68.79899597167969
Linear Weight Rank: 4031
Intra Cos: 0.9521646499633789
Inter Cos: 0.3490370512008667
Norm Quadratic Average: 34.21935272216797
Nearest Class Center Accuracy: 0.9947833333333334

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 27.935501098632812
Linear Weight Rank: 3670
Intra Cos: 0.9529572129249573
Inter Cos: 0.3648226261138916
Norm Quadratic Average: 28.71565818786621
Nearest Class Center Accuracy: 0.9959333333333333

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 3.423367738723755
Linear Weight Rank: 10
Intra Cos: 0.9543848633766174
Inter Cos: 0.3397521376609802
Norm Quadratic Average: 26.003175735473633
Nearest Class Center Accuracy: 0.9975833333333334

Output Layer:
Intra Cos: 0.9762853980064392
Inter Cos: 0.36165332794189453
Norm Quadratic Average: 28.00743293762207
Nearest Class Center Accuracy: 0.9998833333333333

Test Set:
Average Loss: 0.024755040819455457
Accuracy: 0.995
NC1 Within Class Collapse: 0.49028390645980835
NC2 Equinorm: Features: 0.13125932216644287, Weights: 0.1901218742132187
NC2 Equiangle: Features: 0.22837373945448133, Weights: 0.15650172763400608
NC3 Self-Duality: 0.43765988945961
NC4 NCC Mismatch: 0.006700000000000039

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: 0.08907806128263474
Inter Cos: 0.10047689825296402
Norm Quadratic Average: 54.93544387817383
Nearest Class Center Accuracy: 0.8353

Layer 2: Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.13818007707595825
Inter Cos: 0.12574663758277893
Norm Quadratic Average: 47.56037139892578
Nearest Class Center Accuracy: 0.8768

Layer 3: Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.15848328173160553
Inter Cos: 0.14206179976463318
Norm Quadratic Average: 57.46043395996094
Nearest Class Center Accuracy: 0.8915

Layer 4: Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.24578386545181274
Inter Cos: 0.17520570755004883
Norm Quadratic Average: 37.738243103027344
Nearest Class Center Accuracy: 0.9388

Layer 5: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.27929338812828064
Inter Cos: 0.20274780690670013
Norm Quadratic Average: 37.279998779296875
Nearest Class Center Accuracy: 0.9546

Layer 6: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.31783708930015564
Inter Cos: 0.2250516265630722
Norm Quadratic Average: 34.387149810791016
Nearest Class Center Accuracy: 0.9657

Layer 7: Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.3528982996940613
Inter Cos: 0.2241690456867218
Norm Quadratic Average: 31.467077255249023
Nearest Class Center Accuracy: 0.9682

Layer 8: Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.4021231234073639
Inter Cos: 0.22417877614498138
Norm Quadratic Average: 16.2851505279541
Nearest Class Center Accuracy: 0.983

Layer 9: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.5223919153213501
Inter Cos: 0.23928160965442657
Norm Quadratic Average: 13.3726167678833
Nearest Class Center Accuracy: 0.9872

Layer 10: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.619248628616333
Inter Cos: 0.26044517755508423
Norm Quadratic Average: 12.158818244934082
Nearest Class Center Accuracy: 0.9881

Layer 11: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.6969090700149536
Inter Cos: 0.2701810300350189
Norm Quadratic Average: 11.424110412597656
Nearest Class Center Accuracy: 0.989

Layer 12: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.7477623224258423
Inter Cos: 0.18727481365203857
Norm Quadratic Average: 7.465482234954834
Nearest Class Center Accuracy: 0.9855

Layer 13: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.875409722328186
Inter Cos: 0.23859046399593353
Norm Quadratic Average: 6.648075580596924
Nearest Class Center Accuracy: 0.9858

Layer 14: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9214282631874084
Inter Cos: 0.2608563005924225
Norm Quadratic Average: 6.207406520843506
Nearest Class Center Accuracy: 0.9866

Layer 15: Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
Intra Cos: 0.9348257184028625
Inter Cos: 0.3299356997013092
Norm Quadratic Average: 5.856541156768799
Nearest Class Center Accuracy: 0.9874

Layer 16: Linear(in_features=25088, out_features=4096, bias=True)
Linear Weight Norm: 68.79899597167969
Linear Weight Rank: 4031
Intra Cos: 0.9455389380455017
Inter Cos: 0.34869569540023804
Norm Quadratic Average: 34.29499053955078
Nearest Class Center Accuracy: 0.9888

Layer 17: Linear(in_features=4096, out_features=4096, bias=True)
Linear Weight Norm: 27.935501098632812
Linear Weight Rank: 3670
Intra Cos: 0.9462071061134338
Inter Cos: 0.36405861377716064
Norm Quadratic Average: 28.778724670410156
Nearest Class Center Accuracy: 0.9895

Layer 18: Linear(in_features=4096, out_features=10, bias=True)
Linear Weight Norm: 3.423367738723755
Linear Weight Rank: 10
Intra Cos: 0.9473405480384827
Inter Cos: 0.33829832077026367
Norm Quadratic Average: 26.061525344848633
Nearest Class Center Accuracy: 0.9909

Output Layer:
Intra Cos: 0.9686247706413269
Inter Cos: 0.3636656105518341
Norm Quadratic Average: 28.085224151611328
Nearest Class Center Accuracy: 0.9952

