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/seed36/stage_one_erm_model_b_epoch0_seed36.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: 11
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: 1.0
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=11-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/seed36/stage_one_erm_model_b_epoch0_seed36.pt
======
# Calculate probability ...
======
======
p_y_a:  tensor([[0.8309, 0.0200],
        [0.1350, 0.0141]])
p_y:  tensor([0.8509, 0.1491])
# Load biased model ...
======
Epoch:   1 | Train Loss: 0.002 | Train Acc: 90.733 | Val Loss: 0.003 | Val Acc: 79.252
Training:
Accuracies by groups:
0, 0  acc:  5764 / 14545 =  39.629
0, 1  acc:  2754 /  6458 =  42.645
1, 0  acc: 130824 / 133014 =  98.354
1, 1  acc:  8344 /  8753 =  95.327
--------------------------------------
Average acc: 147686 / 162770 =  90.733
Robust  acc:  5764 / 14545 =  39.629
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6193 /  8535 =  72.560
0, 1  acc:  6539 /  8276 =  79.012
1, 0  acc:  2845 /  2874 =  98.991
1, 1  acc:   168 /   182 =  92.308
------------------------------------
Average acc: 15745 / 19867 =  79.252
Robust  acc:  6193 /  8535 =  72.560
------------------------------------
New max robust acc: 72.56004686584652
debias model - Saving best checkpoint at epoch 0
replace: True
-> Updating checkpoint debias-wga-best_seed11.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed11.pt
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 81.360
Robust Acc: 78.120 | Best Acc: 98.548
-------------------------------------
Training, Epoch 0:
Accuracies by groups:
0, 0  acc:  7630 /  9767 =  78.120
0, 1  acc:  5996 /  7535 =  79.575
1, 0  acc:  2444 /  2480 =  98.548
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 16241 / 19962 =  81.360
Robust  acc:  7630 /  9767 =  78.120
------------------------------------
Accuracies by groups:
0, 0  acc:  7630 /  9767 =  78.120
0, 1  acc:  5996 /  7535 =  79.575
1, 0  acc:  2444 /  2480 =  98.548
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 16241 / 19962 =  81.360
Robust  acc:  7630 /  9767 =  78.120
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7630 /  9767 =  78.120
0, 1  acc:  5996 /  7535 =  79.575
1, 0  acc:  2444 /  2480 =  98.548
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 16241 / 19962 =  81.360
Robust  acc:  7630 /  9767 =  78.120
------------------------------------
Epoch:   2 | Train Loss: 0.001 | Train Acc: 94.509 | Val Loss: 0.003 | Val Acc: 83.480
Training:
Accuracies by groups:
0, 0  acc:  9217 / 14515 =  63.500
0, 1  acc:  4628 /  6484 =  71.376
1, 0  acc: 131995 / 133266 =  99.046
1, 1  acc:  7992 /  8505 =  93.968
--------------------------------------
Average acc: 153832 / 162770 =  94.509
Robust  acc:  9217 / 14515 =  63.500
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6437 /  8535 =  75.419
0, 1  acc:  7134 /  8276 =  86.201
1, 0  acc:  2841 /  2874 =  98.852
1, 1  acc:   173 /   182 =  95.055
------------------------------------
Average acc: 16585 / 19867 =  83.480
Robust  acc:  6437 /  8535 =  75.419
------------------------------------
New max robust acc: 75.41886350322203
debias model - Saving best checkpoint at epoch 1
replace: True
-> Updating checkpoint debias-wga-best_seed11.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed11.pt
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 85.087
Robust Acc: 80.690 | Best Acc: 98.710
-------------------------------------
Training, Epoch 1:
Accuracies by groups:
0, 0  acc:  7881 /  9767 =  80.690
0, 1  acc:  6489 /  7535 =  86.118
1, 0  acc:  2448 /  2480 =  98.710
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16985 / 19962 =  85.087
Robust  acc:  7881 /  9767 =  80.690
------------------------------------
Accuracies by groups:
0, 0  acc:  7881 /  9767 =  80.690
0, 1  acc:  6489 /  7535 =  86.118
1, 0  acc:  2448 /  2480 =  98.710
