Arch: resnet50_pt
Bs trn: 128
Bs val: 128
Hidden dim: 256
Dataset: celebA
Resample class: 
Slice with: rep
Rep cluster method: gmm
Num anchor: 32
Num positive: 32
Num negative: 32
Num negative easy: 0
Weight anc by loss: False
Weight pos by loss: False
Weight neg by loss: False
Anc loss temp: 0.5
Pos loss temp: 0.5
Neg loss temp: 0.5
Data wide pos: False
Target sample ratio: 1
Balance targets: False
Additional negatives: False
Hard negative factor: 0
Full contrastive: False
Train encoder: False
No projection head: False
Projection dim: 128
Batch factor: None
Temperature: 0.05
Single pos: False
Supervised linear scale up: False
Supervised update delay: 0
Contrastive weight: 0.5
Classifier update interval: 8
Optim: sgd
Max epoch: 50
Lr: 0.0001
Momentum: 0.9
Weight decay: 0.1
Weight decay c: 0.1
Stopping window: 30
Load encoder: 
Freeze encoder: False
Finetune epochs: 0
Clip grad norm: False
Lr scheduler classifier: 
Lr scheduler: 
Grad clip grad norm: False
Erm: False
Erm only: False
Pretrained spurious path: ./model/celebA/config/stage_one_erm/seed34/stage_one_erm_model_b_epoch0_seed34.pt
Max epoch s: 1
Bs trn s: 32
Lr s: 0.001
Momentum s: 0.9
Weight decay s: 0.0005
Slice temp: 10
Log loss interval: 10
Checkpoint interval: 50
Grad checkpoint interval: 50
Log visual interval: 100
Log grad visual interval: 50
Verbose: True
Seed: 34
Replicate: 0
No cuda: False
Resume: False
New slice: False
Num workers: 12
Evaluate: False
Data cmap: hsv
Test cmap: 
P correlation: 0.9
P corr by class: None
Train classes: ['blond', 'nonblond']
Train class ratios: None
Test shift: random
Flipped: False
Q: 0.7
Pretrained bmodel: True
Cosine: False
Exp: ours
Tau: 1.1
Gamma: None
Remove label noise: False
Model for remove samples: 
Remove ratio: 0.03
Supervised contrast: True
Prioritize spurious pos: False
Contrastive type: cnc
Compute auroc: False
Model type: resnet50_pt_cnc
Criterion: cross_entropy
Pretrained: False
Max grad norm: 1.0
Adam epsilon: 1e-08
Warmup steps: 0
Max grad norm s: 1.0
Adam epsilon s: 1e-08
Warmup steps s: 0
Grad max grad norm: 1.0
Grad adam epsilon: 1e-08
Grad warmup steps: 0
Device: cuda
Img file type: .png
Display image: False
Image path: ./images/celebA/celebA/config/contrastive_umaps
Log interval: 1
Log path: ./logs/celebA/config
Results path: ./results/celebA/config
Model path: ./model/celebA/config
Loss factor: 1
Supersample labels: False
Subsample labels: False
Weigh slice samples by loss: True
Val split: 0.2
Spurious train split: 0.2
Subsample groups: False
Train method: sc
Max robust acc: -1
Max robust epoch: -1
Max robust group acc: (None, None)
Root dir: ./datasets/data/CelebA/
Target name: Blond_Hair
Confounder names: ['Male']
Image mean: 0.449
Image std: 0.226
Augment data: False
Task: celebA
Num classes: 2
Experiment configs: config
Experiment name: cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=None-cw=0.5-sud=0-me=50-bst=128-o=sgd-lr=0.0001-mo=0.9-wd=0.1-wdc=0.1-spur-me=1-bst=32-lr=0.001-mo=0.9-wd=0.0005-sts=0.2-s=34-r=0
Mi resampled: None

Loading checkpoints for train split:
[-1 -1 -1 ... -1 -1 -1]
<class 'numpy.ndarray'>
[0 1 2 3] [71629 66874 22880  1387]
Loading checkpoints for val split:
[-1 -1 -1 ... -1  1 -1]
<class 'numpy.ndarray'>
[0 1 2 3] [8535 8276 2874  182]
Loading checkpoints for test split:
[-1 -1 -1 ... -1 -1  1]
<class 'numpy.ndarray'>
[0 1 2 3] [9767 7535 2480  180]
Train dataset:
    Blond_Hair = 0, Male = 0 : n = 71629
    Blond_Hair = 0, Male = 1 : n = 66874
    Blond_Hair = 1, Male = 0 : n = 22880
    Blond_Hair = 1, Male = 1 : n = 1387
Val dataset:
    Blond_Hair = 0, Male = 0 : n = 8535
    Blond_Hair = 0, Male = 1 : n = 8276
    Blond_Hair = 1, Male = 0 : n = 2874
    Blond_Hair = 1, Male = 1 : n = 182
Test dataset:
    Blond_Hair = 0, Male = 0 : n = 9767
    Blond_Hair = 0, Male = 1 : n = 7535
    Blond_Hair = 1, Male = 0 : n = 2480
    Blond_Hair = 1, Male = 1 : n = 180
------------------------
> Loading spurious model
------------------------
Pretrained model loaded from ./model/celebA/config/stage_one_erm/seed34/stage_one_erm_model_b_epoch0_seed34.pt
======
# Calculate probability ...
======
======
p_y_a:  tensor([[0.8217, 0.0293],
        [0.1305, 0.0186]])
p_y:  tensor([0.8509, 0.1491])
# Load biased model ...
======
Epoch:   1 | Train Loss: 0.002 | Train Acc: 89.678 | Val Loss: 0.003 | Val Acc: 83.843
Training:
Accuracies by groups:
0, 0  acc: 10744 / 20335 =  52.835
0, 1  acc:  5895 /  9623 =  61.259
1, 0  acc: 121696 / 124440 =  97.795
1, 1  acc:  7634 /  8372 =  91.185
--------------------------------------
Average acc: 145969 / 162770 =  89.678
Robust  acc: 10744 / 20335 =  52.835
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6555 /  8535 =  76.801
0, 1  acc:  7103 /  8276 =  85.826
1, 0  acc:  2834 /  2874 =  98.608
1, 1  acc:   165 /   182 =  90.659
------------------------------------
Average acc: 16657 / 19867 =  83.843
Robust  acc:  6555 /  8535 =  76.801
------------------------------------
New max robust acc: 76.80140597539543
debias model - Saving best checkpoint at epoch 0
replace: True
-> Updating checkpoint debias-wga-best_seed34.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed34.pt
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 85.472
Robust Acc: 81.560 | Best Acc: 98.266
-------------------------------------
Training, Epoch 0:
Accuracies by groups:
0, 0  acc:  7966 /  9767 =  81.560
0, 1  acc:  6496 /  7535 =  86.211
1, 0  acc:  2437 /  2480 =  98.266
1, 1  acc:   163 /   180 =  90.556
------------------------------------
Average acc: 17062 / 19962 =  85.472
Robust  acc:  7966 /  9767 =  81.560
------------------------------------
Accuracies by groups:
0, 0  acc:  7966 /  9767 =  81.560
0, 1  acc:  6496 /  7535 =  86.211
1, 0  acc:  2437 /  2480 =  98.266
1, 1  acc:   163 /   180 =  90.556
------------------------------------
Average acc: 17062 / 19962 =  85.472
Robust  acc:  7966 /  9767 =  81.560
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7966 /  9767 =  81.560
0, 1  acc:  6496 /  7535 =  86.211
1, 0  acc:  2437 /  2480 =  98.266
1, 1  acc:   163 /   180 =  90.556
------------------------------------
Average acc: 17062 / 19962 =  85.472
Robust  acc:  7966 /  9767 =  81.560
------------------------------------
Epoch:   2 | Train Loss: 0.001 | Train Acc: 93.466 | Val Loss: 0.002 | Val Acc: 86.772
Training:
Accuracies by groups:
0, 0  acc: 14138 / 20288 =  69.687
0, 1  acc:  7739 /  9557 =  80.977
1, 0  acc: 122878 / 124735 =  98.511
1, 1  acc:  7380 /  8190 =  90.110
--------------------------------------
Average acc: 152135 / 162770 =  93.466
Robust  acc: 14138 / 20288 =  69.687
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6786 /  8535 =  79.508
0, 1  acc:  7452 /  8276 =  90.043
1, 0  acc:  2834 /  2874 =  98.608
1, 1  acc:   167 /   182 =  91.758
------------------------------------
Average acc: 17239 / 19867 =  86.772
Robust  acc:  6786 /  8535 =  79.508
------------------------------------
New max robust acc: 79.5079086115993
debias model - Saving best checkpoint at epoch 1
replace: True
-> Updating checkpoint debias-wga-best_seed34.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed34.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 88.082
Robust Acc: 83.864 | Best Acc: 98.145
-------------------------------------
Training, Epoch 1:
Accuracies by groups:
0, 0  acc:  8191 /  9767 =  83.864
0, 1  acc:  6797 /  7535 =  90.206
1, 0  acc:  2434 /  2480 =  98.145
1, 1  acc:   161 /   180 =  89.444
------------------------------------
Average acc: 17583 / 19962 =  88.082
Robust  acc:  8191 /  9767 =  83.864
------------------------------------
Accuracies by groups:
0, 0  acc:  8191 /  9767 =  83.864
0, 1  acc:  6797 /  7535 =  90.206
1, 0  acc:  2434 /  2480 =  98.145
1, 1  acc:   161 /   180 =  89.444
------------------------------------
Average acc: 17583 / 19962 =  88.082
Robust  acc:  8191 /  9767 =  83.864
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8191 /  9767 =  83.864
0, 1  acc:  6797 /  7535 =  90.206
1, 0  acc:  2434 /  2480 =  98.145
1, 1  acc:   161 /   180 =  89.444
------------------------------------
Average acc: 17583 / 19962 =  88.082
Robust  acc:  8191 /  9767 =  83.864
------------------------------------
Epoch:   3 | Train Loss: 0.001 | Train Acc: 94.350 | Val Loss: 0.002 | Val Acc: 89.148
Training:
Accuracies by groups:
0, 0  acc: 14827 / 20407 =  72.656
0, 1  acc:  8169 /  9489 =  86.089
1, 0  acc: 123109 / 124733 =  98.698
1, 1  acc:  7468 /  8141 =  91.733
--------------------------------------
Average acc: 153573 / 162770 =  94.350
Robust  acc: 14827 / 20407 =  72.656
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7010 /  8535 =  82.132
0, 1  acc:  7715 /  8276 =  93.221
1, 0  acc:  2822 /  2874 =  98.191
1, 1  acc:   164 /   182 =  90.110
------------------------------------
Average acc: 17711 / 19867 =  89.148
Robust  acc:  7010 /  8535 =  82.132
------------------------------------
New max robust acc: 82.13239601640304
debias model - Saving best checkpoint at epoch 2
replace: True
-> Updating checkpoint debias-wga-best_seed34.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed34.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 89.841
Robust Acc: 85.556 | Best Acc: 97.903
-------------------------------------
Training, Epoch 2:
Accuracies by groups:
0, 0  acc:  8365 /  9767 =  85.646
0, 1  acc:  6987 /  7535 =  92.727
1, 0  acc:  2428 /  2480 =  97.903
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 17934 / 19962 =  89.841
Robust  acc:   154 /   180 =  85.556
------------------------------------
Accuracies by groups:
0, 0  acc:  8365 /  9767 =  85.646
0, 1  acc:  6987 /  7535 =  92.727
1, 0  acc:  2428 /  2480 =  97.903
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 17934 / 19962 =  89.841
Robust  acc:   154 /   180 =  85.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8365 /  9767 =  85.646
0, 1  acc:  6987 /  7535 =  92.727
1, 0  acc:  2428 /  2480 =  97.903
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 17934 / 19962 =  89.841
Robust  acc:   154 /   180 =  85.556
------------------------------------
Epoch:   4 | Train Loss: 0.001 | Train Acc: 95.023 | Val Loss: 0.002 | Val Acc: 89.420
Training:
Accuracies by groups:
0, 0  acc: 15259 / 20401 =  74.795
0, 1  acc:  8406 /  9517 =  88.326
1, 0  acc: 123339 / 124667 =  98.935
1, 1  acc:  7665 /  8185 =  93.647
--------------------------------------
Average acc: 154669 / 162770 =  95.023
Robust  acc: 15259 / 20401 =  74.795
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7057 /  8535 =  82.683
0, 1  acc:  7719 /  8276 =  93.270
1, 0  acc:  2821 /  2874 =  98.156
1, 1  acc:   168 /   182 =  92.308
------------------------------------
Average acc: 17765 / 19867 =  89.420
Robust  acc:  7057 /  8535 =  82.683
------------------------------------
New max robust acc: 82.68306971294669
debias model - Saving best checkpoint at epoch 3
replace: True
-> Updating checkpoint debias-wga-best_seed34.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed34.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 90.041
Robust Acc: 86.014 | Best Acc: 97.823
-------------------------------------
Training, Epoch 3:
Accuracies by groups:
0, 0  acc:  8401 /  9767 =  86.014
0, 1  acc:  6992 /  7535 =  92.794
1, 0  acc:  2426 /  2480 =  97.823
1, 1  acc:   155 /   180 =  86.111
------------------------------------
