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/seed31/stage_one_erm_model_b_epoch0_seed31.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: 31
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=31-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/seed31/stage_one_erm_model_b_epoch0_seed31.pt
======
# Calculate probability ...
======
======
p_y_a:  tensor([[0.8182, 0.0327],
        [0.1254, 0.0237]])
p_y:  tensor([0.8509, 0.1491])
# Load biased model ...
======
Epoch:   1 | Train Loss: 0.002 | Train Acc: 88.965 | Val Loss: 0.003 | Val Acc: 84.155
Training:
Accuracies by groups:
0, 0  acc: 12790 / 23270 =  54.963
0, 1  acc:  6489 / 10087 =  64.330
1, 0  acc: 118197 / 121155 =  97.558
1, 1  acc:  7333 /  8258 =  88.799
--------------------------------------
Average acc: 144809 / 162770 =  88.965
Robust  acc: 12790 / 23270 =  54.963
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6608 /  8535 =  77.422
0, 1  acc:  7111 /  8276 =  85.923
1, 0  acc:  2834 /  2874 =  98.608
1, 1  acc:   166 /   182 =  91.209
------------------------------------
Average acc: 16719 / 19867 =  84.155
Robust  acc:  6608 /  8535 =  77.422
------------------------------------
New max robust acc: 77.4223784417106
debias model - Saving best checkpoint at epoch 0
replace: True
-> Updating checkpoint debias-wga-best_seed31.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed31.pt
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 85.838
Robust Acc: 82.298 | Best Acc: 98.427
-------------------------------------
Training, Epoch 0:
Accuracies by groups:
0, 0  acc:  8038 /  9767 =  82.298
0, 1  acc:  6491 /  7535 =  86.145
1, 0  acc:  2441 /  2480 =  98.427
1, 1  acc:   165 /   180 =  91.667
------------------------------------
Average acc: 17135 / 19962 =  85.838
Robust  acc:  8038 /  9767 =  82.298
------------------------------------
Accuracies by groups:
0, 0  acc:  8038 /  9767 =  82.298
0, 1  acc:  6491 /  7535 =  86.145
1, 0  acc:  2441 /  2480 =  98.427
1, 1  acc:   165 /   180 =  91.667
------------------------------------
Average acc: 17135 / 19962 =  85.838
Robust  acc:  8038 /  9767 =  82.298
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8038 /  9767 =  82.298
0, 1  acc:  6491 /  7535 =  86.145
1, 0  acc:  2441 /  2480 =  98.427
1, 1  acc:   165 /   180 =  91.667
------------------------------------
Average acc: 17135 / 19962 =  85.838
Robust  acc:  8038 /  9767 =  82.298
------------------------------------
Epoch:   2 | Train Loss: 0.002 | Train Acc: 92.860 | Val Loss: 0.002 | Val Acc: 87.517
Training:
Accuracies by groups:
0, 0  acc: 16389 / 23266 =  70.442
0, 1  acc:  8320 / 10206 =  81.521
1, 0  acc: 119013 / 121006 =  98.353
1, 1  acc:  7426 /  8292 =  89.556
--------------------------------------
Average acc: 151148 / 162770 =  92.860
Robust  acc: 16389 / 23266 =  70.442
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6830 /  8535 =  80.023
0, 1  acc:  7557 /  8276 =  91.312
1, 0  acc:  2830 /  2874 =  98.469
1, 1  acc:   170 /   182 =  93.407
------------------------------------
Average acc: 17387 / 19867 =  87.517
Robust  acc:  6830 /  8535 =  80.023
------------------------------------
New max robust acc: 80.02343292325718
debias model - Saving best checkpoint at epoch 1
replace: True
-> Updating checkpoint debias-wga-best_seed31.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed31.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 88.739
Robust Acc: 84.581 | Best Acc: 98.306
-------------------------------------
Training, Epoch 1:
Accuracies by groups:
0, 0  acc:  8261 /  9767 =  84.581
0, 1  acc:  6853 /  7535 =  90.949
1, 0  acc:  2438 /  2480 =  98.306
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 17714 / 19962 =  88.739
Robust  acc:  8261 /  9767 =  84.581
------------------------------------
Accuracies by groups:
0, 0  acc:  8261 /  9767 =  84.581
0, 1  acc:  6853 /  7535 =  90.949
1, 0  acc:  2438 /  2480 =  98.306
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 17714 / 19962 =  88.739
Robust  acc:  8261 /  9767 =  84.581
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8261 /  9767 =  84.581
0, 1  acc:  6853 /  7535 =  90.949
1, 0  acc:  2438 /  2480 =  98.306
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 17714 / 19962 =  88.739
Robust  acc:  8261 /  9767 =  84.581
------------------------------------
Epoch:   3 | Train Loss: 0.001 | Train Acc: 93.695 | Val Loss: 0.002 | Val Acc: 89.858
Training:
Accuracies by groups:
0, 0  acc: 17203 / 23436 =  73.404
0, 1  acc:  8804 / 10272 =  85.709
1, 0  acc: 118971 / 120815 =  98.474
1, 1  acc:  7529 /  8247 =  91.294
--------------------------------------
Average acc: 152507 / 162770 =  93.695
Robust  acc: 17203 / 23436 =  73.404
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7143 /  8535 =  83.691
0, 1  acc:  7729 /  8276 =  93.391
1, 0  acc:  2816 /  2874 =  97.982
1, 1  acc:   164 /   182 =  90.110
------------------------------------
Average acc: 17852 / 19867 =  89.858
Robust  acc:  7143 /  8535 =  83.691
------------------------------------
New max robust acc: 83.69068541300527
debias model - Saving best checkpoint at epoch 2
replace: True
-> Updating checkpoint debias-wga-best_seed31.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed31.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 90.562
Robust Acc: 86.966 | Best Acc: 97.339
-------------------------------------
Training, Epoch 2:
Accuracies by groups:
0, 0  acc:  8494 /  9767 =  86.966
0, 1  acc:  7013 /  7535 =  93.072
1, 0  acc:  2414 /  2480 =  97.339
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 18078 / 19962 =  90.562
Robust  acc:  8494 /  9767 =  86.966
------------------------------------
Accuracies by groups:
0, 0  acc:  8494 /  9767 =  86.966
0, 1  acc:  7013 /  7535 =  93.072
1, 0  acc:  2414 /  2480 =  97.339
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 18078 / 19962 =  90.562
Robust  acc:  8494 /  9767 =  86.966
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8494 /  9767 =  86.966
0, 1  acc:  7013 /  7535 =  93.072
1, 0  acc:  2414 /  2480 =  97.339
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 18078 / 19962 =  90.562
Robust  acc:  8494 /  9767 =  86.966
------------------------------------
Epoch:   4 | Train Loss: 0.001 | Train Acc: 94.326 | Val Loss: 0.002 | Val Acc: 90.240
Training:
Accuracies by groups:
0, 0  acc: 17881 / 23657 =  75.584
0, 1  acc:  9042 / 10262 =  88.111
1, 0  acc: 119049 / 120715 =  98.620
1, 1  acc:  7562 /  8136 =  92.945
--------------------------------------
Average acc: 153534 / 162770 =  94.326
Robust  acc: 17881 / 23657 =  75.584
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7209 /  8535 =  84.464
0, 1  acc:  7751 /  8276 =  93.656
1, 0  acc:  2802 /  2874 =  97.495
1, 1  acc:   166 /   182 =  91.209
------------------------------------
Average acc: 17928 / 19867 =  90.240
Robust  acc:  7209 /  8535 =  84.464
------------------------------------
New max robust acc: 84.4639718804921
debias model - Saving best checkpoint at epoch 3
replace: True
-> Updating checkpoint debias-wga-best_seed31.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed31.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 90.823
Robust Acc: 86.667 | Best Acc: 96.935
-------------------------------------
Training, Epoch 3:
Accuracies by groups:
0, 0  acc:  8538 /  9767 =  87.417
0, 1  acc:  7032 /  7535 =  93.324
1, 0  acc:  2404 /  2480 =  96.935
1, 1  acc:   156 /   180 =  86.667
------------------------------------
Average acc: 18130 / 19962 =  90.823
Robust  acc:   156 /   180 =  86.667
------------------------------------
Accuracies by groups:
0, 0  acc:  8538 /  9767 =  87.417
0, 1  acc:  7032 /  7535 =  93.324
1, 0  acc:  2404 /  2480 =  96.935
1, 1  acc:   156 /   180 =  86.667
------------------------------------
Average acc: 18130 / 19962 =  90.823
Robust  acc:   156 /   180 =  86.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8538 /  9767 =  87.417
0, 1  acc:  7032 /  7535 =  93.324
1, 0  acc:  2404 /  2480 =  96.935
1, 1  acc:   156 /   180 =  86.667
------------------------------------
Average acc: 18130 / 19962 =  90.823
Robust  acc:   156 /   180 =  86.667
------------------------------------
Epoch:   5 | Train Loss: 0.001 | Train Acc: 95.188 | Val Loss: 0.002 | Val Acc: 91.564
Training:
Accuracies by groups:
0, 0  acc: 18191 / 23423 =  77.663
0, 1  acc:  9170 / 10212 =  89.796
1, 0  acc: 119704 / 120911 =  99.002
1, 1  acc:  7872 /  8224 =  95.720
--------------------------------------
Average acc: 154937 / 162770 =  95.188
Robust  acc: 18191 / 23423 =  77.663
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7366 /  8535 =  86.303
0, 1  acc:  7889 /  8276 =  95.324
1, 0  acc:  2777 /  2874 =  96.625
1, 1  acc:   159 /   182 =  87.363
------------------------------------
Average acc: 18191 / 19867 =  91.564
Robust  acc:  7366 /  8535 =  86.303
------------------------------------
New max robust acc: 86.30345635618043
debias model - Saving best checkpoint at epoch 4
replace: True
-> Updating checkpoint debias-wga-best_seed31.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed31.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.729
Robust Acc: 79.444 | Best Acc: 96.855
-------------------------------------
Training, Epoch 4:
Accuracies by groups:
0, 0  acc:  8652 /  9767 =  88.584
0, 1  acc:  7114 /  7535 =  94.413
1, 0  acc:  2402 /  2480 =  96.855
1, 1  acc:   143 /   180 =  79.444
------------------------------------
Average acc: 18311 / 19962 =  91.729
Robust  acc:   143 /   180 =  79.444
------------------------------------
Accuracies by groups:
0, 0  acc:  8652 /  9767 =  88.584
0, 1  acc:  7114 /  7535 =  94.413
1, 0  acc:  2402 /  2480 =  96.855
1, 1  acc:   143 /   180 =  79.444
------------------------------------
Average acc: 18311 / 19962 =  91.729
Robust  acc:   143 /   180 =  79.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8652 /  9767 =  88.584
0, 1  acc:  7114 /  7535 =  94.413
1, 0  acc:  2402 /  2480 =  96.855
1, 1  acc:   143 /   180 =  79.444
------------------------------------
Average acc: 18311 / 19962 =  91.729
Robust  acc:   143 /   180 =  79.444
------------------------------------
Epoch:   6 | Train Loss: 0.001 | Train Acc: 96.194 | Val Loss: 0.001 | Val Acc: 92.455
Training:
Accuracies by groups:
0, 0  acc: 19128 / 23593 =  81.075
0, 1  acc:  9545 / 10359 =  92.142
1, 0  acc: 119918 / 120633 =  99.407
1, 1  acc:  7984 /  8185 =  97.544
--------------------------------------
Average acc: 156575 / 162770 =  96.194
Robust  acc: 19128 / 23593 =  81.075
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7507 /  8535 =  87.955
0, 1  acc:  7952 /  8276 =  96.085
1, 0  acc:  2757 /  2874 =  95.929
1, 1  acc:   152 /   182 =  83.516
------------------------------------
Average acc: 18368 / 19867 =  92.455
Robust  acc:   152 /   182 =  83.516