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16985 / 19962 =  85.087
Robust  acc:  7881 /  9767 =  80.690
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7881 /  9767 =  80.690
0, 1  acc:  6489 /  7535 =  86.118
1, 0  acc:  2448 /  2480 =  98.710
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16985 / 19962 =  85.087
Robust  acc:  7881 /  9767 =  80.690
------------------------------------
Epoch:   3 | Train Loss: 0.001 | Train Acc: 95.250 | Val Loss: 0.003 | Val Acc: 86.032
Training:
Accuracies by groups:
0, 0  acc:  9841 / 14596 =  67.423
0, 1  acc:  5244 /  6595 =  79.515
1, 0  acc: 131716 / 132870 =  99.131
1, 1  acc:  8237 /  8709 =  94.580
--------------------------------------
Average acc: 155038 / 162770 =  95.250
Robust  acc:  9841 / 14596 =  67.423
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6682 /  8535 =  78.289
0, 1  acc:  7399 /  8276 =  89.403
1, 0  acc:  2837 /  2874 =  98.713
1, 1  acc:   174 /   182 =  95.604
------------------------------------
Average acc: 17092 / 19867 =  86.032
Robust  acc:  6682 /  8535 =  78.289
------------------------------------
New max robust acc: 78.28939660222612
debias model - Saving best checkpoint at epoch 2
replace: True
-> Updating checkpoint debias-wga-best_seed11.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed11.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 87.010
Robust Acc: 82.646 | Best Acc: 98.306
-------------------------------------
Training, Epoch 2:
Accuracies by groups:
0, 0  acc:  8072 /  9767 =  82.646
0, 1  acc:  6697 /  7535 =  88.879
1, 0  acc:  2438 /  2480 =  98.306
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 17369 / 19962 =  87.010
Robust  acc:  8072 /  9767 =  82.646
------------------------------------
Accuracies by groups:
0, 0  acc:  8072 /  9767 =  82.646
0, 1  acc:  6697 /  7535 =  88.879
1, 0  acc:  2438 /  2480 =  98.306
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 17369 / 19962 =  87.010
Robust  acc:  8072 /  9767 =  82.646
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8072 /  9767 =  82.646
0, 1  acc:  6697 /  7535 =  88.879
1, 0  acc:  2438 /  2480 =  98.306
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 17369 / 19962 =  87.010
Robust  acc:  8072 /  9767 =  82.646
------------------------------------
Epoch:   4 | Train Loss: 0.001 | Train Acc: 95.796 | Val Loss: 0.002 | Val Acc: 87.517
Training:
Accuracies by groups:
0, 0  acc: 10148 / 14490 =  70.035
0, 1  acc:  5400 /  6484 =  83.282
1, 0  acc: 132100 / 133099 =  99.249
1, 1  acc:  8279 /  8697 =  95.194
--------------------------------------
Average acc: 155927 / 162770 =  95.796
Robust  acc: 10148 / 14490 =  70.035
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6808 /  8535 =  79.766
0, 1  acc:  7573 /  8276 =  91.506
1, 0  acc:  2833 /  2874 =  98.573
1, 1  acc:   173 /   182 =  95.055
------------------------------------
Average acc: 17387 / 19867 =  87.517
Robust  acc:  6808 /  8535 =  79.766
------------------------------------
New max robust acc: 79.76567076742823
debias model - Saving best checkpoint at epoch 3
replace: True
-> Updating checkpoint debias-wga-best_seed11.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed11.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 88.433
Robust Acc: 83.956 | Best Acc: 98.105
-------------------------------------
Training, Epoch 3:
Accuracies by groups:
0, 0  acc:  8200 /  9767 =  83.956
0, 1  acc:  6860 /  7535 =  91.042
1, 0  acc:  2433 /  2480 =  98.105
1, 1  acc:   160 /   180 =  88.889
------------------------------------
Average acc: 17653 / 19962 =  88.433
Robust  acc:  8200 /  9767 =  83.956
------------------------------------
Accuracies by groups:
0, 0  acc:  8200 /  9767 =  83.956
0, 1  acc:  6860 /  7535 =  91.042
1, 0  acc:  2433 /  2480 =  98.105
1, 1  acc:   160 /   180 =  88.889
------------------------------------