Average acc: 17974 / 19962 =  90.041
Robust  acc:  8401 /  9767 =  86.014
------------------------------------
Accuracies by groups:
0, 0  acc:  8401 /  9767 =  86.014
0, 1  acc:  6992 /  7535 =  92.794
1, 0  acc:  2426 /  2480 =  97.823
1, 1  acc:   155 /   180 =  86.111
------------------------------------
Average acc: 17974 / 19962 =  90.041
Robust  acc:  8401 /  9767 =  86.014
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8401 /  9767 =  86.014
0, 1  acc:  6992 /  7535 =  92.794
1, 0  acc:  2426 /  2480 =  97.823
1, 1  acc:   155 /   180 =  86.111
------------------------------------
Average acc: 17974 / 19962 =  90.041
Robust  acc:  8401 /  9767 =  86.014
------------------------------------
Epoch:   5 | Train Loss: 0.001 | Train Acc: 95.911 | Val Loss: 0.002 | Val Acc: 91.217
Training:
Accuracies by groups:
0, 0  acc: 15671 / 20136 =  77.826
0, 1  acc:  8512 /  9460 =  89.979
1, 0  acc: 124067 / 125007 =  99.248
1, 1  acc:  7864 /  8167 =  96.290
--------------------------------------
Average acc: 156114 / 162770 =  95.911
Robust  acc: 15671 / 20136 =  77.826
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7273 /  8535 =  85.214
0, 1  acc:  7889 /  8276 =  95.324
1, 0  acc:  2798 /  2874 =  97.356
1, 1  acc:   162 /   182 =  89.011
------------------------------------
Average acc: 18122 / 19867 =  91.217
Robust  acc:  7273 /  8535 =  85.214
------------------------------------
New max robust acc: 85.21382542472173
debias model - Saving best checkpoint at epoch 4
replace: True
-> Updating checkpoint debias-wga-best_seed34.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed34.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.624
Robust Acc: 80.556 | Best Acc: 96.976
-------------------------------------
Training, Epoch 4:
Accuracies by groups:
0, 0  acc:  8585 /  9767 =  87.898
0, 1  acc:  7155 /  7535 =  94.957
1, 0  acc:  2405 /  2480 =  96.976
1, 1  acc:   145 /   180 =  80.556
------------------------------------
Average acc: 18290 / 19962 =  91.624
Robust  acc:   145 /   180 =  80.556
------------------------------------
Accuracies by groups:
0, 0  acc:  8585 /  9767 =  87.898
0, 1  acc:  7155 /  7535 =  94.957
1, 0  acc:  2405 /  2480 =  96.976
1, 1  acc:   145 /   180 =  80.556
------------------------------------
Average acc: 18290 / 19962 =  91.624
Robust  acc:   145 /   180 =  80.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8585 /  9767 =  87.898
0, 1  acc:  7155 /  7535 =  94.957
1, 0  acc:  2405 /  2480 =  96.976
1, 1  acc:   145 /   180 =  80.556
------------------------------------
Average acc: 18290 / 19962 =  91.624
Robust  acc:   145 /   180 =  80.556
------------------------------------
Epoch:   6 | Train Loss: 0.001 | Train Acc: 96.700 | Val Loss: 0.002 | Val Acc: 91.926
Training:
Accuracies by groups:
0, 0  acc: 16517 / 20377 =  81.057
0, 1  acc:  8626 /  9416 =  91.610
1, 0  acc: 124159 / 124693 =  99.572
1, 1  acc:  8096 /  8284 =  97.731
--------------------------------------
Average acc: 157398 / 162770 =  96.700
Robust  acc: 16517 / 20377 =  81.057
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7393 /  8535 =  86.620
0, 1  acc:  7947 /  8276 =  96.025
1, 0  acc:  2768 /  2874 =  96.312
1, 1  acc:   155 /   182 =  85.165
------------------------------------
Average acc: 18263 / 19867 =  91.926
Robust  acc:   155 /   182 =  85.165
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.345
Robust Acc: 80.000 | Best Acc: 96.411
-------------------------------------
Training, Epoch 5:
Accuracies by groups:
0, 0  acc:  8715 /  9767 =  89.229
0, 1  acc:  7184 /  7535 =  95.342
1, 0  acc:  2391 /  2480 =  96.411
1, 1  acc:   144 /   180 =  80.000
------------------------------------
Average acc: 18434 / 19962 =  92.345
Robust  acc:   144 /   180 =  80.000
------------------------------------
Accuracies by groups:
0, 0  acc:  8715 /  9767 =  89.229
0, 1  acc:  7184 /  7535 =  95.342
1, 0  acc:  2391 /  2480 =  96.411
1, 1  acc:   144 /   180 =  80.000
------------------------------------
Average acc: 18434 / 19962 =  92.345
Robust  acc:   144 /   180 =  80.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8715 /  9767 =  89.229
0, 1  acc:  7184 /  7535 =  95.342
1, 0  acc:  2391 /  2480 =  96.411
1, 1  acc:   144 /   180 =  80.000
------------------------------------
Average acc: 18434 / 19962 =  92.345
Robust  acc:   144 /   180 =  80.000
------------------------------------
Epoch:   7 | Train Loss: 0.001 | Train Acc: 97.533 | Val Loss: 0.002 | Val Acc: 92.314
Training:
Accuracies by groups:
0, 0  acc: 17403 / 20578 =  84.571
0, 1  acc:  8824 /  9414 =  93.733
1, 0  acc: 124394 / 124587 =  99.845
1, 1  acc:  8133 /  8191 =  99.292
--------------------------------------
Average acc: 158754 / 162770 =  97.533
Robust  acc: 17403 / 20578 =  84.571
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7488 /  8535 =  87.733
0, 1  acc:  7962 /  8276 =  96.206
1, 0  acc:  2736 /  2874 =  95.198
1, 1  acc:   154 /   182 =  84.615
------------------------------------
Average acc: 18340 / 19867 =  92.314
Robust  acc:   154 /   182 =  84.615
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 92.651
Robust Acc: 76.667 | Best Acc: 95.645
-------------------------------------
Training, Epoch 6:
Accuracies by groups:
0, 0  acc:  8789 /  9767 =  89.987
0, 1  acc:  7196 /  7535 =  95.501
1, 0  acc:  2372 /  2480 =  95.645
1, 1  acc:   138 /   180 =  76.667
------------------------------------
Average acc: 18495 / 19962 =  92.651
Robust  acc:   138 /   180 =  76.667
------------------------------------
Accuracies by groups:
0, 0  acc:  8789 /  9767 =  89.987
0, 1  acc:  7196 /  7535 =  95.501
1, 0  acc:  2372 /  2480 =  95.645
1, 1  acc:   138 /   180 =  76.667
------------------------------------
Average acc: 18495 / 19962 =  92.651
Robust  acc:   138 /   180 =  76.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8789 /  9767 =  89.987
0, 1  acc:  7196 /  7535 =  95.501
1, 0  acc:  2372 /  2480 =  95.645
1, 1  acc:   138 /   180 =  76.667
------------------------------------
Average acc: 18495 / 19962 =  92.651
Robust  acc:   138 /   180 =  76.667
------------------------------------
Epoch:   8 | Train Loss: 0.001 | Train Acc: 98.097 | Val Loss: 0.001 | Val Acc: 92.777
Training:
Accuracies by groups:
0, 0  acc: 18032 / 20510 =  87.918
0, 1  acc:  8990 /  9469 =  94.941
1, 0  acc: 124392 / 124502 =  99.912
1, 1  acc:  8258 /  8289 =  99.626
--------------------------------------
Average acc: 159672 / 162770 =  98.097
Robust  acc: 18032 / 20510 =  87.918
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7538 /  8535 =  88.319
0, 1  acc:  8021 /  8276 =  96.919
1, 0  acc:  2730 /  2874 =  94.990
1, 1  acc:   143 /   182 =  78.571
------------------------------------
Average acc: 18432 / 19867 =  92.777
Robust  acc:   143 /   182 =  78.571
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 93.017
Robust Acc: 70.556 | Best Acc: 96.443
-------------------------------------
Training, Epoch 7:
Accuracies by groups:
0, 0  acc:  8838 /  9767 =  90.488
0, 1  acc:  7267 /  7535 =  96.443
1, 0  acc:  2336 /  2480 =  94.194
1, 1  acc:   127 /   180 =  70.556
------------------------------------
Average acc: 18568 / 19962 =  93.017
Robust  acc:   127 /   180 =  70.556
------------------------------------
Accuracies by groups:
0, 0  acc:  8838 /  9767 =  90.488
0, 1  acc:  7267 /  7535 =  96.443
1, 0  acc:  2336 /  2480 =  94.194
1, 1  acc:   127 /   180 =  70.556
------------------------------------
Average acc: 18568 / 19962 =  93.017
Robust  acc:   127 /   180 =  70.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8838 /  9767 =  90.488
0, 1  acc:  7267 /  7535 =  96.443
1, 0  acc:  2336 /  2480 =  94.194
1, 1  acc:   127 /   180 =  70.556
------------------------------------
Average acc: 18568 / 19962 =  93.017
Robust  acc:   127 /   180 =  70.556
------------------------------------
Epoch:   9 | Train Loss: 0.001 | Train Acc: 98.529 | Val Loss: 0.001 | Val Acc: 93.446
Training:
Accuracies by groups:
0, 0  acc: 18619 / 20549 =  90.608
0, 1  acc:  9223 /  9589 =  96.183
1, 0  acc: 124427 / 124504 =  99.938
1, 1  acc:  8106 /  8128 =  99.729
--------------------------------------
Average acc: 160375 / 162770 =  98.529
Robust  acc: 18619 / 20549 =  90.608
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7695 /  8535 =  90.158
0, 1  acc:  8073 /  8276 =  97.547
1, 0  acc:  2668 /  2874 =  92.832
1, 1  acc:   129 /   182 =  70.879
------------------------------------
Average acc: 18565 / 19867 =  93.446
Robust  acc:   129 /   182 =  70.879
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 93.543
Robust Acc: 66.667 | Best Acc: 96.908
-------------------------------------
Training, Epoch 8:
Accuracies by groups:
0, 0  acc:  8964 /  9767 =  91.778
0, 1  acc:  7302 /  7535 =  96.908
1, 0  acc:  2287 /  2480 =  92.218
1, 1  acc:   120 /   180 =  66.667
------------------------------------
Average acc: 18673 / 19962 =  93.543
Robust  acc:   120 /   180 =  66.667
------------------------------------
Accuracies by groups:
0, 0  acc:  8964 /  9767 =  91.778
0, 1  acc:  7302 /  7535 =  96.908
1, 0  acc:  2287 /  2480 =  92.218
1, 1  acc:   120 /   180 =  66.667
------------------------------------
Average acc: 18673 / 19962 =  93.543
Robust  acc:   120 /   180 =  66.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8964 /  9767 =  91.778
0, 1  acc:  7302 /  7535 =  96.908
1, 0  acc:  2287 /  2480 =  92.218
1, 1  acc:   120 /   180 =  66.667
------------------------------------
Average acc: 18673 / 19962 =  93.543
Robust  acc:   120 /   180 =  66.667
------------------------------------
Epoch:  10 | Train Loss: 0.001 | Train Acc: 98.854 | Val Loss: 0.002 | Val Acc: 92.420
Training:
Accuracies by groups:
0, 0  acc: 18879 / 20396 =  92.562
0, 1  acc:  9357 /  9601 =  97.459
1, 0  acc: 124449 / 124533 =  99.933
1, 1  acc:  8220 /  8240 =  99.757
--------------------------------------
Average acc: 160905 / 162770 =  98.854
Robust  acc: 18879 / 20396 =  92.562
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7441 /  8535 =  87.182
0, 1  acc:  8048 /  8276 =  97.245
1, 0  acc:  2733 /  2874 =  95.094
1, 1  acc:   139 /   182 =  76.374
------------------------------------
Average acc: 18361 / 19867 =  92.420
Robust  acc:   139 /   182 =  76.374
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.846
Robust Acc: 71.667 | Best Acc: 96.709
-------------------------------------
Training, Epoch 9:
Accuracies by groups:
0, 0  acc:  8762 /  9767 =  89.710
0, 1  acc:  7287 /  7535 =  96.709
1, 0  acc:  2356 /  2480 =  95.000
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18534 / 19962 =  92.846
Robust  acc:   129 /   180 =  71.667
------------------------------------
Accuracies by groups:
0, 0  acc:  8762 /  9767 =  89.710
0, 1  acc:  7287 /  7535 =  96.709
1, 0  acc:  2356 /  2480 =  95.000
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18534 / 19962 =  92.846
Robust  acc:   129 /   180 =  71.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8762 /  9767 =  89.710
0, 1  acc:  7287 /  7535 =  96.709
1, 0  acc:  2356 /  2480 =  95.000
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18534 / 19962 =  92.846
Robust  acc:   129 /   180 =  71.667
------------------------------------
Epoch:  11 | Train Loss: 0.001 | Train Acc: 98.781 | Val Loss: 0.001 | Val Acc: 94.267
Training:
Accuracies by groups:
0, 0  acc: 18762 / 20268 =  92.570
0, 1  acc:  9111 /  9374 =  97.194
1, 0  acc: 124645 / 124807 =  99.870
1, 1  acc:  8268 /  8321 =  99.363
--------------------------------------
Average acc: 160786 / 162770 =  98.781
Robust  acc: 18762 / 20268 =  92.570
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7830 /  8535 =  91.740
0, 1  acc:  8164 /  8276 =  98.647
1, 0  acc:  2622 /  2874 =  91.232
1, 1  acc:   112 /   182 =  61.538
------------------------------------
Average acc: 18728 / 19867 =  94.267
Robust  acc:   112 /   182 =  61.538
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.520
Robust Acc: 53.889 | Best Acc: 98.261
-------------------------------------
Training, Epoch 10:
Accuracies by groups:
0, 0  acc:  9141 /  9767 =  93.591
0, 1  acc:  7404 /  7535 =  98.261
1, 0  acc:  2226 /  2480 =  89.758
1, 1  acc:    97 /   180 =  53.889
------------------------------------
Average acc: 18868 / 19962 =  94.520
Robust  acc:    97 /   180 =  53.889
------------------------------------
Accuracies by groups:
0, 0  acc:  9141 /  9767 =  93.591
0, 1  acc:  7404 /  7535 =  98.261
1, 0  acc:  2226 /  2480 =  89.758
1, 1  acc:    97 /   180 =  53.889
------------------------------------
Average acc: 18868 / 19962 =  94.520
Robust  acc:    97 /   180 =  53.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9141 /  9767 =  93.591
0, 1  acc:  7404 /  7535 =  98.261
1, 0  acc:  2226 /  2480 =  89.758
1, 1  acc:    97 /   180 =  53.889
------------------------------------
Average acc: 18868 / 19962 =  94.520
Robust  acc:    97 /   180 =  53.889
------------------------------------
Epoch:  12 | Train Loss: 0.001 | Train Acc: 98.510 | Val Loss: 0.001 | Val Acc: 94.096
Training:
Accuracies by groups:
0, 0  acc: 18641 / 20346 =  91.620
0, 1  acc:  9179 /  9469 =  96.937
1, 0  acc: 124409 / 124739 =  99.735
1, 1  acc:  8116 /  8216 =  98.783
--------------------------------------
Average acc: 160345 / 162770 =  98.510
Robust  acc: 18641 / 20346 =  91.620
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7853 /  8535 =  92.009
0, 1  acc:  8124 /  8276 =  98.163
1, 0  acc:  2596 /  2874 =  90.327
1, 1  acc:   121 /   182 =  66.484
------------------------------------
Average acc: 18694 / 19867 =  94.096
Robust  acc:   121 /   182 =  66.484
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.234
Robust Acc: 59.444 | Best Acc: 97.704
-------------------------------------
Training, Epoch 11:
Accuracies by groups:
0, 0  acc:  9128 /  9767 =  93.458
0, 1  acc:  7362 /  7535 =  97.704
1, 0  acc:  2214 /  2480 =  89.274
1, 1  acc:   107 /   180 =  59.444
------------------------------------
Average acc: 18811 / 19962 =  94.234
Robust  acc:   107 /   180 =  59.444
------------------------------------
Accuracies by groups:
0, 0  acc:  9128 /  9767 =  93.458
0, 1  acc:  7362 /  7535 =  97.704
1, 0  acc:  2214 /  2480 =  89.274
1, 1  acc:   107 /   180 =  59.444
------------------------------------
Average acc: 18811 / 19962 =  94.234
Robust  acc:   107 /   180 =  59.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9128 /  9767 =  93.458
0, 1  acc:  7362 /  7535 =  97.704
1, 0  acc:  2214 /  2480 =  89.274
1, 1  acc:   107 /   180 =  59.444
------------------------------------
Average acc: 18811 / 19962 =  94.234
Robust  acc:   107 /   180 =  59.444
------------------------------------
Epoch:  13 | Train Loss: 0.001 | Train Acc: 97.942 | Val Loss: 0.001 | Val Acc: 93.874
Training:
Accuracies by groups:
0, 0  acc: 17935 / 20162 =  88.954
0, 1  acc:  9066 /  9469 =  95.744
1, 0  acc: 124229 / 124811 =  99.534
1, 1  acc:  8191 /  8328 =  98.355
--------------------------------------
Average acc: 159421 / 162770 =  97.942
Robust  acc: 17935 / 20162 =  88.954
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7735 /  8535 =  90.627
0, 1  acc:  8101 /  8276 =  97.885
1, 0  acc:  2690 /  2874 =  93.598
1, 1  acc:   124 /   182 =  68.132
------------------------------------
Average acc: 18650 / 19867 =  93.874
Robust  acc:   124 /   182 =  68.132
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.169
Robust Acc: 63.333 | Best Acc: 97.598
-------------------------------------
Training, Epoch 12:
Accuracies by groups:
0, 0  acc:  9028 /  9767 =  92.434
0, 1  acc:  7354 /  7535 =  97.598
1, 0  acc:  2302 /  2480 =  92.823
1, 1  acc:   114 /   180 =  63.333
------------------------------------
Average acc: 18798 / 19962 =  94.169
Robust  acc:   114 /   180 =  63.333
------------------------------------
Accuracies by groups:
0, 0  acc:  9028 /  9767 =  92.434
0, 1  acc:  7354 /  7535 =  97.598
1, 0  acc:  2302 /  2480 =  92.823
1, 1  acc:   114 /   180 =  63.333
------------------------------------
Average acc: 18798 / 19962 =  94.169
Robust  acc:   114 /   180 =  63.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9028 /  9767 =  92.434
0, 1  acc:  7354 /  7535 =  97.598
1, 0  acc:  2302 /  2480 =  92.823
1, 1  acc:   114 /   180 =  63.333
------------------------------------
Average acc: 18798 / 19962 =  94.169
Robust  acc:   114 /   180 =  63.333
------------------------------------
Epoch:  14 | Train Loss: 0.001 | Train Acc: 97.237 | Val Loss: 0.002 | Val Acc: 89.943
Training:
Accuracies by groups:
0, 0  acc: 17314 / 20190 =  85.755
0, 1  acc:  8792 /  9324 =  94.294
1, 0  acc: 124239 / 125086 =  99.323
1, 1  acc:  7928 /  8170 =  97.038
--------------------------------------
Average acc: 158273 / 162770 =  97.237
Robust  acc: 17314 / 20190 =  85.755
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7031 /  8535 =  82.378
0, 1  acc:  7855 /  8276 =  94.913
1, 0  acc:  2825 /  2874 =  98.295
1, 1  acc:   158 /   182 =  86.813
------------------------------------
Average acc: 17869 / 19867 =  89.943
Robust  acc:  7031 /  8535 =  82.378
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 90.357
Robust Acc: 85.000 | Best Acc: 97.823
-------------------------------------
Training, Epoch 13:
Accuracies by groups:
0, 0  acc:  8356 /  9767 =  85.553
0, 1  acc:  7102 /  7535 =  94.253
1, 0  acc:  2426 /  2480 =  97.823
1, 1  acc:   153 /   180 =  85.000
------------------------------------
Average acc: 18037 / 19962 =  90.357
Robust  acc:   153 /   180 =  85.000
------------------------------------
Accuracies by groups:
0, 0  acc:  8356 /  9767 =  85.553
0, 1  acc:  7102 /  7535 =  94.253
1, 0  acc:  2426 /  2480 =  97.823
1, 1  acc:   153 /   180 =  85.000
------------------------------------
Average acc: 18037 / 19962 =  90.357
Robust  acc:   153 /   180 =  85.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8356 /  9767 =  85.553
0, 1  acc:  7102 /  7535 =  94.253
1, 0  acc:  2426 /  2480 =  97.823
1, 1  acc:   153 /   180 =  85.000
------------------------------------
Average acc: 18037 / 19962 =  90.357
Robust  acc:   153 /   180 =  85.000
------------------------------------
Epoch:  15 | Train Loss: 0.001 | Train Acc: 96.643 | Val Loss: 0.001 | Val Acc: 94.312
Training:
Accuracies by groups:
0, 0  acc: 16943 / 20347 =  83.270
0, 1  acc:  8768 /  9433 =  92.950
1, 0  acc: 123870 / 124951 =  99.135
1, 1  acc:  7724 /  8039 =  96.082
--------------------------------------
Average acc: 157305 / 162770 =  96.643
Robust  acc: 16943 / 20347 =  83.270
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7867 /  8535 =  92.173
0, 1  acc:  8087 /  8276 =  97.716
1, 0  acc:  2642 /  2874 =  91.928
1, 1  acc:   141 /   182 =  77.473
------------------------------------
Average acc: 18737 / 19867 =  94.312
Robust  acc:   141 /   182 =  77.473
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.309
Robust Acc: 66.667 | Best Acc: 97.319
-------------------------------------
Training, Epoch 14:
Accuracies by groups:
0, 0  acc:  9130 /  9767 =  93.478
0, 1  acc:  7333 /  7535 =  97.319
1, 0  acc:  2243 /  2480 =  90.444
1, 1  acc:   120 /   180 =  66.667
------------------------------------
Average acc: 18826 / 19962 =  94.309
Robust  acc:   120 /   180 =  66.667
------------------------------------
Accuracies by groups:
0, 0  acc:  9130 /  9767 =  93.478
0, 1  acc:  7333 /  7535 =  97.319
1, 0  acc:  2243 /  2480 =  90.444
1, 1  acc:   120 /   180 =  66.667
------------------------------------
Average acc: 18826 / 19962 =  94.309
Robust  acc:   120 /   180 =  66.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9130 /  9767 =  93.478
0, 1  acc:  7333 /  7535 =  97.319
1, 0  acc:  2243 /  2480 =  90.444
1, 1  acc:   120 /   180 =  66.667
------------------------------------
Average acc: 18826 / 19962 =  94.309
Robust  acc:   120 /   180 =  66.667
------------------------------------
Epoch:  16 | Train Loss: 0.001 | Train Acc: 96.131 | Val Loss: 0.001 | Val Acc: 95.158
Training:
Accuracies by groups:
0, 0  acc: 16533 / 20441 =  80.882
0, 1  acc:  8713 /  9521 =  91.513
1, 0  acc: 123392 / 124591 =  99.038
1, 1  acc:  7834 /  8217 =  95.339
--------------------------------------
Average acc: 156472 / 162770 =  96.131
Robust  acc: 16533 / 20441 =  80.882
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8112 /  8535 =  95.044
0, 1  acc:  8199 /  8276 =  99.070
1, 0  acc:  2489 /  2874 =  86.604
1, 1  acc:   105 /   182 =  57.692
------------------------------------
Average acc: 18905 / 19867 =  95.158
Robust  acc:   105 /   182 =  57.692
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.286
Robust Acc: 48.889 | Best Acc: 98.885
-------------------------------------
Training, Epoch 15:
Accuracies by groups:
0, 0  acc:  9380 /  9767 =  96.038
0, 1  acc:  7451 /  7535 =  98.885
1, 0  acc:  2102 /  2480 =  84.758
1, 1  acc:    88 /   180 =  48.889
------------------------------------
Average acc: 19021 / 19962 =  95.286
Robust  acc:    88 /   180 =  48.889
------------------------------------
Accuracies by groups:
0, 0  acc:  9380 /  9767 =  96.038
0, 1  acc:  7451 /  7535 =  98.885
1, 0  acc:  2102 /  2480 =  84.758
1, 1  acc:    88 /   180 =  48.889
------------------------------------
Average acc: 19021 / 19962 =  95.286
Robust  acc:    88 /   180 =  48.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9380 /  9767 =  96.038
0, 1  acc:  7451 /  7535 =  98.885
1, 0  acc:  2102 /  2480 =  84.758
1, 1  acc:    88 /   180 =  48.889
------------------------------------
Average acc: 19021 / 19962 =  95.286
Robust  acc:    88 /   180 =  48.889
------------------------------------
Epoch:  17 | Train Loss: 0.001 | Train Acc: 95.652 | Val Loss: 0.001 | Val Acc: 95.032
Training:
Accuracies by groups:
0, 0  acc: 16062 / 20340 =  78.968
0, 1  acc:  8562 /  9463 =  90.479
1, 0  acc: 123319 / 124771 =  98.836
1, 1  acc:  7749 /  8196 =  94.546
--------------------------------------
Average acc: 155692 / 162770 =  95.652
Robust  acc: 16062 / 20340 =  78.968
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8070 /  8535 =  94.552
0, 1  acc:  8176 /  8276 =  98.792
1, 0  acc:  2527 /  2874 =  87.926
1, 1  acc:   107 /   182 =  58.791
------------------------------------
Average acc: 18880 / 19867 =  95.032
Robust  acc:   107 /   182 =  58.791
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.181
Robust Acc: 51.667 | Best Acc: 98.686
-------------------------------------
Training, Epoch 16:
Accuracies by groups:
0, 0  acc:  9332 /  9767 =  95.546
0, 1  acc:  7436 /  7535 =  98.686
1, 0  acc:  2139 /  2480 =  86.250
1, 1  acc:    93 /   180 =  51.667
------------------------------------
Average acc: 19000 / 19962 =  95.181
Robust  acc:    93 /   180 =  51.667