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 92.761
Robust Acc: 77.222 | Best Acc: 95.766
-------------------------------------
Training, Epoch 5:
Accuracies by groups:
0, 0  acc:  8806 /  9767 =  90.161
0, 1  acc:  7197 /  7535 =  95.514
1, 0  acc:  2375 /  2480 =  95.766
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 18517 / 19962 =  92.761
Robust  acc:   139 /   180 =  77.222
------------------------------------
Accuracies by groups:
0, 0  acc:  8806 /  9767 =  90.161
0, 1  acc:  7197 /  7535 =  95.514
1, 0  acc:  2375 /  2480 =  95.766
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 18517 / 19962 =  92.761
Robust  acc:   139 /   180 =  77.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8806 /  9767 =  90.161
0, 1  acc:  7197 /  7535 =  95.514
1, 0  acc:  2375 /  2480 =  95.766
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 18517 / 19962 =  92.761
Robust  acc:   139 /   180 =  77.222
------------------------------------
Epoch:   7 | Train Loss: 0.001 | Train Acc: 97.199 | Val Loss: 0.001 | Val Acc: 92.752
Training:
Accuracies by groups:
0, 0  acc: 19932 / 23516 =  84.759
0, 1  acc:  9470 / 10093 =  93.827
1, 0  acc: 120750 / 121026 =  99.772
1, 1  acc:  8059 /  8135 =  99.066
--------------------------------------
Average acc: 158211 / 162770 =  97.199
Robust  acc: 19932 / 23516 =  84.759
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7542 /  8535 =  88.366
0, 1  acc:  8013 /  8276 =  96.822
1, 0  acc:  2726 /  2874 =  94.850
1, 1  acc:   146 /   182 =  80.220
------------------------------------
Average acc: 18427 / 19867 =  92.752
Robust  acc:   146 /   182 =  80.220
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 93.097
Robust Acc: 69.444 | Best Acc: 96.417
-------------------------------------
Training, Epoch 6:
Accuracies by groups:
0, 0  acc:  8847 /  9767 =  90.581
0, 1  acc:  7265 /  7535 =  96.417
1, 0  acc:  2347 /  2480 =  94.637
1, 1  acc:   125 /   180 =  69.444
------------------------------------
Average acc: 18584 / 19962 =  93.097
Robust  acc:   125 /   180 =  69.444
------------------------------------
Accuracies by groups:
0, 0  acc:  8847 /  9767 =  90.581
0, 1  acc:  7265 /  7535 =  96.417
1, 0  acc:  2347 /  2480 =  94.637
1, 1  acc:   125 /   180 =  69.444
------------------------------------
Average acc: 18584 / 19962 =  93.097
Robust  acc:   125 /   180 =  69.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8847 /  9767 =  90.581
0, 1  acc:  7265 /  7535 =  96.417
1, 0  acc:  2347 /  2480 =  94.637
1, 1  acc:   125 /   180 =  69.444
------------------------------------
Average acc: 18584 / 19962 =  93.097
Robust  acc:   125 /   180 =  69.444
------------------------------------
Epoch:   8 | Train Loss: 0.001 | Train Acc: 98.072 | Val Loss: 0.001 | Val Acc: 94.388
Training:
Accuracies by groups:
0, 0  acc: 21017 / 23552 =  89.237
0, 1  acc:  9793 / 10241 =  95.625
1, 0  acc: 120522 / 120652 =  99.892
1, 1  acc:  8300 /  8325 =  99.700
--------------------------------------
Average acc: 159632 / 162770 =  98.072
Robust  acc: 21017 / 23552 =  89.237
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7913 /  8535 =  92.712
0, 1  acc:  8115 /  8276 =  98.055
1, 0  acc:  2603 /  2874 =  90.571
1, 1  acc:   121 /   182 =  66.484
------------------------------------
Average acc: 18752 / 19867 =  94.388
Robust  acc:   121 /   182 =  66.484
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.530
Robust Acc: 62.222 | Best Acc: 97.757
-------------------------------------
Training, Epoch 7:
Accuracies by groups:
0, 0  acc:  9175 /  9767 =  93.939
0, 1  acc:  7366 /  7535 =  97.757
1, 0  acc:  2217 /  2480 =  89.395
1, 1  acc:   112 /   180 =  62.222
------------------------------------
Average acc: 18870 / 19962 =  94.530
Robust  acc:   112 /   180 =  62.222
------------------------------------
Accuracies by groups:
0, 0  acc:  9175 /  9767 =  93.939
0, 1  acc:  7366 /  7535 =  97.757
1, 0  acc:  2217 /  2480 =  89.395
1, 1  acc:   112 /   180 =  62.222
------------------------------------
Average acc: 18870 / 19962 =  94.530
Robust  acc:   112 /   180 =  62.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9175 /  9767 =  93.939
0, 1  acc:  7366 /  7535 =  97.757
1, 0  acc:  2217 /  2480 =  89.395
1, 1  acc:   112 /   180 =  62.222
------------------------------------
Average acc: 18870 / 19962 =  94.530
Robust  acc:   112 /   180 =  62.222
------------------------------------
Epoch:   9 | Train Loss: 0.001 | Train Acc: 98.628 | Val Loss: 0.001 | Val Acc: 93.854
Training:
Accuracies by groups:
0, 0  acc: 21584 / 23390 =  92.279
0, 1  acc:  9826 / 10135 =  96.951
1, 0  acc: 120939 / 121035 =  99.921
1, 1  acc:  8187 /  8210 =  99.720
--------------------------------------
Average acc: 160536 / 162770 =  98.628
Robust  acc: 21584 / 23390 =  92.279
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7792 /  8535 =  91.295
0, 1  acc:  8098 /  8276 =  97.849
1, 0  acc:  2633 /  2874 =  91.614
1, 1  acc:   123 /   182 =  67.582
------------------------------------
Average acc: 18646 / 19867 =  93.854
Robust  acc:   123 /   182 =  67.582
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.029
Robust Acc: 62.222 | Best Acc: 97.279
-------------------------------------
Training, Epoch 8:
Accuracies by groups:
0, 0  acc:  9072 /  9767 =  92.884
0, 1  acc:  7330 /  7535 =  97.279
1, 0  acc:  2256 /  2480 =  90.968
1, 1  acc:   112 /   180 =  62.222
------------------------------------
Average acc: 18770 / 19962 =  94.029
Robust  acc:   112 /   180 =  62.222
------------------------------------
Accuracies by groups:
0, 0  acc:  9072 /  9767 =  92.884
0, 1  acc:  7330 /  7535 =  97.279
1, 0  acc:  2256 /  2480 =  90.968
1, 1  acc:   112 /   180 =  62.222
------------------------------------
Average acc: 18770 / 19962 =  94.029
Robust  acc:   112 /   180 =  62.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9072 /  9767 =  92.884
0, 1  acc:  7330 /  7535 =  97.279
1, 0  acc:  2256 /  2480 =  90.968
1, 1  acc:   112 /   180 =  62.222
------------------------------------
Average acc: 18770 / 19962 =  94.029
Robust  acc:   112 /   180 =  62.222
------------------------------------
Epoch:  10 | Train Loss: 0.001 | Train Acc: 98.842 | Val Loss: 0.001 | Val Acc: 94.720
Training:
Accuracies by groups:
0, 0  acc: 22050 / 23506 =  93.806
0, 1  acc: 10042 / 10307 =  97.429
1, 0  acc: 120684 / 120815 =  99.892
1, 1  acc:  8109 /  8142 =  99.595
--------------------------------------
Average acc: 160885 / 162770 =  98.842
Robust  acc: 22050 / 23506 =  93.806
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8025 /  8535 =  94.025
0, 1  acc:  8182 /  8276 =  98.864
1, 0  acc:  2507 /  2874 =  87.230
1, 1  acc:   104 /   182 =  57.143
------------------------------------
Average acc: 18818 / 19867 =  94.720
Robust  acc:   104 /   182 =  57.143
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.810
Robust Acc: 53.889 | Best Acc: 98.500
-------------------------------------
Training, Epoch 9:
Accuracies by groups:
0, 0  acc:  9281 /  9767 =  95.024
0, 1  acc:  7422 /  7535 =  98.500
1, 0  acc:  2126 /  2480 =  85.726
1, 1  acc:    97 /   180 =  53.889
------------------------------------
Average acc: 18926 / 19962 =  94.810
Robust  acc:    97 /   180 =  53.889
------------------------------------
Accuracies by groups:
0, 0  acc:  9281 /  9767 =  95.024
0, 1  acc:  7422 /  7535 =  98.500
1, 0  acc:  2126 /  2480 =  85.726
1, 1  acc:    97 /   180 =  53.889
------------------------------------
Average acc: 18926 / 19962 =  94.810
Robust  acc:    97 /   180 =  53.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9281 /  9767 =  95.024
0, 1  acc:  7422 /  7535 =  98.500
1, 0  acc:  2126 /  2480 =  85.726
1, 1  acc:    97 /   180 =  53.889
------------------------------------
Average acc: 18926 / 19962 =  94.810
Robust  acc:    97 /   180 =  53.889
------------------------------------
Epoch:  11 | Train Loss: 0.001 | Train Acc: 98.790 | Val Loss: 0.001 | Val Acc: 92.837
Training:
Accuracies by groups:
0, 0  acc: 21724 / 23205 =  93.618
0, 1  acc:  9882 / 10106 =  97.783
1, 0  acc: 120858 / 121084 =  99.813
1, 1  acc:  8336 /  8375 =  99.534
--------------------------------------
Average acc: 160800 / 162770 =  98.790
Robust  acc: 21724 / 23205 =  93.618
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7611 /  8535 =  89.174
0, 1  acc:  8014 /  8276 =  96.834
1, 0  acc:  2687 /  2874 =  93.493
1, 1  acc:   132 /   182 =  72.527
------------------------------------
Average acc: 18444 / 19867 =  92.837
Robust  acc:   132 /   182 =  72.527
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 93.107
Robust Acc: 71.667 | Best Acc: 96.138
-------------------------------------
Training, Epoch 10:
Accuracies by groups:
0, 0  acc:  8923 /  9767 =  91.359
0, 1  acc:  7244 /  7535 =  96.138
1, 0  acc:  2290 /  2480 =  92.339
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18586 / 19962 =  93.107
Robust  acc:   129 /   180 =  71.667
------------------------------------
Accuracies by groups:
0, 0  acc:  8923 /  9767 =  91.359
0, 1  acc:  7244 /  7535 =  96.138
1, 0  acc:  2290 /  2480 =  92.339
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18586 / 19962 =  93.107
Robust  acc:   129 /   180 =  71.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8923 /  9767 =  91.359
0, 1  acc:  7244 /  7535 =  96.138
1, 0  acc:  2290 /  2480 =  92.339
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18586 / 19962 =  93.107
Robust  acc:   129 /   180 =  71.667
------------------------------------
Epoch:  12 | Train Loss: 0.001 | Train Acc: 98.403 | Val Loss: 0.001 | Val Acc: 94.267
Training:
Accuracies by groups:
0, 0  acc: 21918 / 23693 =  92.508
0, 1  acc:  9775 / 10076 =  97.013
1, 0  acc: 120375 / 120783 =  99.662
1, 1  acc:  8102 /  8218 =  98.588
--------------------------------------
Average acc: 160170 / 162770 =  98.403
Robust  acc: 21918 / 23693 =  92.508
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7880 /  8535 =  92.326
0, 1  acc:  8127 /  8276 =  98.200
1, 0  acc:  2597 /  2874 =  90.362
1, 1  acc:   124 /   182 =  68.132
------------------------------------
Average acc: 18728 / 19867 =  94.267
Robust  acc:   124 /   182 =  68.132
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.490
Robust Acc: 62.222 | Best Acc: 97.797
-------------------------------------
Training, Epoch 11:
Accuracies by groups:
0, 0  acc:  9173 /  9767 =  93.918
0, 1  acc:  7369 /  7535 =  97.797
1, 0  acc:  2208 /  2480 =  89.032
1, 1  acc:   112 /   180 =  62.222
------------------------------------
Average acc: 18862 / 19962 =  94.490
Robust  acc:   112 /   180 =  62.222
------------------------------------
Accuracies by groups:
0, 0  acc:  9173 /  9767 =  93.918
0, 1  acc:  7369 /  7535 =  97.797
1, 0  acc:  2208 /  2480 =  89.032
1, 1  acc:   112 /   180 =  62.222