Average acc: 17653 / 19962 =  88.433
Robust  acc:  8200 /  9767 =  83.956
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8200 /  9767 =  83.956
0, 1  acc:  6860 /  7535 =  91.042
1, 0  acc:  2433 /  2480 =  98.105
1, 1  acc:   160 /   180 =  88.889
------------------------------------
Average acc: 17653 / 19962 =  88.433
Robust  acc:  8200 /  9767 =  83.956
------------------------------------
Epoch:   5 | Train Loss: 0.001 | Train Acc: 96.288 | Val Loss: 0.002 | Val Acc: 88.896
Training:
Accuracies by groups:
0, 0  acc: 10407 / 14439 =  72.076
0, 1  acc:  5656 /  6551 =  86.338
1, 0  acc: 132536 / 133358 =  99.384
1, 1  acc:  8129 /  8422 =  96.521
--------------------------------------
Average acc: 156728 / 162770 =  96.288
Robust  acc: 10407 / 14439 =  72.076
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6990 /  8535 =  81.898
0, 1  acc:  7676 /  8276 =  92.750
1, 0  acc:  2824 /  2874 =  98.260
1, 1  acc:   171 /   182 =  93.956
------------------------------------
Average acc: 17661 / 19867 =  88.896
Robust  acc:  6990 /  8535 =  81.898
------------------------------------
New max robust acc: 81.89806678383128
debias model - Saving best checkpoint at epoch 4
replace: True
-> Updating checkpoint debias-wga-best_seed11.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed11.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 89.615
Robust Acc: 85.461 | Best Acc: 97.661
-------------------------------------
Training, Epoch 4:
Accuracies by groups:
0, 0  acc:  8347 /  9767 =  85.461
0, 1  acc:  6962 /  7535 =  92.395
1, 0  acc:  2422 /  2480 =  97.661
1, 1  acc:   158 /   180 =  87.778
------------------------------------
Average acc: 17889 / 19962 =  89.615
Robust  acc:  8347 /  9767 =  85.461
------------------------------------
Accuracies by groups:
0, 0  acc:  8347 /  9767 =  85.461
0, 1  acc:  6962 /  7535 =  92.395
1, 0  acc:  2422 /  2480 =  97.661
1, 1  acc:   158 /   180 =  87.778
------------------------------------
Average acc: 17889 / 19962 =  89.615
Robust  acc:  8347 /  9767 =  85.461
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8347 /  9767 =  85.461
0, 1  acc:  6962 /  7535 =  92.395
1, 0  acc:  2422 /  2480 =  97.661
1, 1  acc:   158 /   180 =  87.778
------------------------------------
Average acc: 17889 / 19962 =  89.615
Robust  acc:  8347 /  9767 =  85.461
------------------------------------
Epoch:   6 | Train Loss: 0.001 | Train Acc: 96.739 | Val Loss: 0.002 | Val Acc: 89.898
Training:
Accuracies by groups:
0, 0  acc: 10704 / 14392 =  74.375
0, 1  acc:  5741 /  6544 =  87.729
1, 0  acc: 132750 / 133353 =  99.548
1, 1  acc:  8267 /  8481 =  97.477
--------------------------------------
Average acc: 157462 / 162770 =  96.739
Robust  acc: 10704 / 14392 =  74.375
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7076 /  8535 =  82.906
0, 1  acc:  7794 /  8276 =  94.176
1, 0  acc:  2823 /  2874 =  98.225
1, 1  acc:   167 /   182 =  91.758
------------------------------------
Average acc: 17860 / 19867 =  89.898
Robust  acc:  7076 /  8535 =  82.906
------------------------------------
New max robust acc: 82.90568248388986
debias model - Saving best checkpoint at epoch 5
replace: True
-> Updating checkpoint debias-wga-best_seed11.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed11.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 90.422
Robust Acc: 86.178 | Best Acc: 97.782
-------------------------------------
Training, Epoch 5:
Accuracies by groups:
0, 0  acc:  8417 /  9767 =  86.178
0, 1  acc:  7052 /  7535 =  93.590
1, 0  acc:  2425 /  2480 =  97.782
1, 1  acc:   156 /   180 =  86.667
------------------------------------
Average acc: 18050 / 19962 =  90.422
Robust  acc:  8417 /  9767 =  86.178
------------------------------------
Accuracies by groups:
0, 0  acc:  8417 /  9767 =  86.178
0, 1  acc:  7052 /  7535 =  93.590
1, 0  acc:  2425 /  2480 =  97.782
1, 1  acc:   156 /   180 =  86.667