------------------------------------
Accuracies by groups:
0, 0  acc:  9332 /  9767 =  95.546
0, 1  acc:  7436 /  7535 =  98.686
1, 0  acc:  2139 /  2480 =  86.250
1, 1  acc:    93 /   180 =  51.667
------------------------------------
Average acc: 19000 / 19962 =  95.181
Robust  acc:    93 /   180 =  51.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9332 /  9767 =  95.546
0, 1  acc:  7436 /  7535 =  98.686
1, 0  acc:  2139 /  2480 =  86.250
1, 1  acc:    93 /   180 =  51.667
------------------------------------
Average acc: 19000 / 19962 =  95.181
Robust  acc:    93 /   180 =  51.667
------------------------------------
Epoch:  18 | Train Loss: 0.001 | Train Acc: 95.058 | Val Loss: 0.004 | Val Acc: 79.106
Training:
Accuracies by groups:
0, 0  acc: 15570 / 20258 =  76.859
0, 1  acc:  8394 /  9478 =  88.563
1, 0  acc: 123079 / 124808 =  98.615
1, 1  acc:  7683 /  8226 =  93.399
--------------------------------------
Average acc: 154726 / 162770 =  95.058
Robust  acc: 15570 / 20258 =  76.859
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5740 /  8535 =  67.252
0, 1  acc:  6936 /  8276 =  83.809
1, 0  acc:  2861 /  2874 =  99.548
1, 1  acc:   179 /   182 =  98.352
------------------------------------
Average acc: 15716 / 19867 =  79.106
Robust  acc:  5740 /  8535 =  67.252
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 80.513
Robust Acc: 73.318 | Best Acc: 99.274
-------------------------------------
Training, Epoch 17:
Accuracies by groups:
0, 0  acc:  7161 /  9767 =  73.318
0, 1  acc:  6280 /  7535 =  83.344
1, 0  acc:  2462 /  2480 =  99.274
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 16072 / 19962 =  80.513
Robust  acc:  7161 /  9767 =  73.318
------------------------------------
Accuracies by groups:
0, 0  acc:  7161 /  9767 =  73.318
0, 1  acc:  6280 /  7535 =  83.344
1, 0  acc:  2462 /  2480 =  99.274
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 16072 / 19962 =  80.513
Robust  acc:  7161 /  9767 =  73.318
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7161 /  9767 =  73.318
0, 1  acc:  6280 /  7535 =  83.344
1, 0  acc:  2462 /  2480 =  99.274
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 16072 / 19962 =  80.513
Robust  acc:  7161 /  9767 =  73.318
------------------------------------
Epoch:  19 | Train Loss: 0.001 | Train Acc: 94.684 | Val Loss: 0.003 | Val Acc: 86.284
Training:
Accuracies by groups:
0, 0  acc: 15092 / 20084 =  75.144
0, 1  acc:  8253 /  9378 =  88.004
1, 0  acc: 123099 / 125026 =  98.459
1, 1  acc:  7673 /  8282 =  92.647
--------------------------------------
Average acc: 154117 / 162770 =  94.684
Robust  acc: 15092 / 20084 =  75.144
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6686 /  8535 =  78.336
0, 1  acc:  7449 /  8276 =  90.007
1, 0  acc:  2833 /  2874 =  98.573
1, 1  acc:   174 /   182 =  95.604
------------------------------------
Average acc: 17142 / 19867 =  86.284
Robust  acc:  6686 /  8535 =  78.336
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 87.626
Robust Acc: 82.758 | Best Acc: 98.387
-------------------------------------
Training, Epoch 18:
Accuracies by groups:
0, 0  acc:  8083 /  9767 =  82.758
0, 1  acc:  6811 /  7535 =  90.392
1, 0  acc:  2440 /  2480 =  98.387
1, 1  acc:   158 /   180 =  87.778
------------------------------------
Average acc: 17492 / 19962 =  87.626
Robust  acc:  8083 /  9767 =  82.758
------------------------------------
Accuracies by groups:
0, 0  acc:  8083 /  9767 =  82.758
0, 1  acc:  6811 /  7535 =  90.392
1, 0  acc:  2440 /  2480 =  98.387
1, 1  acc:   158 /   180 =  87.778
------------------------------------
Average acc: 17492 / 19962 =  87.626
Robust  acc:  8083 /  9767 =  82.758
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8083 /  9767 =  82.758
0, 1  acc:  6811 /  7535 =  90.392
1, 0  acc:  2440 /  2480 =  98.387
1, 1  acc:   158 /   180 =  87.778
------------------------------------
Average acc: 17492 / 19962 =  87.626
Robust  acc:  8083 /  9767 =  82.758
------------------------------------
Epoch:  20 | Train Loss: 0.002 | Train Acc: 94.159 | Val Loss: 0.002 | Val Acc: 92.717
Training:
Accuracies by groups:
0, 0  acc: 14827 / 20250 =  73.220
0, 1  acc:  8071 /  9371 =  86.127
1, 0  acc: 122765 / 124892 =  98.297
1, 1  acc:  7600 /  8257 =  92.043
--------------------------------------
Average acc: 153263 / 162770 =  94.159
Robust  acc: 14827 / 20250 =  73.220
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7563 /  8535 =  88.612
0, 1  acc:  7970 /  8276 =  96.303
1, 0  acc:  2739 /  2874 =  95.303
1, 1  acc:   148 /   182 =  81.319
------------------------------------
Average acc: 18420 / 19867 =  92.717
Robust  acc:   148 /   182 =  81.319
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.082
Robust Acc: 73.889 | Best Acc: 96.483
-------------------------------------
Training, Epoch 19:
Accuracies by groups:
0, 0  acc:  8831 /  9767 =  90.417
0, 1  acc:  7270 /  7535 =  96.483
1, 0  acc:  2347 /  2480 =  94.637
1, 1  acc:   133 /   180 =  73.889
------------------------------------
Average acc: 18581 / 19962 =  93.082
Robust  acc:   133 /   180 =  73.889
------------------------------------
Accuracies by groups:
0, 0  acc:  8831 /  9767 =  90.417
0, 1  acc:  7270 /  7535 =  96.483
1, 0  acc:  2347 /  2480 =  94.637
1, 1  acc:   133 /   180 =  73.889
------------------------------------
Average acc: 18581 / 19962 =  93.082
Robust  acc:   133 /   180 =  73.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8831 /  9767 =  90.417
0, 1  acc:  7270 /  7535 =  96.483
1, 0  acc:  2347 /  2480 =  94.637
1, 1  acc:   133 /   180 =  73.889
------------------------------------
Average acc: 18581 / 19962 =  93.082
Robust  acc:   133 /   180 =  73.889
------------------------------------
Epoch:  21 | Train Loss: 0.002 | Train Acc: 93.955 | Val Loss: 0.003 | Val Acc: 92.178
Training:
Accuracies by groups:
0, 0  acc: 14700 / 20281 =  72.482
0, 1  acc:  8017 /  9498 =  84.407
1, 0  acc: 122508 / 124625 =  98.301
1, 1  acc:  7705 /  8366 =  92.099
--------------------------------------
Average acc: 152930 / 162770 =  93.955
Robust  acc: 14700 / 20281 =  72.482
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7421 /  8535 =  86.948
0, 1  acc:  7973 /  8276 =  96.339
1, 0  acc:  2774 /  2874 =  96.521
1, 1  acc:   145 /   182 =  79.670
------------------------------------
Average acc: 18313 / 19867 =  92.178
Robust  acc:   145 /   182 =  79.670
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.942
Robust Acc: 78.333 | Best Acc: 96.257
-------------------------------------
Training, Epoch 20:
Accuracies by groups:
0, 0  acc:  8776 /  9767 =  89.854
0, 1  acc:  7253 /  7535 =  96.257
1, 0  acc:  2383 /  2480 =  96.089
1, 1  acc:   141 /   180 =  78.333
------------------------------------
Average acc: 18553 / 19962 =  92.942
Robust  acc:   141 /   180 =  78.333
------------------------------------
Accuracies by groups:
0, 0  acc:  8776 /  9767 =  89.854
0, 1  acc:  7253 /  7535 =  96.257
1, 0  acc:  2383 /  2480 =  96.089
1, 1  acc:   141 /   180 =  78.333
------------------------------------
Average acc: 18553 / 19962 =  92.942
Robust  acc:   141 /   180 =  78.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8776 /  9767 =  89.854
0, 1  acc:  7253 /  7535 =  96.257
1, 0  acc:  2383 /  2480 =  96.089
1, 1  acc:   141 /   180 =  78.333
------------------------------------
Average acc: 18553 / 19962 =  92.942
Robust  acc:   141 /   180 =  78.333
------------------------------------
Epoch:  22 | Train Loss: 0.002 | Train Acc: 93.735 | Val Loss: 0.003 | Val Acc: 89.153
Training:
Accuracies by groups:
0, 0  acc: 14744 / 20584 =  71.628
0, 1  acc:  7924 /  9471 =  83.666
1, 0  acc: 122275 / 124463 =  98.242
1, 1  acc:  7630 /  8252 =  92.462
--------------------------------------
Average acc: 152573 / 162770 =  93.735
Robust  acc: 14744 / 20584 =  71.628
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7023 /  8535 =  82.285
0, 1  acc:  7749 /  8276 =  93.632
1, 0  acc:  2785 /  2874 =  96.903
1, 1  acc:   155 /   182 =  85.165
------------------------------------
Average acc: 17712 / 19867 =  89.153
Robust  acc:  7023 /  8535 =  82.285
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 90.021
Robust Acc: 79.444 | Best Acc: 96.331
-------------------------------------
Training, Epoch 21:
Accuracies by groups:
0, 0  acc:  8368 /  9767 =  85.676
0, 1  acc:  7070 /  7535 =  93.829
1, 0  acc:  2389 /  2480 =  96.331
1, 1  acc:   143 /   180 =  79.444
------------------------------------
Average acc: 17970 / 19962 =  90.021
Robust  acc:   143 /   180 =  79.444
------------------------------------
Accuracies by groups:
0, 0  acc:  8368 /  9767 =  85.676
0, 1  acc:  7070 /  7535 =  93.829
1, 0  acc:  2389 /  2480 =  96.331
1, 1  acc:   143 /   180 =  79.444
------------------------------------
Average acc: 17970 / 19962 =  90.021
Robust  acc:   143 /   180 =  79.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8368 /  9767 =  85.676
0, 1  acc:  7070 /  7535 =  93.829
1, 0  acc:  2389 /  2480 =  96.331
1, 1  acc:   143 /   180 =  79.444
------------------------------------
Average acc: 17970 / 19962 =  90.021
Robust  acc:   143 /   180 =  79.444
------------------------------------
Epoch:  23 | Train Loss: 0.002 | Train Acc: 93.557 | Val Loss: 0.007 | Val Acc: 59.586
Training:
Accuracies by groups:
0, 0  acc: 14226 / 20362 =  69.865
0, 1  acc:  7781 /  9422 =  82.583
1, 0  acc: 122651 / 124741 =  98.325
1, 1  acc:  7624 /  8245 =  92.468
--------------------------------------
Average acc: 152282 / 162770 =  93.557
Robust  acc: 14226 / 20362 =  69.865
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4079 /  8535 =  47.791
0, 1  acc:  4708 /  8276 =  56.887
1, 0  acc:  2869 /  2874 =  99.826
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc: 11838 / 19867 =  59.586
Robust  acc:  4079 /  8535 =  47.791
------------------------------------
-------------------------------------------
Avg Test Loss: 0.006 | Avg Test Acc: 61.487
Robust Acc: 54.029 | Best Acc: 99.798
-------------------------------------
Training, Epoch 22:
Accuracies by groups:
0, 0  acc:  5277 /  9767 =  54.029
0, 1  acc:  4345 /  7535 =  57.664
1, 0  acc:  2475 /  2480 =  99.798
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 12274 / 19962 =  61.487
Robust  acc:  5277 /  9767 =  54.029
------------------------------------
Accuracies by groups:
0, 0  acc:  5277 /  9767 =  54.029
0, 1  acc:  4345 /  7535 =  57.664
1, 0  acc:  2475 /  2480 =  99.798
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 12274 / 19962 =  61.487
Robust  acc:  5277 /  9767 =  54.029
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  5277 /  9767 =  54.029
0, 1  acc:  4345 /  7535 =  57.664
1, 0  acc:  2475 /  2480 =  99.798
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 12274 / 19962 =  61.487
Robust  acc:  5277 /  9767 =  54.029
------------------------------------
Epoch:  24 | Train Loss: 0.002 | Train Acc: 93.563 | Val Loss: 0.003 | Val Acc: 92.480
Training:
Accuracies by groups:
0, 0  acc: 14182 / 20353 =  69.680
0, 1  acc:  7807 /  9585 =  81.450
1, 0  acc: 122747 / 124701 =  98.433
1, 1  acc:  7556 /  8131 =  92.928
--------------------------------------
Average acc: 152292 / 162770 =  93.563
Robust  acc: 14182 / 20353 =  69.680
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7620 /  8535 =  89.279