------------------------------------
Average acc: 18862 / 19962 =  94.490
Robust  acc:   112 /   180 =  62.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9173 /  9767 =  93.918
0, 1  acc:  7369 /  7535 =  97.797
1, 0  acc:  2208 /  2480 =  89.032
1, 1  acc:   112 /   180 =  62.222
------------------------------------
Average acc: 18862 / 19962 =  94.490
Robust  acc:   112 /   180 =  62.222
------------------------------------
Epoch:  13 | Train Loss: 0.001 | Train Acc: 97.939 | Val Loss: 0.002 | Val Acc: 92.097
Training:
Accuracies by groups:
0, 0  acc: 21497 / 23699 =  90.708
0, 1  acc:  9735 / 10110 =  96.291
1, 0  acc: 120079 / 120712 =  99.476
1, 1  acc:  8104 /  8249 =  98.242
--------------------------------------
Average acc: 159415 / 162770 =  97.939
Robust  acc: 21497 / 23699 =  90.708
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7434 /  8535 =  87.100
0, 1  acc:  8027 /  8276 =  96.991
1, 0  acc:  2702 /  2874 =  94.015
1, 1  acc:   134 /   182 =  73.626
------------------------------------
Average acc: 18297 / 19867 =  92.097
Robust  acc:   134 /   182 =  73.626
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.205
Robust Acc: 71.667 | Best Acc: 95.859
-------------------------------------
Training, Epoch 12:
Accuracies by groups:
0, 0  acc:  8745 /  9767 =  89.536
0, 1  acc:  7223 /  7535 =  95.859
1, 0  acc:  2309 /  2480 =  93.105
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18406 / 19962 =  92.205
Robust  acc:   129 /   180 =  71.667
------------------------------------
Accuracies by groups:
0, 0  acc:  8745 /  9767 =  89.536
0, 1  acc:  7223 /  7535 =  95.859
1, 0  acc:  2309 /  2480 =  93.105
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18406 / 19962 =  92.205
Robust  acc:   129 /   180 =  71.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8745 /  9767 =  89.536
0, 1  acc:  7223 /  7535 =  95.859
1, 0  acc:  2309 /  2480 =  93.105
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18406 / 19962 =  92.205
Robust  acc:   129 /   180 =  71.667
------------------------------------
Epoch:  14 | Train Loss: 0.001 | Train Acc: 97.272 | Val Loss: 0.002 | Val Acc: 88.398
Training:
Accuracies by groups:
0, 0  acc: 20520 / 23279 =  88.148
0, 1  acc:  9691 / 10177 =  95.225
1, 0  acc: 120149 / 121099 =  99.216
1, 1  acc:  7970 /  8215 =  97.018
--------------------------------------
Average acc: 158330 / 162770 =  97.272
Robust  acc: 20520 / 23279 =  88.148
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7181 /  8535 =  84.136
0, 1  acc:  7441 /  8276 =  89.911
1, 0  acc:  2776 /  2874 =  96.590
1, 1  acc:   164 /   182 =  90.110
------------------------------------
Average acc: 17562 / 19867 =  88.398
Robust  acc:  7181 /  8535 =  84.136
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 89.295
Robust Acc: 87.079 | Best Acc: 95.403
-------------------------------------
Training, Epoch 13:
Accuracies by groups:
0, 0  acc:  8505 /  9767 =  87.079
0, 1  acc:  6796 /  7535 =  90.192
1, 0  acc:  2366 /  2480 =  95.403
1, 1  acc:   158 /   180 =  87.778
------------------------------------
Average acc: 17825 / 19962 =  89.295
Robust  acc:  8505 /  9767 =  87.079
------------------------------------
Accuracies by groups:
0, 0  acc:  8505 /  9767 =  87.079
0, 1  acc:  6796 /  7535 =  90.192
1, 0  acc:  2366 /  2480 =  95.403
1, 1  acc:   158 /   180 =  87.778
------------------------------------
Average acc: 17825 / 19962 =  89.295
Robust  acc:  8505 /  9767 =  87.079
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8505 /  9767 =  87.079
0, 1  acc:  6796 /  7535 =  90.192
1, 0  acc:  2366 /  2480 =  95.403
1, 1  acc:   158 /   180 =  87.778
------------------------------------
Average acc: 17825 / 19962 =  89.295
Robust  acc:  8505 /  9767 =  87.079
------------------------------------
Epoch:  15 | Train Loss: 0.001 | Train Acc: 96.739 | Val Loss: 0.002 | Val Acc: 93.240
Training:
Accuracies by groups:
0, 0  acc: 20130 / 23474 =  85.754
0, 1  acc:  9593 / 10181 =  94.225
1, 0  acc: 119715 / 120793 =  99.108
1, 1  acc:  8024 /  8322 =  96.419
--------------------------------------
Average acc: 157462 / 162770 =  96.739
Robust  acc: 20130 / 23474 =  85.754
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7808 /  8535 =  91.482
0, 1  acc:  7947 /  8276 =  96.025
1, 0  acc:  2619 /  2874 =  91.127
1, 1  acc:   150 /   182 =  82.418
------------------------------------
Average acc: 18524 / 19867 =  93.240
Robust  acc:   150 /   182 =  82.418
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.433
Robust Acc: 71.111 | Best Acc: 95.514
-------------------------------------
Training, Epoch 14:
Accuracies by groups:
0, 0  acc:  9071 /  9767 =  92.874
0, 1  acc:  7197 /  7535 =  95.514
1, 0  acc:  2255 /  2480 =  90.927
1, 1  acc:   128 /   180 =  71.111
------------------------------------
Average acc: 18651 / 19962 =  93.433
Robust  acc:   128 /   180 =  71.111
------------------------------------
Accuracies by groups:
0, 0  acc:  9071 /  9767 =  92.874
0, 1  acc:  7197 /  7535 =  95.514
1, 0  acc:  2255 /  2480 =  90.927
1, 1  acc:   128 /   180 =  71.111
------------------------------------
Average acc: 18651 / 19962 =  93.433
Robust  acc:   128 /   180 =  71.111
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9071 /  9767 =  92.874
0, 1  acc:  7197 /  7535 =  95.514
1, 0  acc:  2255 /  2480 =  90.927
1, 1  acc:   128 /   180 =  71.111
------------------------------------
Average acc: 18651 / 19962 =  93.433
Robust  acc:   128 /   180 =  71.111
------------------------------------
Epoch:  16 | Train Loss: 0.001 | Train Acc: 96.109 | Val Loss: 0.001 | Val Acc: 93.774
Training:
Accuracies by groups:
0, 0  acc: 19498 / 23456 =  83.126
0, 1  acc:  9543 / 10274 =  92.885
1, 0  acc: 119464 / 120763 =  98.924
1, 1  acc:  7932 /  8277 =  95.832
--------------------------------------
Average acc: 156437 / 162770 =  96.109
Robust  acc: 19498 / 23456 =  83.126
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7892 /  8535 =  92.466
0, 1  acc:  7995 /  8276 =  96.605
1, 0  acc:  2597 /  2874 =  90.362
1, 1  acc:   146 /   182 =  80.220
------------------------------------
Average acc: 18630 / 19867 =  93.774
Robust  acc:   146 /   182 =  80.220
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.309
Robust Acc: 74.444 | Best Acc: 96.496
-------------------------------------
Training, Epoch 15:
Accuracies by groups:
0, 0  acc:  9187 /  9767 =  94.062
0, 1  acc:  7271 /  7535 =  96.496
1, 0  acc:  2234 /  2480 =  90.081
1, 1  acc:   134 /   180 =  74.444
------------------------------------
Average acc: 18826 / 19962 =  94.309
Robust  acc:   134 /   180 =  74.444
------------------------------------
Accuracies by groups:
0, 0  acc:  9187 /  9767 =  94.062
0, 1  acc:  7271 /  7535 =  96.496
1, 0  acc:  2234 /  2480 =  90.081
1, 1  acc:   134 /   180 =  74.444
------------------------------------
Average acc: 18826 / 19962 =  94.309
Robust  acc:   134 /   180 =  74.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9187 /  9767 =  94.062
0, 1  acc:  7271 /  7535 =  96.496
1, 0  acc:  2234 /  2480 =  90.081
1, 1  acc:   134 /   180 =  74.444
------------------------------------
Average acc: 18826 / 19962 =  94.309
Robust  acc:   134 /   180 =  74.444
------------------------------------
Epoch:  17 | Train Loss: 0.001 | Train Acc: 95.465 | Val Loss: 0.004 | Val Acc: 76.584
Training:
Accuracies by groups:
0, 0  acc: 18954 / 23418 =  80.938
0, 1  acc:  9279 / 10182 =  91.131
1, 0  acc: 119443 / 121038 =  98.682
1, 1  acc:  7712 /  8132 =  94.835
--------------------------------------
Average acc: 155388 / 162770 =  95.465
Robust  acc: 18954 / 23418 =  80.938
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5886 /  8535 =  68.963
0, 1  acc:  6296 /  8276 =  76.075
1, 0  acc:  2855 /  2874 =  99.339
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 15215 / 19867 =  76.584
Robust  acc:  5886 /  8535 =  68.963
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 78.023
Robust Acc: 73.707 | Best Acc: 99.073
-------------------------------------
Training, Epoch 16:
Accuracies by groups:
0, 0  acc:  7199 /  9767 =  73.707
0, 1  acc:  5747 /  7535 =  76.271
1, 0  acc:  2457 /  2480 =  99.073
1, 1  acc:   172 /   180 =  95.556
------------------------------------
Average acc: 15575 / 19962 =  78.023
Robust  acc:  7199 /  9767 =  73.707
------------------------------------
Accuracies by groups:
0, 0  acc:  7199 /  9767 =  73.707
0, 1  acc:  5747 /  7535 =  76.271
1, 0  acc:  2457 /  2480 =  99.073
1, 1  acc:   172 /   180 =  95.556
------------------------------------
Average acc: 15575 / 19962 =  78.023
Robust  acc:  7199 /  9767 =  73.707
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7199 /  9767 =  73.707
0, 1  acc:  5747 /  7535 =  76.271
1, 0  acc:  2457 /  2480 =  99.073
1, 1  acc:   172 /   180 =  95.556
------------------------------------
Average acc: 15575 / 19962 =  78.023
Robust  acc:  7199 /  9767 =  73.707
------------------------------------
Epoch:  18 | Train Loss: 0.001 | Train Acc: 94.819 | Val Loss: 0.002 | Val Acc: 93.879
Training:
Accuracies by groups:
0, 0  acc: 18535 / 23642 =  78.399
0, 1  acc:  9037 / 10078 =  89.671
1, 0  acc: 119018 / 120804 =  98.522
1, 1  acc:  7747 /  8246 =  93.949
--------------------------------------
Average acc: 154337 / 162770 =  94.819
Robust  acc: 18535 / 23642 =  78.399
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7940 /  8535 =  93.029
0, 1  acc:  8037 /  8276 =  97.112
1, 0  acc:  2543 /  2874 =  88.483
1, 1  acc:   131 /   182 =  71.978
------------------------------------
Average acc: 18651 / 19867 =  93.879
Robust  acc:   131 /   182 =  71.978
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 94.610
Robust Acc: 66.667 | Best Acc: 97.213
-------------------------------------
Training, Epoch 17:
Accuracies by groups:
0, 0  acc:  9230 /  9767 =  94.502
0, 1  acc:  7325 /  7535 =  97.213
1, 0  acc:  2211 /  2480 =  89.153
1, 1  acc:   120 /   180 =  66.667
------------------------------------
Average acc: 18886 / 19962 =  94.610
Robust  acc:   120 /   180 =  66.667
------------------------------------
Accuracies by groups:
0, 0  acc:  9230 /  9767 =  94.502
0, 1  acc:  7325 /  7535 =  97.213
1, 0  acc:  2211 /  2480 =  89.153
1, 1  acc:   120 /   180 =  66.667
------------------------------------
Average acc: 18886 / 19962 =  94.610
Robust  acc:   120 /   180 =  66.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9230 /  9767 =  94.502
0, 1  acc:  7325 /  7535 =  97.213
1, 0  acc:  2211 /  2480 =  89.153
1, 1  acc:   120 /   180 =  66.667
------------------------------------
Average acc: 18886 / 19962 =  94.610
Robust  acc:   120 /   180 =  66.667
------------------------------------
Epoch:  19 | Train Loss: 0.002 | Train Acc: 94.337 | Val Loss: 0.002 | Val Acc: 93.210
Training:
Accuracies by groups:
0, 0  acc: 17823 / 23235 =  76.708
0, 1  acc:  9012 / 10213 =  88.240
1, 0  acc: 119164 / 121174 =  98.341
1, 1  acc:  7553 /  8148 =  92.698
--------------------------------------
Average acc: 153552 / 162770 =  94.337
Robust  acc: 17823 / 23235 =  76.708
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7678 /  8535 =  89.959
0, 1  acc:  7982 /  8276 =  96.448
1, 0  acc:  2711 /  2874 =  94.328
1, 1  acc:   147 /   182 =  80.769
------------------------------------
Average acc: 18518 / 19867 =  93.210
Robust  acc:   147 /   182 =  80.769
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.438
Robust Acc: 70.556 | Best Acc: 96.297
-------------------------------------
Training, Epoch 18:
Accuracies by groups:
0, 0  acc:  8946 /  9767 =  91.594
0, 1  acc:  7256 /  7535 =  96.297
1, 0  acc:  2323 /  2480 =  93.669
1, 1  acc:   127 /   180 =  70.556
------------------------------------
Average acc: 18652 / 19962 =  93.438
Robust  acc:   127 /   180 =  70.556
------------------------------------
Accuracies by groups:
0, 0  acc:  8946 /  9767 =  91.594
0, 1  acc:  7256 /  7535 =  96.297
1, 0  acc:  2323 /  2480 =  93.669
1, 1  acc:   127 /   180 =  70.556
------------------------------------
Average acc: 18652 / 19962 =  93.438
Robust  acc:   127 /   180 =  70.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8946 /  9767 =  91.594
0, 1  acc:  7256 /  7535 =  96.297
1, 0  acc:  2323 /  2480 =  93.669
1, 1  acc:   127 /   180 =  70.556
------------------------------------
Average acc: 18652 / 19962 =  93.438
Robust  acc:   127 /   180 =  70.556
------------------------------------
Epoch:  20 | Train Loss: 0.002 | Train Acc: 93.863 | Val Loss: 0.002 | Val Acc: 93.839
Training:
Accuracies by groups:
0, 0  acc: 17710 / 23563 =  75.160
0, 1  acc:  8981 / 10247 =  87.645
1, 0  acc: 118585 / 120840 =  98.134
1, 1  acc:  7505 /  8120 =  92.426
--------------------------------------
Average acc: 152781 / 162770 =  93.863
Robust  acc: 17710 / 23563 =  75.160
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7835 /  8535 =  91.798
0, 1  acc:  8020 /  8276 =  96.907
1, 0  acc:  2645 /  2874 =  92.032
1, 1  acc:   143 /   182 =  78.571
------------------------------------
Average acc: 18643 / 19867 =  93.839
Robust  acc:   143 /   182 =  78.571
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 94.114
Robust Acc: 68.333 | Best Acc: 96.841
-------------------------------------
Training, Epoch 19:
Accuracies by groups:
0, 0  acc:  9089 /  9767 =  93.058
0, 1  acc:  7297 /  7535 =  96.841
1, 0  acc:  2278 /  2480 =  91.855
1, 1  acc:   123 /   180 =  68.333
------------------------------------
Average acc: 18787 / 19962 =  94.114
Robust  acc:   123 /   180 =  68.333
------------------------------------
Accuracies by groups:
0, 0  acc:  9089 /  9767 =  93.058
0, 1  acc:  7297 /  7535 =  96.841
1, 0  acc:  2278 /  2480 =  91.855
1, 1  acc:   123 /   180 =  68.333
------------------------------------
Average acc: 18787 / 19962 =  94.114
Robust  acc:   123 /   180 =  68.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9089 /  9767 =  93.058
0, 1  acc:  7297 /  7535 =  96.841
1, 0  acc:  2278 /  2480 =  91.855
1, 1  acc:   123 /   180 =  68.333
------------------------------------
Average acc: 18787 / 19962 =  94.114
Robust  acc:   123 /   180 =  68.333
------------------------------------
Epoch:  21 | Train Loss: 0.002 | Train Acc: 93.461 | Val Loss: 0.002 | Val Acc: 92.148
Training:
Accuracies by groups:
0, 0  acc: 17563 / 23708 =  74.080
0, 1  acc:  8760 / 10176 =  86.085
1, 0  acc: 118233 / 120629 =  98.014
1, 1  acc:  7570 /  8257 =  91.680
--------------------------------------
Average acc: 152126 / 162770 =  93.461
Robust  acc: 17563 / 23708 =  74.080
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7457 /  8535 =  87.370
0, 1  acc:  7933 /  8276 =  95.855
1, 0  acc:  2765 /  2874 =  96.207
1, 1  acc:   152 /   182 =  83.516
------------------------------------
Average acc: 18307 / 19867 =  92.148
Robust  acc:   152 /   182 =  83.516
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.821
Robust Acc: 80.556 | Best Acc: 95.700
-------------------------------------
Training, Epoch 20:
Accuracies by groups:
0, 0  acc:  8803 /  9767 =  90.130
0, 1  acc:  7211 /  7535 =  95.700
1, 0  acc:  2370 /  2480 =  95.565
1, 1  acc:   145 /   180 =  80.556
------------------------------------
Average acc: 18529 / 19962 =  92.821
Robust  acc:   145 /   180 =  80.556
------------------------------------
Accuracies by groups:
0, 0  acc:  8803 /  9767 =  90.130
0, 1  acc:  7211 /  7535 =  95.700
1, 0  acc:  2370 /  2480 =  95.565
1, 1  acc:   145 /   180 =  80.556
------------------------------------
Average acc: 18529 / 19962 =  92.821
Robust  acc:   145 /   180 =  80.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8803 /  9767 =  90.130
0, 1  acc:  7211 /  7535 =  95.700
1, 0  acc:  2370 /  2480 =  95.565
1, 1  acc:   145 /   180 =  80.556
------------------------------------
Average acc: 18529 / 19962 =  92.821
Robust  acc:   145 /   180 =  80.556
------------------------------------
Epoch:  22 | Train Loss: 0.002 | Train Acc: 93.224 | Val Loss: 0.004 | Val Acc: 78.698
Training:
Accuracies by groups:
0, 0  acc: 17279 / 23677 =  72.978
0, 1  acc:  8506 / 10094 =  84.268
1, 0  acc: 118520 / 120860 =  98.064
1, 1  acc:  7436 /  8139 =  91.363
--------------------------------------
Average acc: 151741 / 162770 =  93.224
Robust  acc: 17279 / 23677 =  72.978
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6043 /  8535 =  70.803
0, 1  acc:  6565 /  8276 =  79.326
1, 0  acc:  2849 /  2874 =  99.130
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 15635 / 19867 =  78.698
Robust  acc:  6043 /  8535 =  70.803
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 80.137
Robust Acc: 76.175 | Best Acc: 99.073
-------------------------------------
Training, Epoch 21:
Accuracies by groups:
0, 0  acc:  7440 /  9767 =  76.175
0, 1  acc:  5927 /  7535 =  78.660
1, 0  acc:  2457 /  2480 =  99.073
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15997 / 19962 =  80.137
Robust  acc:  7440 /  9767 =  76.175
------------------------------------
Accuracies by groups:
0, 0  acc:  7440 /  9767 =  76.175
0, 1  acc:  5927 /  7535 =  78.660
1, 0  acc:  2457 /  2480 =  99.073
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15997 / 19962 =  80.137
Robust  acc:  7440 /  9767 =  76.175
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7440 /  9767 =  76.175
0, 1  acc:  5927 /  7535 =  78.660
1, 0  acc:  2457 /  2480 =  99.073
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15997 / 19962 =  80.137
Robust  acc:  7440 /  9767 =  76.175
------------------------------------
Epoch:  23 | Train Loss: 0.002 | Train Acc: 93.107 | Val Loss: 0.004 | Val Acc: 77.853
Training:
Accuracies by groups:
0, 0  acc: 16629 / 23129 =  71.897
0, 1  acc:  8513 / 10214 =  83.346
1, 0  acc: 118903 / 121250 =  98.064
1, 1  acc:  7505 /  8177 =  91.782
--------------------------------------
Average acc: 151550 / 162770 =  93.107
Robust  acc: 16629 / 23129 =  71.897
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5632 /  8535 =  65.987
0, 1  acc:  6798 /  8276 =  82.141
1, 0  acc:  2861 /  2874 =  99.548
1, 1  acc:   176 /   182 =  96.703
------------------------------------
Average acc: 15467 / 19867 =  77.853
Robust  acc:  5632 /  8535 =  65.987
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 79.411
Robust Acc: 71.762 | Best Acc: 99.476
-------------------------------------
Training, Epoch 22:
Accuracies by groups:
0, 0  acc:  7009 /  9767 =  71.762
0, 1  acc:  6203 /  7535 =  82.322
1, 0  acc:  2467 /  2480 =  99.476
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15852 / 19962 =  79.411
Robust  acc:  7009 /  9767 =  71.762
------------------------------------
Accuracies by groups:
0, 0  acc:  7009 /  9767 =  71.762
0, 1  acc:  6203 /  7535 =  82.322
1, 0  acc:  2467 /  2480 =  99.476
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15852 / 19962 =  79.411
Robust  acc:  7009 /  9767 =  71.762
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7009 /  9767 =  71.762
0, 1  acc:  6203 /  7535 =  82.322
1, 0  acc:  2467 /  2480 =  99.476
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15852 / 19962 =  79.411
Robust  acc:  7009 /  9767 =  71.762
------------------------------------
Epoch:  24 | Train Loss: 0.002 | Train Acc: 92.768 | Val Loss: 0.003 | Val Acc: 90.567
Training:
Accuracies by groups:
0, 0  acc: 16647 / 23546 =  70.700
0, 1  acc:  8513 / 10213 =  83.355
1, 0  acc: 118593 / 121052 =  97.969
1, 1  acc:  7246 /  7959 =  91.042
--------------------------------------
Average acc: 150999 / 162770 =  92.768
Robust  acc: 16647 / 23546 =  70.700
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7433 /  8535 =  87.088
0, 1  acc:  7664 /  8276 =  92.605
1, 0  acc:  2735 /  2874 =  95.164
1, 1  acc:   161 /   182 =  88.462
------------------------------------
Average acc: 17993 / 19867 =  90.567
Robust  acc:  7433 /  8535 =  87.088
------------------------------------
New max robust acc: 87.08845928529584
debias model - Saving best checkpoint at epoch 23
replace: True
-> Updating checkpoint debias-wga-best_seed31.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed31.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.544
Robust Acc: 85.000 | Best Acc: 94.113
-------------------------------------
Training, Epoch 23:
Accuracies by groups:
0, 0  acc:  8784 /  9767 =  89.935
0, 1  acc:  7003 /  7535 =  92.940
1, 0  acc:  2334 /  2480 =  94.113
1, 1  acc:   153 /   180 =  85.000
------------------------------------
Average acc: 18274 / 19962 =  91.544
Robust  acc:   153 /   180 =  85.000
------------------------------------
Accuracies by groups:
0, 0  acc:  8784 /  9767 =  89.935
0, 1  acc:  7003 /  7535 =  92.940
1, 0  acc:  2334 /  2480 =  94.113
1, 1  acc:   153 /   180 =  85.000
------------------------------------
Average acc: 18274 / 19962 =  91.544
Robust  acc:   153 /   180 =  85.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8784 /  9767 =  89.935
0, 1  acc:  7003 /  7535 =  92.940
1, 0  acc:  2334 /  2480 =  94.113
1, 1  acc:   153 /   180 =  85.000
------------------------------------
Average acc: 18274 / 19962 =  91.544
Robust  acc:   153 /   180 =  85.000
------------------------------------
Epoch:  25 | Train Loss: 0.002 | Train Acc: 92.759 | Val Loss: 0.003 | Val Acc: 83.913
Training:
Accuracies by groups:
0, 0  acc: 16573 / 23602 =  70.219
0, 1  acc:  8202 / 10053 =  81.588
1, 0  acc: 118777 / 121032 =  98.137
1, 1  acc:  7432 /  8083 =  91.946
--------------------------------------
Average acc: 150984 / 162770 =  92.759
Robust  acc: 16573 / 23602 =  70.219
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6596 /  8535 =  77.282
0, 1  acc:  7075 /  8276 =  85.488
1, 0  acc:  2827 /  2874 =  98.365