------------------------------------
Average acc: 18050 / 19962 =  90.422
Robust  acc:  8417 /  9767 =  86.178
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8417 /  9767 =  86.178
0, 1  acc:  7052 /  7535 =  93.590
1, 0  acc:  2425 /  2480 =  97.782
1, 1  acc:   156 /   180 =  86.667
------------------------------------
Average acc: 18050 / 19962 =  90.422
Robust  acc:  8417 /  9767 =  86.178
------------------------------------
Epoch:   7 | Train Loss: 0.001 | Train Acc: 97.230 | Val Loss: 0.002 | Val Acc: 90.175
Training:
Accuracies by groups:
0, 0  acc: 11003 / 14392 =  76.452
0, 1  acc:  5927 /  6603 =  89.762
1, 0  acc: 132792 / 133140 =  99.739
1, 1  acc:  8539 /  8635 =  98.888
--------------------------------------
Average acc: 158261 / 162770 =  97.230
Robust  acc: 11003 / 14392 =  76.452
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7131 /  8535 =  83.550
0, 1  acc:  7807 /  8276 =  94.333
1, 0  acc:  2812 /  2874 =  97.843
1, 1  acc:   165 /   182 =  90.659
------------------------------------
Average acc: 17915 / 19867 =  90.175
Robust  acc:  7131 /  8535 =  83.550
------------------------------------
New max robust acc: 83.55008787346222
debias model - Saving best checkpoint at epoch 6
replace: True
-> Updating checkpoint debias-wga-best_seed11.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed11.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 90.672
Robust Acc: 85.556 | Best Acc: 97.419
-------------------------------------
Training, Epoch 6:
Accuracies by groups:
0, 0  acc:  8462 /  9767 =  86.639
0, 1  acc:  7068 /  7535 =  93.802
1, 0  acc:  2416 /  2480 =  97.419
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 18100 / 19962 =  90.672
Robust  acc:   154 /   180 =  85.556
------------------------------------
Accuracies by groups:
0, 0  acc:  8462 /  9767 =  86.639
0, 1  acc:  7068 /  7535 =  93.802
1, 0  acc:  2416 /  2480 =  97.419
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 18100 / 19962 =  90.672
Robust  acc:   154 /   180 =  85.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8462 /  9767 =  86.639
0, 1  acc:  7068 /  7535 =  93.802
1, 0  acc:  2416 /  2480 =  97.419
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 18100 / 19962 =  90.672
Robust  acc:   154 /   180 =  85.556
------------------------------------
Epoch:   8 | Train Loss: 0.001 | Train Acc: 97.739 | Val Loss: 0.002 | Val Acc: 91.458
Training:
Accuracies by groups:
0, 0  acc: 11595 / 14503 =  79.949
0, 1  acc:  6121 /  6687 =  91.536
1, 0  acc: 132606 / 132775 =  99.873
1, 1  acc:  8767 /  8805 =  99.568
--------------------------------------
Average acc: 159089 / 162770 =  97.739
Robust  acc: 11595 / 14503 =  79.949
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7298 /  8535 =  85.507
0, 1  acc:  7921 /  8276 =  95.710
1, 0  acc:  2793 /  2874 =  97.182
1, 1  acc:   158 /   182 =  86.813
------------------------------------
Average acc: 18170 / 19867 =  91.458
Robust  acc:  7298 /  8535 =  85.507
------------------------------------
New max robust acc: 85.50673696543643
debias model - Saving best checkpoint at epoch 7
replace: True
-> Updating checkpoint debias-wga-best_seed11.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed11.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.990
Robust Acc: 77.778 | Best Acc: 97.016
-------------------------------------
Training, Epoch 7:
Accuracies by groups:
0, 0  acc:  8643 /  9767 =  88.492
0, 1  acc:  7174 /  7535 =  95.209
1, 0  acc:  2406 /  2480 =  97.016
1, 1  acc:   140 /   180 =  77.778
------------------------------------
Average acc: 18363 / 19962 =  91.990
Robust  acc:   140 /   180 =  77.778
------------------------------------
Accuracies by groups:
0, 0  acc:  8643 /  9767 =  88.492
0, 1  acc:  7174 /  7535 =  95.209
1, 0  acc:  2406 /  2480 =  97.016
1, 1  acc:   140 /   180 =  77.778
------------------------------------