0, 1  acc:  7930 /  8276 =  95.819
1, 0  acc:  2689 /  2874 =  93.563
1, 1  acc:   134 /   182 =  73.626
------------------------------------
Average acc: 18373 / 19867 =  92.480
Robust  acc:   134 /   182 =  73.626
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 93.127
Robust Acc: 68.333 | Best Acc: 95.461
-------------------------------------
Training, Epoch 23:
Accuracies by groups:
0, 0  acc:  8964 /  9767 =  91.778
0, 1  acc:  7193 /  7535 =  95.461
1, 0  acc:  2310 /  2480 =  93.145
1, 1  acc:   123 /   180 =  68.333
------------------------------------
Average acc: 18590 / 19962 =  93.127
Robust  acc:   123 /   180 =  68.333
------------------------------------
Accuracies by groups:
0, 0  acc:  8964 /  9767 =  91.778
0, 1  acc:  7193 /  7535 =  95.461
1, 0  acc:  2310 /  2480 =  93.145
1, 1  acc:   123 /   180 =  68.333
------------------------------------
Average acc: 18590 / 19962 =  93.127
Robust  acc:   123 /   180 =  68.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8964 /  9767 =  91.778
0, 1  acc:  7193 /  7535 =  95.461
1, 0  acc:  2310 /  2480 =  93.145
1, 1  acc:   123 /   180 =  68.333
------------------------------------
Average acc: 18590 / 19962 =  93.127
Robust  acc:   123 /   180 =  68.333
------------------------------------
Epoch:  25 | Train Loss: 0.002 | Train Acc: 93.486 | Val Loss: 0.003 | Val Acc: 88.911
Training:
Accuracies by groups:
0, 0  acc: 13961 / 20201 =  69.110
0, 1  acc:  7536 /  9413 =  80.059
1, 0  acc: 122939 / 124839 =  98.478
1, 1  acc:  7731 /  8317 =  92.954
--------------------------------------
Average acc: 152167 / 162770 =  93.486
Robust  acc: 13961 / 20201 =  69.110
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7014 /  8535 =  82.179
0, 1  acc:  7678 /  8276 =  92.774
1, 0  acc:  2811 /  2874 =  97.808
1, 1  acc:   161 /   182 =  88.462
------------------------------------
Average acc: 17664 / 19867 =  88.911
Robust  acc:  7014 /  8535 =  82.179
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 90.001
Robust Acc: 83.333 | Best Acc: 97.097
-------------------------------------
Training, Epoch 24:
Accuracies by groups:
0, 0  acc:  8372 /  9767 =  85.717
0, 1  acc:  7036 /  7535 =  93.378
1, 0  acc:  2408 /  2480 =  97.097
1, 1  acc:   150 /   180 =  83.333
------------------------------------
Average acc: 17966 / 19962 =  90.001
Robust  acc:   150 /   180 =  83.333
------------------------------------
Accuracies by groups:
0, 0  acc:  8372 /  9767 =  85.717
0, 1  acc:  7036 /  7535 =  93.378
1, 0  acc:  2408 /  2480 =  97.097
1, 1  acc:   150 /   180 =  83.333
------------------------------------
Average acc: 17966 / 19962 =  90.001
Robust  acc:   150 /   180 =  83.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8372 /  9767 =  85.717
0, 1  acc:  7036 /  7535 =  93.378
1, 0  acc:  2408 /  2480 =  97.097
1, 1  acc:   150 /   180 =  83.333
------------------------------------
Average acc: 17966 / 19962 =  90.001
Robust  acc:   150 /   180 =  83.333
------------------------------------
Epoch:  26 | Train Loss: 0.002 | Train Acc: 93.282 | Val Loss: 0.003 | Val Acc: 90.910
Training:
Accuracies by groups:
0, 0  acc: 13970 / 20480 =  68.213
0, 1  acc:  7675 /  9588 =  80.048
1, 0  acc: 122531 / 124441 =  98.465
1, 1  acc:  7659 /  8261 =  92.713
--------------------------------------
Average acc: 151835 / 162770 =  93.282
Robust  acc: 13970 / 20480 =  68.213
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7326 /  8535 =  85.835
0, 1  acc:  7826 /  8276 =  94.563
1, 0  acc:  2758 /  2874 =  95.964
1, 1  acc:   151 /   182 =  82.967
------------------------------------
Average acc: 18061 / 19867 =  90.910
Robust  acc:   151 /   182 =  82.967
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 91.689
Robust Acc: 76.667 | Best Acc: 94.798
-------------------------------------
Training, Epoch 25:
Accuracies by groups:
0, 0  acc:  8676 /  9767 =  88.830
0, 1  acc:  7143 /  7535 =  94.798
1, 0  acc:  2346 /  2480 =  94.597
1, 1  acc:   138 /   180 =  76.667
------------------------------------
Average acc: 18303 / 19962 =  91.689
Robust  acc:   138 /   180 =  76.667
------------------------------------
Accuracies by groups:
0, 0  acc:  8676 /  9767 =  88.830
0, 1  acc:  7143 /  7535 =  94.798
1, 0  acc:  2346 /  2480 =  94.597
1, 1  acc:   138 /   180 =  76.667
------------------------------------
Average acc: 18303 / 19962 =  91.689
Robust  acc:   138 /   180 =  76.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8676 /  9767 =  88.830
0, 1  acc:  7143 /  7535 =  94.798
1, 0  acc:  2346 /  2480 =  94.597
1, 1  acc:   138 /   180 =  76.667
------------------------------------
Average acc: 18303 / 19962 =  91.689
Robust  acc:   138 /   180 =  76.667
------------------------------------
Epoch:  27 | Train Loss: 0.002 | Train Acc: 93.287 | Val Loss: 0.003 | Val Acc: 88.312
Training:
Accuracies by groups:
0, 0  acc: 13919 / 20359 =  68.368
0, 1  acc:  7598 /  9558 =  79.494
1, 0  acc: 122710 / 124598 =  98.485
1, 1  acc:  7616 /  8255 =  92.259
--------------------------------------
Average acc: 151843 / 162770 =  93.287
Robust  acc: 13919 / 20359 =  68.368
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7062 /  8535 =  82.742
0, 1  acc:  7549 /  8276 =  91.216
1, 0  acc:  2781 /  2874 =  96.764
1, 1  acc:   153 /   182 =  84.066
------------------------------------
Average acc: 17545 / 19867 =  88.312
Robust  acc:  7062 /  8535 =  82.742
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 89.766
Robust Acc: 81.111 | Best Acc: 96.573
-------------------------------------
Training, Epoch 26:
Accuracies by groups:
0, 0  acc:  8472 /  9767 =  86.741
0, 1  acc:  6906 /  7535 =  91.652
1, 0  acc:  2395 /  2480 =  96.573
1, 1  acc:   146 /   180 =  81.111
------------------------------------
Average acc: 17919 / 19962 =  89.766
Robust  acc:   146 /   180 =  81.111
------------------------------------
Accuracies by groups:
0, 0  acc:  8472 /  9767 =  86.741
0, 1  acc:  6906 /  7535 =  91.652
1, 0  acc:  2395 /  2480 =  96.573
1, 1  acc:   146 /   180 =  81.111
------------------------------------
Average acc: 17919 / 19962 =  89.766
Robust  acc:   146 /   180 =  81.111
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8472 /  9767 =  86.741
0, 1  acc:  6906 /  7535 =  91.652
1, 0  acc:  2395 /  2480 =  96.573
1, 1  acc:   146 /   180 =  81.111
------------------------------------
Average acc: 17919 / 19962 =  89.766
Robust  acc:   146 /   180 =  81.111
------------------------------------
Epoch:  28 | Train Loss: 0.002 | Train Acc: 93.476 | Val Loss: 0.006 | Val Acc: 67.454
Training:
Accuracies by groups:
0, 0  acc: 13835 / 20176 =  68.572
0, 1  acc:  7426 /  9412 =  78.899
1, 0  acc: 123302 / 125049 =  98.603
1, 1  acc:  7588 /  8133 =  93.299
--------------------------------------
Average acc: 152151 / 162770 =  93.476
Robust  acc: 13835 / 20176 =  68.572
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4886 /  8535 =  57.247
0, 1  acc:  5467 /  8276 =  66.058
1, 0  acc:  2867 /  2874 =  99.756
1, 1  acc:   181 /   182 =  99.451
------------------------------------
Average acc: 13401 / 19867 =  67.454
Robust  acc:  4886 /  8535 =  57.247
------------------------------------
-------------------------------------------
Avg Test Loss: 0.006 | Avg Test Acc: 69.392
Robust Acc: 63.919 | Best Acc: 99.597
-------------------------------------
Training, Epoch 27:
Accuracies by groups:
0, 0  acc:  6243 /  9767 =  63.919
0, 1  acc:  4962 /  7535 =  65.853
1, 0  acc:  2470 /  2480 =  99.597
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 13852 / 19962 =  69.392
Robust  acc:  6243 /  9767 =  63.919
------------------------------------
Accuracies by groups:
0, 0  acc:  6243 /  9767 =  63.919
0, 1  acc:  4962 /  7535 =  65.853
1, 0  acc:  2470 /  2480 =  99.597
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 13852 / 19962 =  69.392
Robust  acc:  6243 /  9767 =  63.919
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6243 /  9767 =  63.919
0, 1  acc:  4962 /  7535 =  65.853
1, 0  acc:  2470 /  2480 =  99.597
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 13852 / 19962 =  69.392
Robust  acc:  6243 /  9767 =  63.919
------------------------------------
Epoch:  29 | Train Loss: 0.002 | Train Acc: 93.187 | Val Loss: 0.006 | Val Acc: 69.331
Training:
Accuracies by groups:
0, 0  acc: 13747 / 20416 =  67.334
0, 1  acc:  7523 /  9567 =  78.635
1, 0  acc: 122875 / 124734 =  98.510
1, 1  acc:  7535 /  8053 =  93.568
--------------------------------------
Average acc: 151680 / 162770 =  93.187
Robust  acc: 13747 / 20416 =  67.334
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5196 /  8535 =  60.879
0, 1  acc:  5533 /  8276 =  66.856
1, 0  acc:  2863 /  2874 =  99.617
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc: 13774 / 19867 =  69.331
Robust  acc:  5196 /  8535 =  60.879
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 71.290
Robust Acc: 66.714 | Best Acc: 99.516
-------------------------------------
Training, Epoch 28:
Accuracies by groups:
0, 0  acc:  6516 /  9767 =  66.714
0, 1  acc:  5070 /  7535 =  67.286
1, 0  acc:  2468 /  2480 =  99.516
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 14231 / 19962 =  71.290
Robust  acc:  6516 /  9767 =  66.714
------------------------------------
Accuracies by groups:
0, 0  acc:  6516 /  9767 =  66.714
0, 1  acc:  5070 /  7535 =  67.286
1, 0  acc:  2468 /  2480 =  99.516
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 14231 / 19962 =  71.290
Robust  acc:  6516 /  9767 =  66.714
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6516 /  9767 =  66.714
0, 1  acc:  5070 /  7535 =  67.286
1, 0  acc:  2468 /  2480 =  99.516
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 14231 / 19962 =  71.290
Robust  acc:  6516 /  9767 =  66.714
------------------------------------
Epoch:  30 | Train Loss: 0.002 | Train Acc: 93.273 | Val Loss: 0.004 | Val Acc: 82.871
Training:
Accuracies by groups:
0, 0  acc: 13567 / 20351 =  66.665
0, 1  acc:  7344 /  9339 =  78.638
1, 0  acc: 123198 / 124861 =  98.668
1, 1  acc:  7711 /  8219 =  93.819
--------------------------------------
Average acc: 151820 / 162770 =  93.273
Robust  acc: 13567 / 20351 =  66.665
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6355 /  8535 =  74.458
0, 1  acc:  7093 /  8276 =  85.706
1, 0  acc:  2839 /  2874 =  98.782
1, 1  acc:   177 /   182 =  97.253
------------------------------------
Average acc: 16464 / 19867 =  82.871
Robust  acc:  6355 /  8535 =  74.458
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 84.806
Robust Acc: 79.451 | Best Acc: 98.548
-------------------------------------
Training, Epoch 29:
Accuracies by groups:
0, 0  acc:  7760 /  9767 =  79.451
0, 1  acc:  6563 /  7535 =  87.100
1, 0  acc:  2444 /  2480 =  98.548
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 16929 / 19962 =  84.806
Robust  acc:  7760 /  9767 =  79.451
------------------------------------
Accuracies by groups:
0, 0  acc:  7760 /  9767 =  79.451
0, 1  acc:  6563 /  7535 =  87.100
1, 0  acc:  2444 /  2480 =  98.548
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 16929 / 19962 =  84.806
Robust  acc:  7760 /  9767 =  79.451
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7760 /  9767 =  79.451
0, 1  acc:  6563 /  7535 =  87.100
1, 0  acc:  2444 /  2480 =  98.548
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 16929 / 19962 =  84.806