1, 1  acc:   173 /   182 =  95.055
------------------------------------
Average acc: 16671 / 19867 =  83.913
Robust  acc:  6596 /  8535 =  77.282
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 85.482
Robust Acc: 82.400 | Best Acc: 98.145
-------------------------------------
Training, Epoch 24:
Accuracies by groups:
0, 0  acc:  8048 /  9767 =  82.400
0, 1  acc:  6416 /  7535 =  85.149
1, 0  acc:  2434 /  2480 =  98.145
1, 1  acc:   166 /   180 =  92.222
------------------------------------
Average acc: 17064 / 19962 =  85.482
Robust  acc:  8048 /  9767 =  82.400
------------------------------------
Accuracies by groups:
0, 0  acc:  8048 /  9767 =  82.400
0, 1  acc:  6416 /  7535 =  85.149
1, 0  acc:  2434 /  2480 =  98.145
1, 1  acc:   166 /   180 =  92.222
------------------------------------
Average acc: 17064 / 19962 =  85.482
Robust  acc:  8048 /  9767 =  82.400
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8048 /  9767 =  82.400
0, 1  acc:  6416 /  7535 =  85.149
1, 0  acc:  2434 /  2480 =  98.145
1, 1  acc:   166 /   180 =  92.222
------------------------------------
Average acc: 17064 / 19962 =  85.482
Robust  acc:  8048 /  9767 =  82.400
------------------------------------
Epoch:  26 | Train Loss: 0.002 | Train Acc: 92.699 | Val Loss: 0.009 | Val Acc: 35.798
Training:
Accuracies by groups:
0, 0  acc: 16548 / 23628 =  70.036
0, 1  acc:  8377 / 10235 =  81.847
1, 0  acc: 118458 / 120727 =  98.121
1, 1  acc:  7503 /  8180 =  91.724
--------------------------------------
Average acc: 150886 / 162770 =  92.699
Robust  acc: 16548 / 23628 =  70.036
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  1687 /  8535 =  19.766
0, 1  acc:  2371 /  8276 =  28.649
1, 0  acc:  2872 /  2874 =  99.930
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc:  7112 / 19867 =  35.798
Robust  acc:  1687 /  8535 =  19.766
------------------------------------
-------------------------------------------
Avg Test Loss: 0.009 | Avg Test Acc: 34.921
Robust Acc: 23.733 | Best Acc: 99.960
-------------------------------------
Training, Epoch 25:
Accuracies by groups:
0, 0  acc:  2318 /  9767 =  23.733
0, 1  acc:  1996 /  7535 =  26.490
1, 0  acc:  2479 /  2480 =  99.960
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  6971 / 19962 =  34.921
Robust  acc:  2318 /  9767 =  23.733
------------------------------------
Accuracies by groups:
0, 0  acc:  2318 /  9767 =  23.733
0, 1  acc:  1996 /  7535 =  26.490
1, 0  acc:  2479 /  2480 =  99.960
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  6971 / 19962 =  34.921
Robust  acc:  2318 /  9767 =  23.733
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  2318 /  9767 =  23.733
0, 1  acc:  1996 /  7535 =  26.490
1, 0  acc:  2479 /  2480 =  99.960
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  6971 / 19962 =  34.921
Robust  acc:  2318 /  9767 =  23.733
------------------------------------
Epoch:  27 | Train Loss: 0.002 | Train Acc: 92.806 | Val Loss: 0.004 | Val Acc: 79.010
Training:
Accuracies by groups:
0, 0  acc: 16374 / 23426 =  69.897
0, 1  acc:  8258 / 10135 =  81.480
1, 0  acc: 118773 / 120929 =  98.217
1, 1  acc:  7656 /  8280 =  92.464
--------------------------------------
Average acc: 151061 / 162770 =  92.806
Robust  acc: 16374 / 23426 =  69.897
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5730 /  8535 =  67.135
0, 1  acc:  6931 /  8276 =  83.748
1, 0  acc:  2858 /  2874 =  99.443
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 15697 / 19867 =  79.010
Robust  acc:  5730 /  8535 =  67.135
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 80.573
Robust Acc: 73.114 | Best Acc: 99.315
-------------------------------------
Training, Epoch 26:
Accuracies by groups:
0, 0  acc:  7141 /  9767 =  73.114
0, 1  acc:  6312 /  7535 =  83.769
1, 0  acc:  2463 /  2480 =  99.315
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 16084 / 19962 =  80.573
Robust  acc:  7141 /  9767 =  73.114
------------------------------------
Accuracies by groups:
0, 0  acc:  7141 /  9767 =  73.114
0, 1  acc:  6312 /  7535 =  83.769
1, 0  acc:  2463 /  2480 =  99.315
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 16084 / 19962 =  80.573
Robust  acc:  7141 /  9767 =  73.114
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7141 /  9767 =  73.114
0, 1  acc:  6312 /  7535 =  83.769
1, 0  acc:  2463 /  2480 =  99.315
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 16084 / 19962 =  80.573
Robust  acc:  7141 /  9767 =  73.114
------------------------------------
Epoch:  28 | Train Loss: 0.002 | Train Acc: 92.647 | Val Loss: 0.002 | Val Acc: 93.769
Training:
Accuracies by groups:
0, 0  acc: 16330 / 23514 =  69.448
0, 1  acc:  8282 / 10235 =  80.918
1, 0  acc: 118656 / 120822 =  98.207
1, 1  acc:  7534 /  8199 =  91.889
--------------------------------------
Average acc: 150802 / 162770 =  92.647
Robust  acc: 16330 / 23514 =  69.448
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8013 /  8535 =  93.884
0, 1  acc:  8029 /  8276 =  97.015
1, 0  acc:  2467 /  2874 =  85.839
1, 1  acc:   120 /   182 =  65.934
------------------------------------
Average acc: 18629 / 19867 =  93.769
Robust  acc:   120 /   182 =  65.934
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 94.269
Robust Acc: 54.444 | Best Acc: 96.868
-------------------------------------
Training, Epoch 27:
Accuracies by groups:
0, 0  acc:  9319 /  9767 =  95.413
0, 1  acc:  7299 /  7535 =  96.868
1, 0  acc:  2102 /  2480 =  84.758
1, 1  acc:    98 /   180 =  54.444
------------------------------------
Average acc: 18818 / 19962 =  94.269
Robust  acc:    98 /   180 =  54.444
------------------------------------
Accuracies by groups:
0, 0  acc:  9319 /  9767 =  95.413
0, 1  acc:  7299 /  7535 =  96.868
1, 0  acc:  2102 /  2480 =  84.758
1, 1  acc:    98 /   180 =  54.444
------------------------------------
Average acc: 18818 / 19962 =  94.269
Robust  acc:    98 /   180 =  54.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9319 /  9767 =  95.413
0, 1  acc:  7299 /  7535 =  96.868
1, 0  acc:  2102 /  2480 =  84.758
1, 1  acc:    98 /   180 =  54.444
------------------------------------
Average acc: 18818 / 19962 =  94.269
Robust  acc:    98 /   180 =  54.444
------------------------------------
Epoch:  29 | Train Loss: 0.002 | Train Acc: 92.746 | Val Loss: 0.004 | Val Acc: 84.210
Training:
Accuracies by groups:
0, 0  acc: 16369 / 23544 =  69.525
0, 1  acc:  8373 / 10327 =  81.079
1, 0  acc: 118594 / 120654 =  98.293
1, 1  acc:  7627 /  8245 =  92.505
--------------------------------------
Average acc: 150963 / 162770 =  92.746
Robust  acc: 16369 / 23544 =  69.525
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6530 /  8535 =  76.508
0, 1  acc:  7205 /  8276 =  87.059
1, 0  acc:  2828 /  2874 =  98.399
1, 1  acc:   167 /   182 =  91.758
------------------------------------
Average acc: 16730 / 19867 =  84.210
Robust  acc:  6530 /  8535 =  76.508
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 85.548
Robust Acc: 80.997 | Best Acc: 98.024
-------------------------------------
Training, Epoch 28:
Accuracies by groups:
0, 0  acc:  7911 /  9767 =  80.997
0, 1  acc:  6573 /  7535 =  87.233
1, 0  acc:  2431 /  2480 =  98.024
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 17077 / 19962 =  85.548
Robust  acc:  7911 /  9767 =  80.997
------------------------------------
Accuracies by groups:
0, 0  acc:  7911 /  9767 =  80.997
0, 1  acc:  6573 /  7535 =  87.233
1, 0  acc:  2431 /  2480 =  98.024
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 17077 / 19962 =  85.548
Robust  acc:  7911 /  9767 =  80.997
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7911 /  9767 =  80.997
0, 1  acc:  6573 /  7535 =  87.233
1, 0  acc:  2431 /  2480 =  98.024
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 17077 / 19962 =  85.548
Robust  acc:  7911 /  9767 =  80.997
------------------------------------
Epoch:  30 | Train Loss: 0.002 | Train Acc: 92.668 | Val Loss: 0.003 | Val Acc: 89.163
Training:
Accuracies by groups:
0, 0  acc: 16498 / 23732 =  69.518
0, 1  acc:  8230 / 10254 =  80.261
1, 0  acc: 118698 / 120776 =  98.279
1, 1  acc:  7409 /  8008 =  92.520
--------------------------------------
Average acc: 150835 / 162770 =  92.668
Robust  acc: 16498 / 23732 =  69.518
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7169 /  8535 =  83.995
0, 1  acc:  7653 /  8276 =  92.472
1, 0  acc:  2731 /  2874 =  95.024
1, 1  acc:   161 /   182 =  88.462
------------------------------------
Average acc: 17714 / 19867 =  89.163
Robust  acc:  7169 /  8535 =  83.995
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 90.196
Robust Acc: 80.000 | Best Acc: 95.444
-------------------------------------
Training, Epoch 29:
Accuracies by groups:
0, 0  acc:  8553 /  9767 =  87.570
0, 1  acc:  6941 /  7535 =  92.117
1, 0  acc:  2367 /  2480 =  95.444
1, 1  acc:   144 /   180 =  80.000
------------------------------------
Average acc: 18005 / 19962 =  90.196
Robust  acc:   144 /   180 =  80.000
------------------------------------
Accuracies by groups:
0, 0  acc:  8553 /  9767 =  87.570
0, 1  acc:  6941 /  7535 =  92.117
1, 0  acc:  2367 /  2480 =  95.444
1, 1  acc:   144 /   180 =  80.000
------------------------------------
Average acc: 18005 / 19962 =  90.196
Robust  acc:   144 /   180 =  80.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8553 /  9767 =  87.570
0, 1  acc:  6941 /  7535 =  92.117
1, 0  acc:  2367 /  2480 =  95.444
1, 1  acc:   144 /   180 =  80.000
------------------------------------
Average acc: 18005 / 19962 =  90.196
Robust  acc:   144 /   180 =  80.000
------------------------------------
Epoch:  31 | Train Loss: 0.002 | Train Acc: 92.809 | Val Loss: 0.006 | Val Acc: 66.739
Training:
Accuracies by groups:
0, 0  acc: 16362 / 23498 =  69.631
0, 1  acc:  8085 / 10069 =  80.296
1, 0  acc: 118952 / 120909 =  98.381
1, 1  acc:  7667 /  8294 =  92.440
--------------------------------------
Average acc: 151066 / 162770 =  92.809
Robust  acc: 16362 / 23498 =  69.631
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4494 /  8535 =  52.654
0, 1  acc:  5718 /  8276 =  69.091
1, 0  acc:  2867 /  2874 =  99.756
1, 1  acc:   180 /   182 =  98.901
------------------------------------
Average acc: 13259 / 19867 =  66.739
Robust  acc:  4494 /  8535 =  52.654
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 68.170
Robust Acc: 59.128 | Best Acc: 99.677
-------------------------------------
Training, Epoch 30:
Accuracies by groups:
0, 0  acc:  5775 /  9767 =  59.128
0, 1  acc:  5185 /  7535 =  68.812
1, 0  acc:  2472 /  2480 =  99.677
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 13608 / 19962 =  68.170
Robust  acc:  5775 /  9767 =  59.128
------------------------------------
Accuracies by groups:
0, 0  acc:  5775 /  9767 =  59.128
0, 1  acc:  5185 /  7535 =  68.812
