Model save path: /content/drive/MyDrive/Neural Collapse/New_Models/bn_True_dataset_mlp3_epochs_100_lr_0.001_model_depth_MLP_3_model_type_MLP_rand_seed_314159_weight_decay_0.02.pth.tar
Training Set:
Layer 0: Linear(in_features=16, out_features=200, bias=True)
Linear Weight Norm: 3.1915433406829834
Linear Weight Rank: 16
Intra Cos: 0.03299567848443985
Inter Cos: -0.006222445517778397
Norm Quadratic Average: 3.9938883781433105
Nearest Class Center Accuracy: 0.583375

Layer 1: Linear(in_features=200, out_features=200, bias=True)
Linear Weight Norm: 3.299037456512451
Linear Weight Rank: 179
Intra Cos: 0.029276756569743156
Inter Cos: -0.004401979967951775
Norm Quadratic Average: 5.388193607330322
Nearest Class Center Accuracy: 0.60225

Layer 2: Linear(in_features=200, out_features=4, bias=True)
Linear Weight Norm: 2.9813499450683594
Linear Weight Rank: 4
Intra Cos: 0.11351806670427322
Inter Cos: -0.0281389020383358
Norm Quadratic Average: 6.062290191650391
Nearest Class Center Accuracy: 0.94325

Output Layer:
Intra Cos: 0.6748807430267334
Inter Cos: -0.15942323207855225
Norm Quadratic Average: 4.430618762969971
Nearest Class Center Accuracy: 0.98175

Test Set:
Average Loss: 0.5622610206604004
Accuracy: 0.767
NC1 Within Class Collapse: 3.377164363861084
NC2 Equinorm: Features: 0.1114029735326767, Weights: 0.0447852797806263
NC2 Equiangle: Features: 0.12533365686734518, Weights: 0.021201983094215393
NC3 Self-Duality: 0.2484211027622223
NC4 NCC Mismatch: 0.08850000000000002

Layer 0: Linear(in_features=16, out_features=200, bias=True)
Linear Weight Norm: 3.1915433406829834
Linear Weight Rank: 16
Intra Cos: 0.030431468039751053
Inter Cos: -0.0045497240498661995
Norm Quadratic Average: 3.9951601028442383
Nearest Class Center Accuracy: 0.583

Layer 1: Linear(in_features=200, out_features=200, bias=True)
Linear Weight Norm: 3.299037456512451
Linear Weight Rank: 179
Intra Cos: 0.027937013655900955
Inter Cos: -0.0036807931028306484
Norm Quadratic Average: 5.406029224395752
Nearest Class Center Accuracy: 0.5915

Layer 2: Linear(in_features=200, out_features=4, bias=True)
Linear Weight Norm: 2.9813499450683594
Linear Weight Rank: 4
Intra Cos: 0.07996249943971634
Inter Cos: -0.008352609351277351
Norm Quadratic Average: 6.075219631195068
Nearest Class Center Accuracy: 0.766

Output Layer:
Intra Cos: 0.43756750226020813
Inter Cos: -0.05044006556272507
Norm Quadratic Average: 4.325603485107422
Nearest Class Center Accuracy: 0.7675