Robust  acc:  7760 /  9767 =  79.451
------------------------------------
Epoch:  31 | Train Loss: 0.002 | Train Acc: 93.264 | Val Loss: 0.007 | Val Acc: 59.949
Training:
Accuracies by groups:
0, 0  acc: 13529 / 20184 =  67.028
0, 1  acc:  7539 /  9678 =  77.898
1, 0  acc: 123020 / 124637 =  98.703
1, 1  acc:  7717 /  8271 =  93.302
--------------------------------------
Average acc: 151805 / 162770 =  93.264
Robust  acc: 13529 / 20184 =  67.028
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4106 /  8535 =  48.108
0, 1  acc:  4753 /  8276 =  57.431
1, 0  acc:  2869 /  2874 =  99.826
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc: 11910 / 19867 =  59.949
Robust  acc:  4106 /  8535 =  48.108
------------------------------------
-------------------------------------------
Avg Test Loss: 0.007 | Avg Test Acc: 61.792
Robust Acc: 54.643 | Best Acc: 99.758
-------------------------------------
Training, Epoch 30:
Accuracies by groups:
0, 0  acc:  5337 /  9767 =  54.643
0, 1  acc:  4346 /  7535 =  57.678
1, 0  acc:  2474 /  2480 =  99.758
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 12335 / 19962 =  61.792
Robust  acc:  5337 /  9767 =  54.643
------------------------------------
Accuracies by groups:
0, 0  acc:  5337 /  9767 =  54.643
0, 1  acc:  4346 /  7535 =  57.678
1, 0  acc:  2474 /  2480 =  99.758
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 12335 / 19962 =  61.792
Robust  acc:  5337 /  9767 =  54.643
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  5337 /  9767 =  54.643
0, 1  acc:  4346 /  7535 =  57.678
1, 0  acc:  2474 /  2480 =  99.758
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 12335 / 19962 =  61.792
Robust  acc:  5337 /  9767 =  54.643
------------------------------------
Epoch:  32 | Train Loss: 0.002 | Train Acc: 93.200 | Val Loss: 0.004 | Val Acc: 87.160
Training:
Accuracies by groups:
0, 0  acc: 13612 / 20355 =  66.873
0, 1  acc:  7341 /  9397 =  78.121
1, 0  acc: 123012 / 124772 =  98.589
1, 1  acc:  7736 /  8246 =  93.815
--------------------------------------
Average acc: 151701 / 162770 =  93.200
Robust  acc: 13612 / 20355 =  66.873
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6962 /  8535 =  81.570
0, 1  acc:  7417 /  8276 =  89.621
1, 0  acc:  2781 /  2874 =  96.764
1, 1  acc:   156 /   182 =  85.714
------------------------------------
Average acc: 17316 / 19867 =  87.160
Robust  acc:  6962 /  8535 =  81.570
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 88.518
Robust Acc: 77.778 | Best Acc: 95.605
-------------------------------------
Training, Epoch 31:
Accuracies by groups:
0, 0  acc:  8413 /  9767 =  86.137
0, 1  acc:  6746 /  7535 =  89.529
1, 0  acc:  2371 /  2480 =  95.605
1, 1  acc:   140 /   180 =  77.778
------------------------------------
Average acc: 17670 / 19962 =  88.518
Robust  acc:   140 /   180 =  77.778
------------------------------------
Accuracies by groups:
0, 0  acc:  8413 /  9767 =  86.137
0, 1  acc:  6746 /  7535 =  89.529
1, 0  acc:  2371 /  2480 =  95.605
1, 1  acc:   140 /   180 =  77.778
------------------------------------
Average acc: 17670 / 19962 =  88.518
Robust  acc:   140 /   180 =  77.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8413 /  9767 =  86.137
0, 1  acc:  6746 /  7535 =  89.529
1, 0  acc:  2371 /  2480 =  95.605
1, 1  acc:   140 /   180 =  77.778
------------------------------------
Average acc: 17670 / 19962 =  88.518
Robust  acc:   140 /   180 =  77.778
------------------------------------
Epoch:  33 | Train Loss: 0.002 | Train Acc: 93.332 | Val Loss: 0.006 | Val Acc: 64.308
Training:
Accuracies by groups:
0, 0  acc: 13586 / 20302 =  66.920
0, 1  acc:  7154 /  9209 =  77.685
1, 0  acc: 123357 / 124954 =  98.722
1, 1  acc:  7819 /  8305 =  94.148
--------------------------------------
Average acc: 151916 / 162770 =  93.332
Robust  acc: 13586 / 20302 =  66.920
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4317 /  8535 =  50.580
0, 1  acc:  5413 /  8276 =  65.406
1, 0  acc:  2865 /  2874 =  99.687
1, 1  acc:   181 /   182 =  99.451
------------------------------------
Average acc: 12776 / 19867 =  64.308
Robust  acc:  4317 /  8535 =  50.580
------------------------------------
-------------------------------------------
Avg Test Loss: 0.006 | Avg Test Acc: 66.241
Robust Acc: 57.152 | Best Acc: 99.556
-------------------------------------
Training, Epoch 32:
Accuracies by groups:
0, 0  acc:  5582 /  9767 =  57.152
0, 1  acc:  4994 /  7535 =  66.277
1, 0  acc:  2469 /  2480 =  99.556
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 13223 / 19962 =  66.241
Robust  acc:  5582 /  9767 =  57.152
------------------------------------
Accuracies by groups:
0, 0  acc:  5582 /  9767 =  57.152
0, 1  acc:  4994 /  7535 =  66.277
1, 0  acc:  2469 /  2480 =  99.556
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 13223 / 19962 =  66.241
Robust  acc:  5582 /  9767 =  57.152
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  5582 /  9767 =  57.152
0, 1  acc:  4994 /  7535 =  66.277
1, 0  acc:  2469 /  2480 =  99.556
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 13223 / 19962 =  66.241
Robust  acc:  5582 /  9767 =  57.152
------------------------------------
Epoch:  34 | Train Loss: 0.002 | Train Acc: 93.213 | Val Loss: 0.004 | Val Acc: 78.910
Training:
Accuracies by groups:
0, 0  acc: 13499 / 20243 =  66.685
0, 1  acc:  7428 /  9587 =  77.480
1, 0  acc: 123252 / 124899 =  98.681
1, 1  acc:  7544 /  8041 =  93.819
--------------------------------------
Average acc: 151723 / 162770 =  93.213
Robust  acc: 13499 / 20243 =  66.685
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5961 /  8535 =  69.842
0, 1  acc:  6687 /  8276 =  80.800
1, 0  acc:  2849 /  2874 =  99.130
1, 1  acc:   180 /   182 =  98.901
------------------------------------
Average acc: 15677 / 19867 =  78.910
Robust  acc:  5961 /  8535 =  69.842
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 80.733
Robust Acc: 75.489 | Best Acc: 98.750
-------------------------------------
Training, Epoch 33:
Accuracies by groups:
0, 0  acc:  7373 /  9767 =  75.489
0, 1  acc:  6126 /  7535 =  81.301
1, 0  acc:  2449 /  2480 =  98.750
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 16116 / 19962 =  80.733
Robust  acc:  7373 /  9767 =  75.489
------------------------------------
Accuracies by groups:
0, 0  acc:  7373 /  9767 =  75.489
0, 1  acc:  6126 /  7535 =  81.301
1, 0  acc:  2449 /  2480 =  98.750
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 16116 / 19962 =  80.733
Robust  acc:  7373 /  9767 =  75.489
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7373 /  9767 =  75.489
0, 1  acc:  6126 /  7535 =  81.301
1, 0  acc:  2449 /  2480 =  98.750
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 16116 / 19962 =  80.733
Robust  acc:  7373 /  9767 =  75.489
------------------------------------
Epoch:  35 | Train Loss: 0.002 | Train Acc: 93.252 | Val Loss: 0.008 | Val Acc: 46.862
Training:
Accuracies by groups:
0, 0  acc: 13682 / 20360 =  67.200
0, 1  acc:  7325 /  9415 =  77.801
1, 0  acc: 122908 / 124591 =  98.649
1, 1  acc:  7871 /  8404 =  93.658
--------------------------------------
Average acc: 151786 / 162770 =  93.252
Robust  acc: 13682 / 20360 =  67.200
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  2601 /  8535 =  30.475
0, 1  acc:  3656 /  8276 =  44.176
1, 0  acc:  2871 /  2874 =  99.896
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc:  9310 / 19867 =  46.862
Robust  acc:  2601 /  8535 =  30.475
------------------------------------
-------------------------------------------
Avg Test Loss: 0.008 | Avg Test Acc: 47.686
Robust Acc: 35.753 | Best Acc: 99.919
-------------------------------------
Training, Epoch 34:
Accuracies by groups:
0, 0  acc:  3492 /  9767 =  35.753
0, 1  acc:  3371 /  7535 =  44.738
1, 0  acc:  2478 /  2480 =  99.919
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  9519 / 19962 =  47.686
Robust  acc:  3492 /  9767 =  35.753
------------------------------------
Accuracies by groups:
0, 0  acc:  3492 /  9767 =  35.753
0, 1  acc:  3371 /  7535 =  44.738
1, 0  acc:  2478 /  2480 =  99.919
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  9519 / 19962 =  47.686
Robust  acc:  3492 /  9767 =  35.753
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  3492 /  9767 =  35.753
0, 1  acc:  3371 /  7535 =  44.738
1, 0  acc:  2478 /  2480 =  99.919
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  9519 / 19962 =  47.686
Robust  acc:  3492 /  9767 =  35.753
------------------------------------
Epoch:  36 | Train Loss: 0.002 | Train Acc: 93.265 | Val Loss: 0.003 | Val Acc: 92.369
Training:
Accuracies by groups:
0, 0  acc: 13802 / 20474 =  67.412
0, 1  acc:  7361 /  9492 =  77.550
1, 0  acc: 122797 / 124446 =  98.675
1, 1  acc:  7847 /  8358 =  93.886
--------------------------------------
Average acc: 151807 / 162770 =  93.265
Robust  acc: 13802 / 20474 =  67.412
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7933 /  8535 =  92.947
0, 1  acc:  7950 /  8276 =  96.061
1, 0  acc:  2363 /  2874 =  82.220
1, 1  acc:   105 /   182 =  57.692
------------------------------------
Average acc: 18351 / 19867 =  92.369
Robust  acc:   105 /   182 =  57.692
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 93.082
Robust Acc: 53.333 | Best Acc: 96.032
-------------------------------------
Training, Epoch 35:
Accuracies by groups:
0, 0  acc:  9267 /  9767 =  94.881
0, 1  acc:  7236 /  7535 =  96.032
1, 0  acc:  1982 /  2480 =  79.919
1, 1  acc:    96 /   180 =  53.333
------------------------------------
Average acc: 18581 / 19962 =  93.082
Robust  acc:    96 /   180 =  53.333
------------------------------------
Accuracies by groups:
0, 0  acc:  9267 /  9767 =  94.881
0, 1  acc:  7236 /  7535 =  96.032
1, 0  acc:  1982 /  2480 =  79.919
1, 1  acc:    96 /   180 =  53.333
------------------------------------
Average acc: 18581 / 19962 =  93.082
Robust  acc:    96 /   180 =  53.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9267 /  9767 =  94.881
0, 1  acc:  7236 /  7535 =  96.032
1, 0  acc:  1982 /  2480 =  79.919
1, 1  acc:    96 /   180 =  53.333
------------------------------------
Average acc: 18581 / 19962 =  93.082
Robust  acc:    96 /   180 =  53.333
------------------------------------
Epoch:  37 | Train Loss: 0.002 | Train Acc: 93.361 | Val Loss: 0.005 | Val Acc: 76.247
Training:
Accuracies by groups:
0, 0  acc: 13657 / 20388 =  66.985
0, 1  acc:  7440 /  9460 =  78.647
1, 0  acc: 123248 / 124844 =  98.722
1, 1  acc:  7619 /  8078 =  94.318
--------------------------------------
Average acc: 151964 / 162770 =  93.361
Robust  acc: 13657 / 20388 =  66.985
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5852 /  8535 =  68.565
0, 1  acc:  6265 /  8276 =  75.701
1, 0  acc:  2852 /  2874 =  99.235
1, 1  acc:   179 /   182 =  98.352
------------------------------------
Average acc: 15148 / 19867 =  76.247
Robust  acc:  5852 /  8535 =  68.565
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 78.264
Robust Acc: 74.363 | Best Acc: 99.032
-------------------------------------
Training, Epoch 36:
Accuracies by groups:
0, 0  acc:  7263 /  9767 =  74.363
0, 1  acc:  5734 /  7535 =  76.098
1, 0  acc:  2456 /  2480 =  99.032
1, 1  acc:   170 /   180 =  94.444
------------------------------------
Average acc: 15623 / 19962 =  78.264
Robust  acc:  7263 /  9767 =  74.363