1, 0  acc:  2472 /  2480 =  99.677
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 13608 / 19962 =  68.170
Robust  acc:  5775 /  9767 =  59.128
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  5775 /  9767 =  59.128
0, 1  acc:  5185 /  7535 =  68.812
1, 0  acc:  2472 /  2480 =  99.677
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 13608 / 19962 =  68.170
Robust  acc:  5775 /  9767 =  59.128
------------------------------------
Epoch:  32 | Train Loss: 0.002 | Train Acc: 92.788 | Val Loss: 0.004 | Val Acc: 79.061
Training:
Accuracies by groups:
0, 0  acc: 16126 / 23292 =  69.234
0, 1  acc:  8336 / 10376 =  80.339
1, 0  acc: 118909 / 120878 =  98.371
1, 1  acc:  7660 /  8224 =  93.142
--------------------------------------
Average acc: 151031 / 162770 =  92.788
Robust  acc: 16126 / 23292 =  69.234
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6259 /  8535 =  73.333
0, 1  acc:  6421 /  8276 =  77.586
1, 0  acc:  2847 /  2874 =  99.061
1, 1  acc:   180 /   182 =  98.901
------------------------------------
Average acc: 15707 / 19867 =  79.061
Robust  acc:  6259 /  8535 =  73.333
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 80.548
Robust Acc: 77.226 | Best Acc: 98.468
-------------------------------------
Training, Epoch 31:
Accuracies by groups:
0, 0  acc:  7645 /  9767 =  78.274
0, 1  acc:  5819 /  7535 =  77.226
1, 0  acc:  2442 /  2480 =  98.468
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 16079 / 19962 =  80.548
Robust  acc:  5819 /  7535 =  77.226
------------------------------------
Accuracies by groups:
0, 0  acc:  7645 /  9767 =  78.274
0, 1  acc:  5819 /  7535 =  77.226
1, 0  acc:  2442 /  2480 =  98.468
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 16079 / 19962 =  80.548
Robust  acc:  5819 /  7535 =  77.226
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7645 /  9767 =  78.274
0, 1  acc:  5819 /  7535 =  77.226
1, 0  acc:  2442 /  2480 =  98.468
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 16079 / 19962 =  80.548
Robust  acc:  5819 /  7535 =  77.226
------------------------------------
Epoch:  33 | Train Loss: 0.002 | Train Acc: 92.746 | Val Loss: 0.005 | Val Acc: 74.601
Training:
Accuracies by groups:
0, 0  acc: 16247 / 23510 =  69.107
0, 1  acc:  8135 / 10154 =  80.116
1, 0  acc: 118841 / 120763 =  98.408
1, 1  acc:  7739 /  8343 =  92.760
--------------------------------------
Average acc: 150962 / 162770 =  92.746
Robust  acc: 16247 / 23510 =  69.107
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5612 /  8535 =  65.753
0, 1  acc:  6170 /  8276 =  74.553
1, 0  acc:  2861 /  2874 =  99.548
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 14821 / 19867 =  74.601
Robust  acc:  5612 /  8535 =  65.753
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 76.470
Robust Acc: 71.731 | Best Acc: 99.274
-------------------------------------
Training, Epoch 32:
Accuracies by groups:
0, 0  acc:  7006 /  9767 =  71.731
0, 1  acc:  5625 /  7535 =  74.652
1, 0  acc:  2462 /  2480 =  99.274
1, 1  acc:   172 /   180 =  95.556
------------------------------------
Average acc: 15265 / 19962 =  76.470
Robust  acc:  7006 /  9767 =  71.731
------------------------------------
Accuracies by groups:
0, 0  acc:  7006 /  9767 =  71.731
0, 1  acc:  5625 /  7535 =  74.652
1, 0  acc:  2462 /  2480 =  99.274
1, 1  acc:   172 /   180 =  95.556
------------------------------------
Average acc: 15265 / 19962 =  76.470
Robust  acc:  7006 /  9767 =  71.731
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7006 /  9767 =  71.731
0, 1  acc:  5625 /  7535 =  74.652
1, 0  acc:  2462 /  2480 =  99.274
1, 1  acc:   172 /   180 =  95.556
------------------------------------
Average acc: 15265 / 19962 =  76.470
Robust  acc:  7006 /  9767 =  71.731
------------------------------------
Epoch:  34 | Train Loss: 0.002 | Train Acc: 92.796 | Val Loss: 0.004 | Val Acc: 83.193
Training:
Accuracies by groups:
0, 0  acc: 16167 / 23460 =  68.913
0, 1  acc:  8208 / 10181 =  80.621
1, 0  acc: 119014 / 120895 =  98.444
1, 1  acc:  7655 /  8234 =  92.968
--------------------------------------
Average acc: 151044 / 162770 =  92.796
Robust  acc: 16167 / 23460 =  68.913
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6488 /  8535 =  76.016
0, 1  acc:  7043 /  8276 =  85.101
1, 0  acc:  2822 /  2874 =  98.191
1, 1  acc:   175 /   182 =  96.154
------------------------------------
Average acc: 16528 / 19867 =  83.193
Robust  acc:  6488 /  8535 =  76.016
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 84.896
Robust Acc: 81.509 | Best Acc: 98.024
-------------------------------------
Training, Epoch 33:
Accuracies by groups:
0, 0  acc:  7961 /  9767 =  81.509
0, 1  acc:  6388 /  7535 =  84.778
1, 0  acc:  2431 /  2480 =  98.024
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16947 / 19962 =  84.896
Robust  acc:  7961 /  9767 =  81.509
------------------------------------
Accuracies by groups:
0, 0  acc:  7961 /  9767 =  81.509
0, 1  acc:  6388 /  7535 =  84.778
1, 0  acc:  2431 /  2480 =  98.024
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16947 / 19962 =  84.896
Robust  acc:  7961 /  9767 =  81.509
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7961 /  9767 =  81.509
0, 1  acc:  6388 /  7535 =  84.778
1, 0  acc:  2431 /  2480 =  98.024
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16947 / 19962 =  84.896
Robust  acc:  7961 /  9767 =  81.509
------------------------------------
Epoch:  35 | Train Loss: 0.002 | Train Acc: 92.827 | Val Loss: 0.004 | Val Acc: 83.958
Training:
Accuracies by groups:
0, 0  acc: 15969 / 23165 =  68.936
0, 1  acc:  8208 / 10170 =  80.708
1, 0  acc: 119252 / 121206 =  98.388
1, 1  acc:  7666 /  8229 =  93.158
--------------------------------------
Average acc: 151095 / 162770 =  92.827
Robust  acc: 15969 / 23165 =  68.936
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6791 /  8535 =  79.566
0, 1  acc:  6903 /  8276 =  83.410
1, 0  acc:  2813 /  2874 =  97.878
1, 1  acc:   173 /   182 =  95.055
------------------------------------
Average acc: 16680 / 19867 =  83.958
Robust  acc:  6791 /  8535 =  79.566
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 85.472
Robust Acc: 83.570 | Best Acc: 97.661
-------------------------------------
Training, Epoch 34:
Accuracies by groups:
0, 0  acc:  8176 /  9767 =  83.710
0, 1  acc:  6297 /  7535 =  83.570
1, 0  acc:  2422 /  2480 =  97.661
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 17062 / 19962 =  85.472
Robust  acc:  6297 /  7535 =  83.570
------------------------------------
Accuracies by groups:
0, 0  acc:  8176 /  9767 =  83.710
0, 1  acc:  6297 /  7535 =  83.570
1, 0  acc:  2422 /  2480 =  97.661
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 17062 / 19962 =  85.472
Robust  acc:  6297 /  7535 =  83.570
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8176 /  9767 =  83.710
0, 1  acc:  6297 /  7535 =  83.570
1, 0  acc:  2422 /  2480 =  97.661
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 17062 / 19962 =  85.472
Robust  acc:  6297 /  7535 =  83.570
------------------------------------
Epoch:  36 | Train Loss: 0.002 | Train Acc: 92.800 | Val Loss: 0.003 | Val Acc: 93.472
Training:
Accuracies by groups:
0, 0  acc: 15999 / 23259 =  68.786
0, 1  acc:  8343 / 10346 =  80.640
1, 0  acc: 119004 / 120879 =  98.449
1, 1  acc:  7705 /  8286 =  92.988
--------------------------------------
Average acc: 151051 / 162770 =  92.800
Robust  acc: 15999 / 23259 =  68.786
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7744 /  8535 =  90.732
0, 1  acc:  8069 /  8276 =  97.499
1, 0  acc:  2631 /  2874 =  91.545
1, 1  acc:   126 /   182 =  69.231
------------------------------------
Average acc: 18570 / 19867 =  93.472
Robust  acc:   126 /   182 =  69.231
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 93.984
Robust Acc: 62.778 | Best Acc: 97.001
-------------------------------------
Training, Epoch 35:
Accuracies by groups:
0, 0  acc:  9083 /  9767 =  92.997
0, 1  acc:  7309 /  7535 =  97.001
1, 0  acc:  2256 /  2480 =  90.968
1, 1  acc:   113 /   180 =  62.778
------------------------------------
Average acc: 18761 / 19962 =  93.984
Robust  acc:   113 /   180 =  62.778
------------------------------------
Accuracies by groups:
0, 0  acc:  9083 /  9767 =  92.997
0, 1  acc:  7309 /  7535 =  97.001
1, 0  acc:  2256 /  2480 =  90.968
1, 1  acc:   113 /   180 =  62.778
------------------------------------
Average acc: 18761 / 19962 =  93.984
Robust  acc:   113 /   180 =  62.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9083 /  9767 =  92.997
0, 1  acc:  7309 /  7535 =  97.001
1, 0  acc:  2256 /  2480 =  90.968
1, 1  acc:   113 /   180 =  62.778
------------------------------------
Average acc: 18761 / 19962 =  93.984
Robust  acc:   113 /   180 =  62.778
------------------------------------
Epoch:  37 | Train Loss: 0.002 | Train Acc: 92.797 | Val Loss: 0.004 | Val Acc: 85.358
Training:
Accuracies by groups:
0, 0  acc: 16154 / 23389 =  69.067
0, 1  acc:  8327 / 10343 =  80.509
1, 0  acc: 119060 / 120930 =  98.454
1, 1  acc:  7505 /  8108 =  92.563
--------------------------------------
Average acc: 151046 / 162770 =  92.797
Robust  acc: 16154 / 23389 =  69.067
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6547 /  8535 =  76.708
0, 1  acc:  7406 /  8276 =  89.488
1, 0  acc:  2835 /  2874 =  98.643
1, 1  acc:   170 /   182 =  93.407
------------------------------------
Average acc: 16958 / 19867 =  85.358
Robust  acc:  6547 /  8535 =  76.708
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 86.735
Robust Acc: 81.847 | Best Acc: 98.427
-------------------------------------
Training, Epoch 36:
Accuracies by groups:
0, 0  acc:  7994 /  9767 =  81.847
0, 1  acc:  6715 /  7535 =  89.117
1, 0  acc:  2441 /  2480 =  98.427
1, 1  acc:   164 /   180 =  91.111
------------------------------------
Average acc: 17314 / 19962 =  86.735
Robust  acc:  7994 /  9767 =  81.847
------------------------------------
Accuracies by groups:
0, 0  acc:  7994 /  9767 =  81.847
0, 1  acc:  6715 /  7535 =  89.117
1, 0  acc:  2441 /  2480 =  98.427
1, 1  acc:   164 /   180 =  91.111
------------------------------------
Average acc: 17314 / 19962 =  86.735
Robust  acc:  7994 /  9767 =  81.847
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7994 /  9767 =  81.847
0, 1  acc:  6715 /  7535 =  89.117
1, 0  acc:  2441 /  2480 =  98.427
1, 1  acc:   164 /   180 =  91.111
------------------------------------
Average acc: 17314 / 19962 =  86.735
Robust  acc:  7994 /  9767 =  81.847
------------------------------------
Epoch:  38 | Train Loss: 0.002 | Train Acc: 92.751 | Val Loss: 0.003 | Val Acc: 90.482
Training:
Accuracies by groups:
0, 0  acc: 16049 / 23273 =  68.960
0, 1  acc:  8291 / 10324 =  80.308
1, 0  acc: 118912 / 120851 =  98.396
1, 1  acc:  7719 /  8322 =  92.754
--------------------------------------