Average acc: 18363 / 19962 =  91.990
Robust  acc:   140 /   180 =  77.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8643 /  9767 =  88.492
0, 1  acc:  7174 /  7535 =  95.209
1, 0  acc:  2406 /  2480 =  97.016
1, 1  acc:   140 /   180 =  77.778
------------------------------------
Average acc: 18363 / 19962 =  91.990
Robust  acc:   140 /   180 =  77.778
------------------------------------
Epoch:   9 | Train Loss: 0.001 | Train Acc: 98.176 | Val Loss: 0.001 | Val Acc: 93.527
Training:
Accuracies by groups:
0, 0  acc: 12186 / 14571 =  83.632
0, 1  acc:  6221 /  6671 =  93.254
1, 0  acc: 132672 / 132787 =  99.913
1, 1  acc:  8722 /  8741 =  99.783
--------------------------------------
Average acc: 159801 / 162770 =  98.176
Robust  acc: 12186 / 14571 =  83.632
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7662 /  8535 =  89.772
0, 1  acc:  8051 /  8276 =  97.281
1, 0  acc:  2728 /  2874 =  94.920
1, 1  acc:   140 /   182 =  76.923
------------------------------------
Average acc: 18581 / 19867 =  93.527
Robust  acc:   140 /   182 =  76.923
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 93.813
Robust Acc: 72.778 | Best Acc: 97.147
-------------------------------------
Training, Epoch 8:
Accuracies by groups:
0, 0  acc:  8955 /  9767 =  91.686
0, 1  acc:  7320 /  7535 =  97.147
1, 0  acc:  2321 /  2480 =  93.589
1, 1  acc:   131 /   180 =  72.778
------------------------------------
Average acc: 18727 / 19962 =  93.813
Robust  acc:   131 /   180 =  72.778
------------------------------------
Accuracies by groups:
0, 0  acc:  8955 /  9767 =  91.686
0, 1  acc:  7320 /  7535 =  97.147
1, 0  acc:  2321 /  2480 =  93.589
1, 1  acc:   131 /   180 =  72.778
------------------------------------
Average acc: 18727 / 19962 =  93.813
Robust  acc:   131 /   180 =  72.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8955 /  9767 =  91.686
0, 1  acc:  7320 /  7535 =  97.147
1, 0  acc:  2321 /  2480 =  93.589
1, 1  acc:   131 /   180 =  72.778
------------------------------------
Average acc: 18727 / 19962 =  93.813
Robust  acc:   131 /   180 =  72.778
------------------------------------
Epoch:  10 | Train Loss: 0.001 | Train Acc: 98.477 | Val Loss: 0.001 | Val Acc: 93.965
Training:
Accuracies by groups:
0, 0  acc: 12455 / 14456 =  86.158
0, 1  acc:  6129 /  6469 =  94.744
1, 0  acc: 133103 / 133208 =  99.921
1, 1  acc:  8604 /  8637 =  99.618
--------------------------------------
Average acc: 160291 / 162770 =  98.477
Robust  acc: 12455 / 14456 =  86.158
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7788 /  8535 =  91.248
0, 1  acc:  8101 /  8276 =  97.885
1, 0  acc:  2655 /  2874 =  92.380
1, 1  acc:   124 /   182 =  68.132
------------------------------------
Average acc: 18668 / 19867 =  93.965
Robust  acc:   124 /   182 =  68.132
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.169
Robust Acc: 67.222 | Best Acc: 97.465
-------------------------------------
Training, Epoch 9:
Accuracies by groups:
0, 0  acc:  9067 /  9767 =  92.833
0, 1  acc:  7344 /  7535 =  97.465
1, 0  acc:  2266 /  2480 =  91.371
1, 1  acc:   121 /   180 =  67.222
------------------------------------
Average acc: 18798 / 19962 =  94.169
Robust  acc:   121 /   180 =  67.222
------------------------------------
Accuracies by groups:
0, 0  acc:  9067 /  9767 =  92.833
0, 1  acc:  7344 /  7535 =  97.465
1, 0  acc:  2266 /  2480 =  91.371
1, 1  acc:   121 /   180 =  67.222
------------------------------------
Average acc: 18798 / 19962 =  94.169
Robust  acc:   121 /   180 =  67.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9067 /  9767 =  92.833
0, 1  acc:  7344 /  7535 =  97.465
1, 0  acc:  2266 /  2480 =  91.371
1, 1  acc:   121 /   180 =  67.222
------------------------------------
Average acc: 18798 / 19962 =  94.169
Robust  acc:   121 /   180 =  67.222
------------------------------------