------------------------------------
Accuracies by groups:
0, 0  acc:  7263 /  9767 =  74.363
0, 1  acc:  5734 /  7535 =  76.098
1, 0  acc:  2456 /  2480 =  99.032
1, 1  acc:   170 /   180 =  94.444
------------------------------------
Average acc: 15623 / 19962 =  78.264
Robust  acc:  7263 /  9767 =  74.363
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7263 /  9767 =  74.363
0, 1  acc:  5734 /  7535 =  76.098
1, 0  acc:  2456 /  2480 =  99.032
1, 1  acc:   170 /   180 =  94.444
------------------------------------
Average acc: 15623 / 19962 =  78.264
Robust  acc:  7263 /  9767 =  74.363
------------------------------------
Epoch:  38 | Train Loss: 0.002 | Train Acc: 93.319 | Val Loss: 0.004 | Val Acc: 83.908
Training:
Accuracies by groups:
0, 0  acc: 13848 / 20538 =  67.426
0, 1  acc:  7308 /  9298 =  78.598
1, 0  acc: 123024 / 124710 =  98.648
1, 1  acc:  7715 /  8224 =  93.811
--------------------------------------
Average acc: 151895 / 162770 =  93.319
Robust  acc: 13848 / 20538 =  67.426
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6689 /  8535 =  78.371
0, 1  acc:  6984 /  8276 =  84.389
1, 0  acc:  2826 /  2874 =  98.330
1, 1  acc:   171 /   182 =  93.956
------------------------------------
Average acc: 16670 / 19867 =  83.908
Robust  acc:  6689 /  8535 =  78.371
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 85.517
Robust Acc: 82.707 | Best Acc: 97.661
-------------------------------------
Training, Epoch 37:
Accuracies by groups:
0, 0  acc:  8078 /  9767 =  82.707
0, 1  acc:  6414 /  7535 =  85.123
1, 0  acc:  2422 /  2480 =  97.661
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 17071 / 19962 =  85.517
Robust  acc:  8078 /  9767 =  82.707
------------------------------------
Accuracies by groups:
0, 0  acc:  8078 /  9767 =  82.707
0, 1  acc:  6414 /  7535 =  85.123
1, 0  acc:  2422 /  2480 =  97.661
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 17071 / 19962 =  85.517
Robust  acc:  8078 /  9767 =  82.707
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8078 /  9767 =  82.707
0, 1  acc:  6414 /  7535 =  85.123
1, 0  acc:  2422 /  2480 =  97.661
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 17071 / 19962 =  85.517
Robust  acc:  8078 /  9767 =  82.707
------------------------------------
Epoch:  39 | Train Loss: 0.002 | Train Acc: 93.260 | Val Loss: 0.004 | Val Acc: 81.980
Training:
Accuracies by groups:
0, 0  acc: 13507 / 20303 =  66.527
0, 1  acc:  7503 /  9535 =  78.689
1, 0  acc: 123358 / 124981 =  98.701
1, 1  acc:  7432 /  7951 =  93.473
--------------------------------------
Average acc: 151800 / 162770 =  93.260
Robust  acc: 13507 / 20303 =  66.527
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6444 /  8535 =  75.501
0, 1  acc:  6835 /  8276 =  82.588
1, 0  acc:  2830 /  2874 =  98.469
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 16287 / 19867 =  81.980
Robust  acc:  6444 /  8535 =  75.501
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 83.774
Robust Acc: 80.106 | Best Acc: 98.226
-------------------------------------
Training, Epoch 38:
Accuracies by groups:
0, 0  acc:  7824 /  9767 =  80.106
0, 1  acc:  6294 /  7535 =  83.530
1, 0  acc:  2436 /  2480 =  98.226
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 16723 / 19962 =  83.774
Robust  acc:  7824 /  9767 =  80.106
------------------------------------
Accuracies by groups:
0, 0  acc:  7824 /  9767 =  80.106
0, 1  acc:  6294 /  7535 =  83.530
1, 0  acc:  2436 /  2480 =  98.226
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 16723 / 19962 =  83.774
Robust  acc:  7824 /  9767 =  80.106
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7824 /  9767 =  80.106
0, 1  acc:  6294 /  7535 =  83.530
1, 0  acc:  2436 /  2480 =  98.226
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 16723 / 19962 =  83.774
Robust  acc:  7824 /  9767 =  80.106
------------------------------------
Epoch:  40 | Train Loss: 0.002 | Train Acc: 93.357 | Val Loss: 0.004 | Val Acc: 87.185
Training:
Accuracies by groups:
0, 0  acc: 13677 / 20342 =  67.235
0, 1  acc:  7509 /  9558 =  78.562
1, 0  acc: 122980 / 124606 =  98.695
1, 1  acc:  7791 /  8264 =  94.276
--------------------------------------
Average acc: 151957 / 162770 =  93.357
Robust  acc: 13677 / 20342 =  67.235
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6949 /  8535 =  81.418
0, 1  acc:  7417 /  8276 =  89.621
1, 0  acc:  2791 /  2874 =  97.112
1, 1  acc:   164 /   182 =  90.110
------------------------------------
Average acc: 17321 / 19867 =  87.185
Robust  acc:  6949 /  8535 =  81.418
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 88.458
Robust Acc: 82.222 | Best Acc: 96.734
-------------------------------------
Training, Epoch 39:
Accuracies by groups:
0, 0  acc:  8322 /  9767 =  85.205
0, 1  acc:  6789 /  7535 =  90.100
1, 0  acc:  2399 /  2480 =  96.734
1, 1  acc:   148 /   180 =  82.222
------------------------------------
Average acc: 17658 / 19962 =  88.458
Robust  acc:   148 /   180 =  82.222
------------------------------------
Accuracies by groups:
0, 0  acc:  8322 /  9767 =  85.205
0, 1  acc:  6789 /  7535 =  90.100
1, 0  acc:  2399 /  2480 =  96.734
1, 1  acc:   148 /   180 =  82.222
------------------------------------
Average acc: 17658 / 19962 =  88.458
Robust  acc:   148 /   180 =  82.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8322 /  9767 =  85.205
0, 1  acc:  6789 /  7535 =  90.100
1, 0  acc:  2399 /  2480 =  96.734
1, 1  acc:   148 /   180 =  82.222
------------------------------------
Average acc: 17658 / 19962 =  88.458
Robust  acc:   148 /   180 =  82.222
------------------------------------
Epoch:  41 | Train Loss: 0.002 | Train Acc: 93.454 | Val Loss: 0.004 | Val Acc: 80.415
Training:
Accuracies by groups:
0, 0  acc: 13520 / 20128 =  67.170
0, 1  acc:  7472 /  9442 =  79.136
1, 0  acc: 123361 / 124949 =  98.729
1, 1  acc:  7762 /  8251 =  94.073
--------------------------------------
Average acc: 152115 / 162770 =  93.454
Robust  acc: 13520 / 20128 =  67.170
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6348 /  8535 =  74.376
0, 1  acc:  6620 /  8276 =  79.990
1, 0  acc:  2832 /  2874 =  98.539
1, 1  acc:   176 /   182 =  96.703
------------------------------------
Average acc: 15976 / 19867 =  80.415
Robust  acc:  6348 /  8535 =  74.376
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 82.372
Robust Acc: 79.390 | Best Acc: 98.306
-------------------------------------
Training, Epoch 40:
Accuracies by groups:
0, 0  acc:  7754 /  9767 =  79.390
0, 1  acc:  6083 /  7535 =  80.730
1, 0  acc:  2438 /  2480 =  98.306
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 16443 / 19962 =  82.372
Robust  acc:  7754 /  9767 =  79.390
------------------------------------
Accuracies by groups:
0, 0  acc:  7754 /  9767 =  79.390
0, 1  acc:  6083 /  7535 =  80.730
1, 0  acc:  2438 /  2480 =  98.306
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 16443 / 19962 =  82.372
Robust  acc:  7754 /  9767 =  79.390
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7754 /  9767 =  79.390
0, 1  acc:  6083 /  7535 =  80.730
1, 0  acc:  2438 /  2480 =  98.306
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 16443 / 19962 =  82.372
Robust  acc:  7754 /  9767 =  79.390
------------------------------------
Epoch:  42 | Train Loss: 0.002 | Train Acc: 93.502 | Val Loss: 0.004 | Val Acc: 80.566
Training:
Accuracies by groups:
0, 0  acc: 13843 / 20354 =  68.011
0, 1  acc:  7480 /  9491 =  78.812
1, 0  acc: 123388 / 124943 =  98.755
1, 1  acc:  7482 /  7982 =  93.736
--------------------------------------
Average acc: 152193 / 162770 =  93.502
Robust  acc: 13843 / 20354 =  68.011
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6157 /  8535 =  72.138
0, 1  acc:  6832 /  8276 =  82.552
1, 0  acc:  2840 /  2874 =  98.817
1, 1  acc:   177 /   182 =  97.253
------------------------------------
Average acc: 16006 / 19867 =  80.566
Robust  acc:  6157 /  8535 =  72.138
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 82.121
Robust Acc: 77.588 | Best Acc: 98.387
-------------------------------------
Training, Epoch 41:
Accuracies by groups:
0, 0  acc:  7578 /  9767 =  77.588
0, 1  acc:  6212 /  7535 =  82.442
1, 0  acc:  2440 /  2480 =  98.387
1, 1  acc:   163 /   180 =  90.556
------------------------------------
Average acc: 16393 / 19962 =  82.121
Robust  acc:  7578 /  9767 =  77.588
------------------------------------
Accuracies by groups:
0, 0  acc:  7578 /  9767 =  77.588
0, 1  acc:  6212 /  7535 =  82.442
1, 0  acc:  2440 /  2480 =  98.387
1, 1  acc:   163 /   180 =  90.556
------------------------------------
Average acc: 16393 / 19962 =  82.121
Robust  acc:  7578 /  9767 =  77.588
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7578 /  9767 =  77.588
0, 1  acc:  6212 /  7535 =  82.442
1, 0  acc:  2440 /  2480 =  98.387
1, 1  acc:   163 /   180 =  90.556
------------------------------------
Average acc: 16393 / 19962 =  82.121
Robust  acc:  7578 /  9767 =  77.588
------------------------------------
Epoch:  43 | Train Loss: 0.002 | Train Acc: 93.441 | Val Loss: 0.005 | Val Acc: 76.252
Training:
Accuracies by groups:
0, 0  acc: 13560 / 20165 =  67.245
0, 1  acc:  7354 /  9381 =  78.392
1, 0  acc: 123508 / 125026 =  98.786
1, 1  acc:  7672 /  8198 =  93.584
--------------------------------------
Average acc: 152094 / 162770 =  93.441
Robust  acc: 13560 / 20165 =  67.245
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5780 /  8535 =  67.721
0, 1  acc:  6336 /  8276 =  76.559
1, 0  acc:  2855 /  2874 =  99.339
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 15149 / 19867 =  76.252
Robust  acc:  5780 /  8535 =  67.721
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 78.659
Robust Acc: 73.830 | Best Acc: 99.355
-------------------------------------
Training, Epoch 42:
Accuracies by groups:
0, 0  acc:  7211 /  9767 =  73.830
0, 1  acc:  5857 /  7535 =  77.731
1, 0  acc:  2464 /  2480 =  99.355
1, 1  acc:   170 /   180 =  94.444
------------------------------------
Average acc: 15702 / 19962 =  78.659
Robust  acc:  7211 /  9767 =  73.830
------------------------------------
Accuracies by groups:
0, 0  acc:  7211 /  9767 =  73.830
0, 1  acc:  5857 /  7535 =  77.731
1, 0  acc:  2464 /  2480 =  99.355
1, 1  acc:   170 /   180 =  94.444
------------------------------------
Average acc: 15702 / 19962 =  78.659
Robust  acc:  7211 /  9767 =  73.830
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7211 /  9767 =  73.830
0, 1  acc:  5857 /  7535 =  77.731
1, 0  acc:  2464 /  2480 =  99.355
1, 1  acc:   170 /   180 =  94.444
------------------------------------
Average acc: 15702 / 19962 =  78.659
Robust  acc:  7211 /  9767 =  73.830
------------------------------------
Epoch:  44 | Train Loss: 0.002 | Train Acc: 93.413 | Val Loss: 0.005 | Val Acc: 72.683
Training:
Accuracies by groups:
0, 0  acc: 13631 / 20159 =  67.617
0, 1  acc:  7514 /  9556 =  78.631
1, 0  acc: 123226 / 124871 =  98.683
1, 1  acc:  7677 /  8184 =  93.805
--------------------------------------
Average acc: 152048 / 162770 =  93.413
Robust  acc: 13631 / 20159 =  67.617
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5325 /  8535 =  62.390
0, 1  acc:  6074 /  8276 =  73.393
1, 0  acc:  2863 /  2874 =  99.617
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 14440 / 19867 =  72.683
Robust  acc:  5325 /  8535 =  62.390