Average acc: 150971 / 162770 =  92.751
Robust  acc: 16049 / 23273 =  68.960
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7408 /  8535 =  86.796
0, 1  acc:  7694 /  8276 =  92.968
1, 0  acc:  2721 /  2874 =  94.676
1, 1  acc:   153 /   182 =  84.066
------------------------------------
Average acc: 17976 / 19867 =  90.482
Robust  acc:   153 /   182 =  84.066
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 91.208
Robust Acc: 77.222 | Best Acc: 93.669
-------------------------------------
Training, Epoch 37:
Accuracies by groups:
0, 0  acc:  8713 /  9767 =  89.209
0, 1  acc:  7032 /  7535 =  93.324
1, 0  acc:  2323 /  2480 =  93.669
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 18207 / 19962 =  91.208
Robust  acc:   139 /   180 =  77.222
------------------------------------
Accuracies by groups:
0, 0  acc:  8713 /  9767 =  89.209
0, 1  acc:  7032 /  7535 =  93.324
1, 0  acc:  2323 /  2480 =  93.669
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 18207 / 19962 =  91.208
Robust  acc:   139 /   180 =  77.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8713 /  9767 =  89.209
0, 1  acc:  7032 /  7535 =  93.324
1, 0  acc:  2323 /  2480 =  93.669
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 18207 / 19962 =  91.208
Robust  acc:   139 /   180 =  77.222
------------------------------------
Epoch:  39 | Train Loss: 0.002 | Train Acc: 92.914 | Val Loss: 0.005 | Val Acc: 70.861
Training:
Accuracies by groups:
0, 0  acc: 16153 / 23309 =  69.299
0, 1  acc:  8369 / 10284 =  81.379
1, 0  acc: 118949 / 120865 =  98.415
1, 1  acc:  7765 /  8312 =  93.419
--------------------------------------
Average acc: 151236 / 162770 =  92.914
Robust  acc: 16153 / 23309 =  69.299
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5380 /  8535 =  63.035
0, 1  acc:  5660 /  8276 =  68.391
1, 0  acc:  2859 /  2874 =  99.478
1, 1  acc:   179 /   182 =  98.352
------------------------------------
Average acc: 14078 / 19867 =  70.861
Robust  acc:  5380 /  8535 =  63.035
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 72.563
Robust Acc: 67.538 | Best Acc: 99.194
-------------------------------------
Training, Epoch 38:
Accuracies by groups:
0, 0  acc:  6759 /  9767 =  69.202
0, 1  acc:  5089 /  7535 =  67.538
1, 0  acc:  2460 /  2480 =  99.194
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 14485 / 19962 =  72.563
Robust  acc:  5089 /  7535 =  67.538
------------------------------------
Accuracies by groups:
0, 0  acc:  6759 /  9767 =  69.202
0, 1  acc:  5089 /  7535 =  67.538
1, 0  acc:  2460 /  2480 =  99.194
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 14485 / 19962 =  72.563
Robust  acc:  5089 /  7535 =  67.538
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6759 /  9767 =  69.202
0, 1  acc:  5089 /  7535 =  67.538
1, 0  acc:  2460 /  2480 =  99.194
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 14485 / 19962 =  72.563
Robust  acc:  5089 /  7535 =  67.538
------------------------------------
Epoch:  40 | Train Loss: 0.002 | Train Acc: 92.993 | Val Loss: 0.003 | Val Acc: 90.653
Training:
Accuracies by groups:
0, 0  acc: 16445 / 23500 =  69.979
0, 1  acc:  8360 / 10222 =  81.784
1, 0  acc: 118850 / 120735 =  98.439
1, 1  acc:  7709 /  8313 =  92.734
--------------------------------------
Average acc: 151364 / 162770 =  92.993
Robust  acc: 16445 / 23500 =  69.979
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7487 /  8535 =  87.721
0, 1  acc:  7646 /  8276 =  92.388
1, 0  acc:  2724 /  2874 =  94.781
1, 1  acc:   153 /   182 =  84.066
------------------------------------
Average acc: 18010 / 19867 =  90.653
Robust  acc:   153 /   182 =  84.066
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 91.519
Robust Acc: 77.778 | Best Acc: 93.427
-------------------------------------
Training, Epoch 39:
Accuracies by groups:
0, 0  acc:  8815 /  9767 =  90.253
0, 1  acc:  6997 /  7535 =  92.860
1, 0  acc:  2317 /  2480 =  93.427
1, 1  acc:   140 /   180 =  77.778
------------------------------------
Average acc: 18269 / 19962 =  91.519
Robust  acc:   140 /   180 =  77.778
------------------------------------
Accuracies by groups:
0, 0  acc:  8815 /  9767 =  90.253
0, 1  acc:  6997 /  7535 =  92.860
1, 0  acc:  2317 /  2480 =  93.427
1, 1  acc:   140 /   180 =  77.778
------------------------------------
Average acc: 18269 / 19962 =  91.519
Robust  acc:   140 /   180 =  77.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8815 /  9767 =  90.253
0, 1  acc:  6997 /  7535 =  92.860
1, 0  acc:  2317 /  2480 =  93.427
1, 1  acc:   140 /   180 =  77.778
------------------------------------
Average acc: 18269 / 19962 =  91.519
Robust  acc:   140 /   180 =  77.778
------------------------------------
Epoch:  41 | Train Loss: 0.002 | Train Acc: 92.971 | Val Loss: 0.004 | Val Acc: 85.312
Training:
Accuracies by groups:
0, 0  acc: 16339 / 23468 =  69.622
0, 1  acc:  8294 / 10193 =  81.370
1, 0  acc: 118999 / 120838 =  98.478
1, 1  acc:  7697 /  8271 =  93.060
--------------------------------------
Average acc: 151329 / 162770 =  92.971
Robust  acc: 16339 / 23468 =  69.622
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6604 /  8535 =  77.376
0, 1  acc:  7348 /  8276 =  88.787
1, 0  acc:  2825 /  2874 =  98.295
1, 1  acc:   172 /   182 =  94.505
------------------------------------
Average acc: 16949 / 19867 =  85.312
Robust  acc:  6604 /  8535 =  77.376
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 86.549
Robust Acc: 82.298 | Best Acc: 97.661
-------------------------------------
Training, Epoch 40:
Accuracies by groups:
0, 0  acc:  8038 /  9767 =  82.298
0, 1  acc:  6658 /  7535 =  88.361
1, 0  acc:  2422 /  2480 =  97.661
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17277 / 19962 =  86.549
Robust  acc:  8038 /  9767 =  82.298
------------------------------------
Accuracies by groups:
0, 0  acc:  8038 /  9767 =  82.298
0, 1  acc:  6658 /  7535 =  88.361
1, 0  acc:  2422 /  2480 =  97.661
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17277 / 19962 =  86.549
Robust  acc:  8038 /  9767 =  82.298
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8038 /  9767 =  82.298
0, 1  acc:  6658 /  7535 =  88.361
1, 0  acc:  2422 /  2480 =  97.661
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17277 / 19962 =  86.549
Robust  acc:  8038 /  9767 =  82.298
------------------------------------
Epoch:  42 | Train Loss: 0.002 | Train Acc: 92.940 | Val Loss: 0.003 | Val Acc: 89.183
Training:
Accuracies by groups:
0, 0  acc: 16671 / 23809 =  70.020
0, 1  acc:  8203 / 10083 =  81.355
1, 0  acc: 118690 / 120596 =  98.420
1, 1  acc:  7715 /  8282 =  93.154
--------------------------------------
Average acc: 151279 / 162770 =  92.940
Robust  acc: 16671 / 23809 =  70.020
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7004 /  8535 =  82.062
0, 1  acc:  7763 /  8276 =  93.801
1, 0  acc:  2797 /  2874 =  97.321
1, 1  acc:   154 /   182 =  84.615
------------------------------------
Average acc: 17718 / 19867 =  89.183
Robust  acc:  7004 /  8535 =  82.062
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 90.191
Robust Acc: 82.222 | Best Acc: 96.532
-------------------------------------
Training, Epoch 41:
Accuracies by groups:
0, 0  acc:  8419 /  9767 =  86.198
0, 1  acc:  7043 /  7535 =  93.470
1, 0  acc:  2394 /  2480 =  96.532
1, 1  acc:   148 /   180 =  82.222
------------------------------------
Average acc: 18004 / 19962 =  90.191
Robust  acc:   148 /   180 =  82.222
------------------------------------
Accuracies by groups:
0, 0  acc:  8419 /  9767 =  86.198
0, 1  acc:  7043 /  7535 =  93.470
1, 0  acc:  2394 /  2480 =  96.532
1, 1  acc:   148 /   180 =  82.222
------------------------------------
Average acc: 18004 / 19962 =  90.191
Robust  acc:   148 /   180 =  82.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8419 /  9767 =  86.198
0, 1  acc:  7043 /  7535 =  93.470
1, 0  acc:  2394 /  2480 =  96.532
1, 1  acc:   148 /   180 =  82.222
------------------------------------
Average acc: 18004 / 19962 =  90.191
Robust  acc:   148 /   180 =  82.222
------------------------------------
Epoch:  43 | Train Loss: 0.002 | Train Acc: 92.908 | Val Loss: 0.003 | Val Acc: 91.156
Training:
Accuracies by groups:
0, 0  acc: 16559 / 23682 =  69.922
0, 1  acc:  8322 / 10225 =  81.389
1, 0  acc: 118706 / 120617 =  98.416
1, 1  acc:  7640 /  8246 =  92.651
--------------------------------------
Average acc: 151227 / 162770 =  92.908
Robust  acc: 16559 / 23682 =  69.922
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7456 /  8535 =  87.358
0, 1  acc:  7796 /  8276 =  94.200
1, 0  acc:  2711 /  2874 =  94.328
1, 1  acc:   147 /   182 =  80.769
------------------------------------
Average acc: 18110 / 19867 =  91.156
Robust  acc:   147 /   182 =  80.769
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 91.855
Robust Acc: 78.333 | Best Acc: 94.121
-------------------------------------
Training, Epoch 42:
Accuracies by groups:
0, 0  acc:  8801 /  9767 =  90.110
0, 1  acc:  7092 /  7535 =  94.121
1, 0  acc:  2302 /  2480 =  92.823
1, 1  acc:   141 /   180 =  78.333
------------------------------------
Average acc: 18336 / 19962 =  91.855
Robust  acc:   141 /   180 =  78.333
------------------------------------
Accuracies by groups:
0, 0  acc:  8801 /  9767 =  90.110
0, 1  acc:  7092 /  7535 =  94.121
1, 0  acc:  2302 /  2480 =  92.823
1, 1  acc:   141 /   180 =  78.333
------------------------------------
Average acc: 18336 / 19962 =  91.855
Robust  acc:   141 /   180 =  78.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8801 /  9767 =  90.110
0, 1  acc:  7092 /  7535 =  94.121
1, 0  acc:  2302 /  2480 =  92.823
1, 1  acc:   141 /   180 =  78.333
------------------------------------
Average acc: 18336 / 19962 =  91.855
Robust  acc:   141 /   180 =  78.333
------------------------------------
Epoch:  44 | Train Loss: 0.002 | Train Acc: 93.074 | Val Loss: 0.003 | Val Acc: 85.332
Training:
Accuracies by groups:
0, 0  acc: 16187 / 23190 =  69.802
0, 1  acc:  8242 / 10101 =  81.596
1, 0  acc: 119459 / 121318 =  98.468
1, 1  acc:  7609 /  8161 =  93.236
--------------------------------------
Average acc: 151497 / 162770 =  93.074
Robust  acc: 16187 / 23190 =  69.802
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6490 /  8535 =  76.040
0, 1  acc:  7471 /  8276 =  90.273
1, 0  acc:  2832 /  2874 =  98.539
1, 1  acc:   160 /   182 =  87.912
------------------------------------
Average acc: 16953 / 19867 =  85.332
Robust  acc:  6490 /  8535 =  76.040
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 86.244
Robust Acc: 80.721 | Best Acc: 97.863
-------------------------------------
Training, Epoch 43:
Accuracies by groups:
0, 0  acc:  7884 /  9767 =  80.721
0, 1  acc:  6746 /  7535 =  89.529
1, 0  acc:  2427 /  2480 =  97.863