Epoch:  11 | Train Loss: 0.001 | Train Acc: 98.492 | Val Loss: 0.001 | Val Acc: 93.356
Training:
Accuracies by groups:
0, 0  acc: 12434 / 14378 =  86.479
0, 1  acc:  6045 /  6342 =  95.317
1, 0  acc: 133213 / 133373 =  99.880
1, 1  acc:  8623 /  8677 =  99.378
--------------------------------------
Average acc: 160315 / 162770 =  98.492
Robust  acc: 12434 / 14378 =  86.479
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7680 /  8535 =  89.982
0, 1  acc:  8032 /  8276 =  97.052
1, 0  acc:  2693 /  2874 =  93.702
1, 1  acc:   142 /   182 =  78.022
------------------------------------
Average acc: 18547 / 19867 =  93.356
Robust  acc:   142 /   182 =  78.022
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 93.613
Robust Acc: 71.667 | Best Acc: 96.802
-------------------------------------
Training, Epoch 10:
Accuracies by groups:
0, 0  acc:  8980 /  9767 =  91.942
0, 1  acc:  7294 /  7535 =  96.802
1, 0  acc:  2284 /  2480 =  92.097
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18687 / 19962 =  93.613
Robust  acc:   129 /   180 =  71.667
------------------------------------
Accuracies by groups:
0, 0  acc:  8980 /  9767 =  91.942
0, 1  acc:  7294 /  7535 =  96.802
1, 0  acc:  2284 /  2480 =  92.097
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18687 / 19962 =  93.613
Robust  acc:   129 /   180 =  71.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8980 /  9767 =  91.942
0, 1  acc:  7294 /  7535 =  96.802
1, 0  acc:  2284 /  2480 =  92.097
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18687 / 19962 =  93.613
Robust  acc:   129 /   180 =  71.667
------------------------------------
Epoch:  12 | Train Loss: 0.001 | Train Acc: 98.291 | Val Loss: 0.002 | Val Acc: 91.765
Training:
Accuracies by groups:
0, 0  acc: 12320 / 14428 =  85.390
0, 1  acc:  6138 /  6494 =  94.518
1, 0  acc: 133024 / 133266 =  99.818
1, 1  acc:  8506 /  8582 =  99.114
--------------------------------------
Average acc: 159988 / 162770 =  98.291
Robust  acc: 12320 / 14428 =  85.390
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7413 /  8535 =  86.854
0, 1  acc:  7927 /  8276 =  95.783
1, 0  acc:  2741 /  2874 =  95.372
1, 1  acc:   150 /   182 =  82.418
------------------------------------
Average acc: 18231 / 19867 =  91.765
Robust  acc:   150 /   182 =  82.418
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.125
Robust Acc: 71.667 | Best Acc: 95.408
-------------------------------------
Training, Epoch 11:
Accuracies by groups:
0, 0  acc:  8726 /  9767 =  89.342
0, 1  acc:  7189 /  7535 =  95.408
1, 0  acc:  2346 /  2480 =  94.597
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18390 / 19962 =  92.125
Robust  acc:   129 /   180 =  71.667
------------------------------------
Accuracies by groups:
0, 0  acc:  8726 /  9767 =  89.342
0, 1  acc:  7189 /  7535 =  95.408
1, 0  acc:  2346 /  2480 =  94.597
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18390 / 19962 =  92.125
Robust  acc:   129 /   180 =  71.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8726 /  9767 =  89.342
0, 1  acc:  7189 /  7535 =  95.408
1, 0  acc:  2346 /  2480 =  94.597
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18390 / 19962 =  92.125
Robust  acc:   129 /   180 =  71.667
------------------------------------
Epoch:  13 | Train Loss: 0.001 | Train Acc: 97.802 | Val Loss: 0.002 | Val Acc: 91.916
Training:
Accuracies by groups:
0, 0  acc: 12199 / 14693 =  83.026
0, 1  acc:  6180 /  6674 =  92.598
1, 0  acc: 132275 / 132729 =  99.658
1, 1  acc:  8538 /  8674 =  98.432
--------------------------------------
Average acc: 159192 / 162770 =  97.802
Robust  acc: 12199 / 14693 =  83.026
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7406 /  8535 =  86.772
0, 1  acc:  7916 /  8276 =  95.650
1, 0  acc:  2783 /  2874 =  96.834
1, 1  acc:   156 /   182 =  85.714
------------------------------------
Average acc: 18261 / 19867 =  91.916
Robust  acc:   156 /   182 =  85.714