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 74.707
Robust Acc: 68.875 | Best Acc: 99.315
-------------------------------------
Training, Epoch 43:
Accuracies by groups:
0, 0  acc:  6727 /  9767 =  68.875
0, 1  acc:  5549 /  7535 =  73.643
1, 0  acc:  2463 /  2480 =  99.315
1, 1  acc:   174 /   180 =  96.667
------------------------------------
Average acc: 14913 / 19962 =  74.707
Robust  acc:  6727 /  9767 =  68.875
------------------------------------
Accuracies by groups:
0, 0  acc:  6727 /  9767 =  68.875
0, 1  acc:  5549 /  7535 =  73.643
1, 0  acc:  2463 /  2480 =  99.315
1, 1  acc:   174 /   180 =  96.667
------------------------------------
Average acc: 14913 / 19962 =  74.707
Robust  acc:  6727 /  9767 =  68.875
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6727 /  9767 =  68.875
0, 1  acc:  5549 /  7535 =  73.643
1, 0  acc:  2463 /  2480 =  99.315
1, 1  acc:   174 /   180 =  96.667
------------------------------------
Average acc: 14913 / 19962 =  74.707
Robust  acc:  6727 /  9767 =  68.875
------------------------------------
Epoch:  45 | Train Loss: 0.002 | Train Acc: 93.544 | Val Loss: 0.003 | Val Acc: 92.968
Training:
Accuracies by groups:
0, 0  acc: 13865 / 20322 =  68.227
0, 1  acc:  7377 /  9328 =  79.084
1, 0  acc: 123346 / 124940 =  98.724
1, 1  acc:  7674 /  8180 =  93.814
--------------------------------------
Average acc: 152262 / 162770 =  93.544
Robust  acc: 13865 / 20322 =  68.227
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7704 /  8535 =  90.264
0, 1  acc:  8004 /  8276 =  96.713
1, 0  acc:  2639 /  2874 =  91.823
1, 1  acc:   123 /   182 =  67.582
------------------------------------
Average acc: 18470 / 19867 =  92.968
Robust  acc:   123 /   182 =  67.582
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 93.713
Robust Acc: 66.667 | Best Acc: 96.669
-------------------------------------
Training, Epoch 44:
Accuracies by groups:
0, 0  acc:  9063 /  9767 =  92.792
0, 1  acc:  7284 /  7535 =  96.669
1, 0  acc:  2240 /  2480 =  90.323
1, 1  acc:   120 /   180 =  66.667
------------------------------------
Average acc: 18707 / 19962 =  93.713
Robust  acc:   120 /   180 =  66.667
------------------------------------
Accuracies by groups:
0, 0  acc:  9063 /  9767 =  92.792
0, 1  acc:  7284 /  7535 =  96.669
1, 0  acc:  2240 /  2480 =  90.323
1, 1  acc:   120 /   180 =  66.667
------------------------------------
Average acc: 18707 / 19962 =  93.713
Robust  acc:   120 /   180 =  66.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9063 /  9767 =  92.792
0, 1  acc:  7284 /  7535 =  96.669
1, 0  acc:  2240 /  2480 =  90.323
1, 1  acc:   120 /   180 =  66.667
------------------------------------
Average acc: 18707 / 19962 =  93.713
Robust  acc:   120 /   180 =  66.667
------------------------------------
Epoch:  46 | Train Loss: 0.002 | Train Acc: 93.383 | Val Loss: 0.004 | Val Acc: 82.871
Training:
Accuracies by groups:
0, 0  acc: 13666 / 20433 =  66.882
0, 1  acc:  7463 /  9393 =  79.453
1, 0  acc: 123028 / 124619 =  98.723
1, 1  acc:  7842 /  8325 =  94.198
--------------------------------------
Average acc: 151999 / 162770 =  93.383
Robust  acc: 13666 / 20433 =  66.882
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6504 /  8535 =  76.204
0, 1  acc:  6957 /  8276 =  84.062
1, 0  acc:  2829 /  2874 =  98.434
1, 1  acc:   174 /   182 =  95.604
------------------------------------
Average acc: 16464 / 19867 =  82.871
Robust  acc:  6504 /  8535 =  76.204
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 84.385
Robust Acc: 80.915 | Best Acc: 98.105
-------------------------------------
Training, Epoch 45:
Accuracies by groups:
0, 0  acc:  7903 /  9767 =  80.915
0, 1  acc:  6350 /  7535 =  84.273
1, 0  acc:  2433 /  2480 =  98.105
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 16845 / 19962 =  84.385
Robust  acc:  7903 /  9767 =  80.915
------------------------------------
Accuracies by groups:
0, 0  acc:  7903 /  9767 =  80.915
0, 1  acc:  6350 /  7535 =  84.273
1, 0  acc:  2433 /  2480 =  98.105
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 16845 / 19962 =  84.385
Robust  acc:  7903 /  9767 =  80.915
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7903 /  9767 =  80.915
0, 1  acc:  6350 /  7535 =  84.273
1, 0  acc:  2433 /  2480 =  98.105
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 16845 / 19962 =  84.385
Robust  acc:  7903 /  9767 =  80.915
------------------------------------
Epoch:  47 | Train Loss: 0.002 | Train Acc: 93.472 | Val Loss: 0.006 | Val Acc: 68.098
Training:
Accuracies by groups:
0, 0  acc: 13774 / 20354 =  67.672
0, 1  acc:  7450 /  9414 =  79.137
1, 0  acc: 123240 / 124839 =  98.719
1, 1  acc:  7681 /  8163 =  94.095
--------------------------------------
Average acc: 152145 / 162770 =  93.472
Robust  acc: 13774 / 20354 =  67.672
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4894 /  8535 =  57.340
0, 1  acc:  5588 /  8276 =  67.521
1, 0  acc:  2866 /  2874 =  99.722
1, 1  acc:   181 /   182 =  99.451
------------------------------------
Average acc: 13529 / 19867 =  68.098
Robust  acc:  4894 /  8535 =  57.340
------------------------------------
-------------------------------------------
Avg Test Loss: 0.006 | Avg Test Acc: 70.309
Robust Acc: 64.482 | Best Acc: 99.435
-------------------------------------
Training, Epoch 46:
Accuracies by groups:
0, 0  acc:  6298 /  9767 =  64.482
0, 1  acc:  5095 /  7535 =  67.618
1, 0  acc:  2466 /  2480 =  99.435
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 14035 / 19962 =  70.309
Robust  acc:  6298 /  9767 =  64.482
------------------------------------
Accuracies by groups:
0, 0  acc:  6298 /  9767 =  64.482
0, 1  acc:  5095 /  7535 =  67.618
1, 0  acc:  2466 /  2480 =  99.435
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 14035 / 19962 =  70.309
Robust  acc:  6298 /  9767 =  64.482
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6298 /  9767 =  64.482
0, 1  acc:  5095 /  7535 =  67.618
1, 0  acc:  2466 /  2480 =  99.435
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 14035 / 19962 =  70.309
Robust  acc:  6298 /  9767 =  64.482
------------------------------------
Epoch:  48 | Train Loss: 0.002 | Train Acc: 93.514 | Val Loss: 0.003 | Val Acc: 94.352
Training:
Accuracies by groups:
0, 0  acc: 13738 / 20351 =  67.505
0, 1  acc:  7559 /  9486 =  79.686
1, 0  acc: 123041 / 124572 =  98.771
1, 1  acc:  7874 /  8361 =  94.175
--------------------------------------
Average acc: 152212 / 162770 =  93.514
Robust  acc: 13738 / 20351 =  67.505
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7951 /  8535 =  93.158
0, 1  acc:  8165 /  8276 =  98.659
1, 0  acc:  2530 /  2874 =  88.031
1, 1  acc:    99 /   182 =  54.396
------------------------------------
Average acc: 18745 / 19867 =  94.352
Robust  acc:    99 /   182 =  54.396
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 94.805
Robust Acc: 46.111 | Best Acc: 98.660
-------------------------------------
Training, Epoch 47:
Accuracies by groups:
0, 0  acc:  9259 /  9767 =  94.799
0, 1  acc:  7434 /  7535 =  98.660
1, 0  acc:  2149 /  2480 =  86.653
1, 1  acc:    83 /   180 =  46.111
------------------------------------
Average acc: 18925 / 19962 =  94.805
Robust  acc:    83 /   180 =  46.111
------------------------------------
Accuracies by groups:
0, 0  acc:  9259 /  9767 =  94.799
0, 1  acc:  7434 /  7535 =  98.660
1, 0  acc:  2149 /  2480 =  86.653
1, 1  acc:    83 /   180 =  46.111
------------------------------------
Average acc: 18925 / 19962 =  94.805
Robust  acc:    83 /   180 =  46.111
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9259 /  9767 =  94.799
0, 1  acc:  7434 /  7535 =  98.660
1, 0  acc:  2149 /  2480 =  86.653
1, 1  acc:    83 /   180 =  46.111
------------------------------------
Average acc: 18925 / 19962 =  94.805
Robust  acc:    83 /   180 =  46.111
------------------------------------
Epoch:  49 | Train Loss: 0.002 | Train Acc: 93.499 | Val Loss: 0.005 | Val Acc: 76.232
Training:
Accuracies by groups:
0, 0  acc: 13949 / 20465 =  68.160
0, 1  acc:  7489 /  9481 =  78.990
1, 0  acc: 122931 / 124452 =  98.778
1, 1  acc:  7819 /  8372 =  93.395
--------------------------------------
Average acc: 152188 / 162770 =  93.499
Robust  acc: 13949 / 20465 =  68.160
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5705 /  8535 =  66.842
0, 1  acc:  6409 /  8276 =  77.441
1, 0  acc:  2853 /  2874 =  99.269
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 15145 / 19867 =  76.232
Robust  acc:  5705 /  8535 =  66.842
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 78.329
Robust Acc: 72.182 | Best Acc: 99.395
-------------------------------------
Training, Epoch 48:
Accuracies by groups:
0, 0  acc:  7050 /  9767 =  72.182
0, 1  acc:  5952 /  7535 =  78.991
1, 0  acc:  2465 /  2480 =  99.395
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 15636 / 19962 =  78.329
Robust  acc:  7050 /  9767 =  72.182
------------------------------------
Accuracies by groups:
0, 0  acc:  7050 /  9767 =  72.182
0, 1  acc:  5952 /  7535 =  78.991
1, 0  acc:  2465 /  2480 =  99.395
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 15636 / 19962 =  78.329
Robust  acc:  7050 /  9767 =  72.182
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7050 /  9767 =  72.182
0, 1  acc:  5952 /  7535 =  78.991
1, 0  acc:  2465 /  2480 =  99.395
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 15636 / 19962 =  78.329
Robust  acc:  7050 /  9767 =  72.182
------------------------------------
Epoch:  50 | Train Loss: 0.002 | Train Acc: 93.507 | Val Loss: 0.004 | Val Acc: 88.378
Training:
Accuracies by groups:
0, 0  acc: 13881 / 20393 =  68.067
0, 1  acc:  7603 /  9530 =  79.780
1, 0  acc: 123067 / 124661 =  98.721
1, 1  acc:  7651 /  8186 =  93.464
--------------------------------------
Average acc: 152202 / 162770 =  93.507
Robust  acc: 13881 / 20393 =  68.067
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7131 /  8535 =  83.550
0, 1  acc:  7475 /  8276 =  90.321
1, 0  acc:  2787 /  2874 =  96.973
1, 1  acc:   165 /   182 =  90.659
------------------------------------
Average acc: 17558 / 19867 =  88.378
Robust  acc:  7131 /  8535 =  83.550
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 89.350
Robust Acc: 82.778 | Best Acc: 95.766
-------------------------------------
Training, Epoch 49:
Accuracies by groups:
0, 0  acc:  8512 /  9767 =  87.151
0, 1  acc:  6800 /  7535 =  90.246
1, 0  acc:  2375 /  2480 =  95.766
1, 1  acc:   149 /   180 =  82.778
------------------------------------
Average acc: 17836 / 19962 =  89.350
Robust  acc:   149 /   180 =  82.778
------------------------------------
Accuracies by groups:
0, 0  acc:  8512 /  9767 =  87.151
0, 1  acc:  6800 /  7535 =  90.246
1, 0  acc:  2375 /  2480 =  95.766
1, 1  acc:   149 /   180 =  82.778
------------------------------------
Average acc: 17836 / 19962 =  89.350
Robust  acc:   149 /   180 =  82.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8512 /  9767 =  87.151
0, 1  acc:  6800 /  7535 =  90.246
1, 0  acc:  2375 /  2480 =  95.766
1, 1  acc:   149 /   180 =  82.778
------------------------------------
Average acc: 17836 / 19962 =  89.350
Robust  acc:   149 /   180 =  82.778
------------------------------------
replace: True
-> Updating checkpoint debias-end_seed34.pt...
Checkpoint saved at ./model/celebA/config/debias-end_seed34.pt