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17216 / 19962 =  86.244
Robust  acc:  7884 /  9767 =  80.721
------------------------------------
Accuracies by groups:
0, 0  acc:  7884 /  9767 =  80.721
0, 1  acc:  6746 /  7535 =  89.529
1, 0  acc:  2427 /  2480 =  97.863
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17216 / 19962 =  86.244
Robust  acc:  7884 /  9767 =  80.721
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7884 /  9767 =  80.721
0, 1  acc:  6746 /  7535 =  89.529
1, 0  acc:  2427 /  2480 =  97.863
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17216 / 19962 =  86.244
Robust  acc:  7884 /  9767 =  80.721
------------------------------------
Epoch:  45 | Train Loss: 0.002 | Train Acc: 93.023 | Val Loss: 0.004 | Val Acc: 78.678
Training:
Accuracies by groups:
0, 0  acc: 16412 / 23364 =  70.245
0, 1  acc:  8117 /  9994 =  81.219
1, 0  acc: 119252 / 121163 =  98.423
1, 1  acc:  7633 /  8249 =  92.532
--------------------------------------
Average acc: 151414 / 162770 =  93.023
Robust  acc: 16412 / 23364 =  70.245
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5753 /  8535 =  67.405
0, 1  acc:  6848 /  8276 =  82.745
1, 0  acc:  2855 /  2874 =  99.339
1, 1  acc:   175 /   182 =  96.154
------------------------------------
Average acc: 15631 / 19867 =  78.678
Robust  acc:  5753 /  8535 =  67.405
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 80.303
Robust Acc: 73.523 | Best Acc: 99.032
-------------------------------------
Training, Epoch 44:
Accuracies by groups:
0, 0  acc:  7181 /  9767 =  73.523
0, 1  acc:  6227 /  7535 =  82.641
1, 0  acc:  2456 /  2480 =  99.032
1, 1  acc:   166 /   180 =  92.222
------------------------------------
Average acc: 16030 / 19962 =  80.303
Robust  acc:  7181 /  9767 =  73.523
------------------------------------
Accuracies by groups:
0, 0  acc:  7181 /  9767 =  73.523
0, 1  acc:  6227 /  7535 =  82.641
1, 0  acc:  2456 /  2480 =  99.032
1, 1  acc:   166 /   180 =  92.222
------------------------------------
Average acc: 16030 / 19962 =  80.303
Robust  acc:  7181 /  9767 =  73.523
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7181 /  9767 =  73.523
0, 1  acc:  6227 /  7535 =  82.641
1, 0  acc:  2456 /  2480 =  99.032
1, 1  acc:   166 /   180 =  92.222
------------------------------------
Average acc: 16030 / 19962 =  80.303
Robust  acc:  7181 /  9767 =  73.523
------------------------------------
Epoch:  46 | Train Loss: 0.002 | Train Acc: 93.042 | Val Loss: 0.007 | Val Acc: 59.319
Training:
Accuracies by groups:
0, 0  acc: 16456 / 23446 =  70.187
0, 1  acc:  8242 / 10124 =  81.411
1, 0  acc: 119141 / 121014 =  98.452
1, 1  acc:  7605 /  8186 =  92.903
--------------------------------------
Average acc: 151444 / 162770 =  93.042
Robust  acc: 16456 / 23446 =  70.187
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4332 /  8535 =  50.756
0, 1  acc:  4403 /  8276 =  53.202
1, 0  acc:  2868 /  2874 =  99.791
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc: 11785 / 19867 =  59.319
Robust  acc:  4332 /  8535 =  50.756
------------------------------------
-------------------------------------------
Avg Test Loss: 0.007 | Avg Test Acc: 60.836
Robust Acc: 53.232 | Best Acc: 99.839
-------------------------------------
Training, Epoch 45:
Accuracies by groups:
0, 0  acc:  5479 /  9767 =  56.097
0, 1  acc:  4011 /  7535 =  53.232
1, 0  acc:  2476 /  2480 =  99.839
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 12144 / 19962 =  60.836
Robust  acc:  4011 /  7535 =  53.232
------------------------------------
Accuracies by groups:
0, 0  acc:  5479 /  9767 =  56.097
0, 1  acc:  4011 /  7535 =  53.232
1, 0  acc:  2476 /  2480 =  99.839
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 12144 / 19962 =  60.836
Robust  acc:  4011 /  7535 =  53.232
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  5479 /  9767 =  56.097
0, 1  acc:  4011 /  7535 =  53.232
1, 0  acc:  2476 /  2480 =  99.839
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 12144 / 19962 =  60.836
Robust  acc:  4011 /  7535 =  53.232
------------------------------------
Epoch:  47 | Train Loss: 0.002 | Train Acc: 93.120 | Val Loss: 0.004 | Val Acc: 78.844
Training:
Accuracies by groups:
0, 0  acc: 16502 / 23380 =  70.582
0, 1  acc:  8363 / 10186 =  82.103
1, 0  acc: 119082 / 120980 =  98.431
1, 1  acc:  7625 /  8224 =  92.716
--------------------------------------
Average acc: 151572 / 162770 =  93.120
Robust  acc: 16502 / 23380 =  70.582
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6137 /  8535 =  71.904
0, 1  acc:  6507 /  8276 =  78.625
1, 0  acc:  2842 /  2874 =  98.887
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 15664 / 19867 =  78.844
Robust  acc:  6137 /  8535 =  71.904
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 80.563
Robust Acc: 76.943 | Best Acc: 98.669
-------------------------------------
Training, Epoch 46:
Accuracies by groups:
0, 0  acc:  7515 /  9767 =  76.943
0, 1  acc:  5949 /  7535 =  78.952
1, 0  acc:  2447 /  2480 =  98.669
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 16082 / 19962 =  80.563
Robust  acc:  7515 /  9767 =  76.943
------------------------------------
Accuracies by groups:
0, 0  acc:  7515 /  9767 =  76.943
0, 1  acc:  5949 /  7535 =  78.952
1, 0  acc:  2447 /  2480 =  98.669
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 16082 / 19962 =  80.563
Robust  acc:  7515 /  9767 =  76.943
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7515 /  9767 =  76.943
0, 1  acc:  5949 /  7535 =  78.952
1, 0  acc:  2447 /  2480 =  98.669
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 16082 / 19962 =  80.563
Robust  acc:  7515 /  9767 =  76.943
------------------------------------
Epoch:  48 | Train Loss: 0.002 | Train Acc: 93.119 | Val Loss: 0.006 | Val Acc: 66.855
Training:
Accuracies by groups:
0, 0  acc: 16569 / 23406 =  70.790
0, 1  acc:  8208 / 10094 =  81.316
1, 0  acc: 119365 / 121233 =  98.459
1, 1  acc:  7428 /  8037 =  92.423
--------------------------------------
Average acc: 151570 / 162770 =  93.119
Robust  acc: 16569 / 23406 =  70.790
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4779 /  8535 =  55.993
0, 1  acc:  5461 /  8276 =  65.986
1, 0  acc:  2862 /  2874 =  99.582
1, 1  acc:   180 /   182 =  98.901
------------------------------------
Average acc: 13282 / 19867 =  66.855
Robust  acc:  4779 /  8535 =  55.993
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 68.595
Robust Acc: 62.486 | Best Acc: 99.395
-------------------------------------
Training, Epoch 47:
Accuracies by groups:
0, 0  acc:  6103 /  9767 =  62.486
0, 1  acc:  4952 /  7535 =  65.720
1, 0  acc:  2465 /  2480 =  99.395
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 13693 / 19962 =  68.595
Robust  acc:  6103 /  9767 =  62.486
------------------------------------
Accuracies by groups:
0, 0  acc:  6103 /  9767 =  62.486
0, 1  acc:  4952 /  7535 =  65.720
1, 0  acc:  2465 /  2480 =  99.395
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 13693 / 19962 =  68.595
Robust  acc:  6103 /  9767 =  62.486
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6103 /  9767 =  62.486
0, 1  acc:  4952 /  7535 =  65.720
1, 0  acc:  2465 /  2480 =  99.395
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 13693 / 19962 =  68.595
Robust  acc:  6103 /  9767 =  62.486
------------------------------------
Epoch:  49 | Train Loss: 0.002 | Train Acc: 92.980 | Val Loss: 0.005 | Val Acc: 71.188
Training:
Accuracies by groups:
0, 0  acc: 16567 / 23550 =  70.348
0, 1  acc:  8303 / 10175 =  81.602
1, 0  acc: 118867 / 120858 =  98.353
1, 1  acc:  7606 /  8187 =  92.903
--------------------------------------
Average acc: 151343 / 162770 =  92.980
Robust  acc: 16567 / 23550 =  70.348
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5172 /  8535 =  60.598
0, 1  acc:  5931 /  8276 =  71.665
1, 0  acc:  2863 /  2874 =  99.617
1, 1  acc:   177 /   182 =  97.253
------------------------------------
Average acc: 14143 / 19867 =  71.188
Robust  acc:  5172 /  8535 =  60.598
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 73.450
Robust Acc: 67.298 | Best Acc: 99.718
-------------------------------------
Training, Epoch 48:
Accuracies by groups:
0, 0  acc:  6573 /  9767 =  67.298
0, 1  acc:  5442 /  7535 =  72.223
1, 0  acc:  2473 /  2480 =  99.718
1, 1  acc:   174 /   180 =  96.667
------------------------------------
Average acc: 14662 / 19962 =  73.450
Robust  acc:  6573 /  9767 =  67.298
------------------------------------
Accuracies by groups:
0, 0  acc:  6573 /  9767 =  67.298
0, 1  acc:  5442 /  7535 =  72.223
1, 0  acc:  2473 /  2480 =  99.718
1, 1  acc:   174 /   180 =  96.667
------------------------------------
Average acc: 14662 / 19962 =  73.450
Robust  acc:  6573 /  9767 =  67.298
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6573 /  9767 =  67.298
0, 1  acc:  5442 /  7535 =  72.223
1, 0  acc:  2473 /  2480 =  99.718
1, 1  acc:   174 /   180 =  96.667
------------------------------------
Average acc: 14662 / 19962 =  73.450
Robust  acc:  6573 /  9767 =  67.298
------------------------------------
Epoch:  50 | Train Loss: 0.002 | Train Acc: 93.060 | Val Loss: 0.003 | Val Acc: 91.272
Training:
Accuracies by groups:
0, 0  acc: 16476 / 23437 =  70.299
0, 1  acc:  8363 / 10176 =  82.184
1, 0  acc: 119156 / 121030 =  98.452
1, 1  acc:  7478 /  8127 =  92.014
--------------------------------------
Average acc: 151473 / 162770 =  93.060
Robust  acc: 16476 / 23437 =  70.299
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7415 /  8535 =  86.878
0, 1  acc:  7819 /  8276 =  94.478
1, 0  acc:  2750 /  2874 =  95.685
1, 1  acc:   149 /   182 =  81.868
------------------------------------
Average acc: 18133 / 19867 =  91.272
Robust  acc:   149 /   182 =  81.868
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 92.175
Robust Acc: 77.222 | Best Acc: 95.050
-------------------------------------
Training, Epoch 49:
Accuracies by groups:
0, 0  acc:  8749 /  9767 =  89.577
0, 1  acc:  7162 /  7535 =  95.050
1, 0  acc:  2350 /  2480 =  94.758
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 18400 / 19962 =  92.175
Robust  acc:   139 /   180 =  77.222
------------------------------------
Accuracies by groups:
0, 0  acc:  8749 /  9767 =  89.577
0, 1  acc:  7162 /  7535 =  95.050
1, 0  acc:  2350 /  2480 =  94.758
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 18400 / 19962 =  92.175
Robust  acc:   139 /   180 =  77.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8749 /  9767 =  89.577
0, 1  acc:  7162 /  7535 =  95.050
1, 0  acc:  2350 /  2480 =  94.758
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 18400 / 19962 =  92.175
Robust  acc:   139 /   180 =  77.222
------------------------------------
replace: True
-> Updating checkpoint debias-end_seed31.pt...
Checkpoint saved at ./model/celebA/config/debias-end_seed31.pt