------------------------------------
New max robust acc: 85.71428571428571
debias model - Saving best checkpoint at epoch 12
replace: True
-> Updating checkpoint debias-wga-best_seed11.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed11.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.401
Robust Acc: 76.111 | Best Acc: 95.820
-------------------------------------
Training, Epoch 12:
Accuracies by groups:
0, 0  acc:  8714 /  9767 =  89.219
0, 1  acc:  7220 /  7535 =  95.820
1, 0  acc:  2374 /  2480 =  95.726
1, 1  acc:   137 /   180 =  76.111
------------------------------------
Average acc: 18445 / 19962 =  92.401
Robust  acc:   137 /   180 =  76.111
------------------------------------
Accuracies by groups:
0, 0  acc:  8714 /  9767 =  89.219
0, 1  acc:  7220 /  7535 =  95.820
1, 0  acc:  2374 /  2480 =  95.726
1, 1  acc:   137 /   180 =  76.111
------------------------------------
Average acc: 18445 / 19962 =  92.401
Robust  acc:   137 /   180 =  76.111
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8714 /  9767 =  89.219
0, 1  acc:  7220 /  7535 =  95.820
1, 0  acc:  2374 /  2480 =  95.726
1, 1  acc:   137 /   180 =  76.111
------------------------------------
Average acc: 18445 / 19962 =  92.401
Robust  acc:   137 /   180 =  76.111
------------------------------------
Epoch:  14 | Train Loss: 0.001 | Train Acc: 97.397 | Val Loss: 0.002 | Val Acc: 91.045
Training:
Accuracies by groups:
0, 0  acc: 11600 / 14457 =  80.238
0, 1  acc:  5964 /  6521 =  91.458
1, 0  acc: 132550 / 133178 =  99.528
1, 1  acc:  8419 /  8614 =  97.736
--------------------------------------
Average acc: 158533 / 162770 =  97.397
Robust  acc: 11600 / 14457 =  80.238
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7232 /  8535 =  84.733
0, 1  acc:  7907 /  8276 =  95.541
1, 0  acc:  2795 /  2874 =  97.251
1, 1  acc:   154 /   182 =  84.615
------------------------------------
Average acc: 18088 / 19867 =  91.045
Robust  acc:   154 /   182 =  84.615
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.479
Robust Acc: 78.889 | Best Acc: 96.815
-------------------------------------
Training, Epoch 13:
Accuracies by groups:
0, 0  acc:  8546 /  9767 =  87.499
0, 1  acc:  7172 /  7535 =  95.182
1, 0  acc:  2401 /  2480 =  96.815
1, 1  acc:   142 /   180 =  78.889
------------------------------------
Average acc: 18261 / 19962 =  91.479
Robust  acc:   142 /   180 =  78.889
------------------------------------
Accuracies by groups:
0, 0  acc:  8546 /  9767 =  87.499
0, 1  acc:  7172 /  7535 =  95.182
1, 0  acc:  2401 /  2480 =  96.815
1, 1  acc:   142 /   180 =  78.889
------------------------------------
Average acc: 18261 / 19962 =  91.479
Robust  acc:   142 /   180 =  78.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8546 /  9767 =  87.499
0, 1  acc:  7172 /  7535 =  95.182
1, 0  acc:  2401 /  2480 =  96.815
1, 1  acc:   142 /   180 =  78.889
------------------------------------
Average acc: 18261 / 19962 =  91.479
Robust  acc:   142 /   180 =  78.889
------------------------------------
Epoch:  15 | Train Loss: 0.001 | Train Acc: 96.893 | Val Loss: 0.003 | Val Acc: 83.767
Training:
Accuracies by groups:
0, 0  acc: 11170 / 14537 =  76.838
0, 1  acc:  5963 /  6617 =  90.116
1, 0  acc: 132235 / 133014 =  99.414
1, 1  acc:  8344 /  8602 =  97.001
--------------------------------------
Average acc: 157712 / 162770 =  96.893
Robust  acc: 11170 / 14537 =  76.838
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6284 /  8535 =  73.626
0, 1  acc:  7337 /  8276 =  88.654
1, 0  acc:  2850 /  2874 =  99.165
1, 1  acc:   171 /   182 =  93.956
------------------------------------
Average acc: 16642 / 19867 =  83.767
Robust  acc:  6284 /  8535 =  73.626
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 84.901
Robust Acc: 78.468 | Best Acc: 99.153
-------------------------------------
Training, Epoch 14:
Accuracies by groups:
0, 0  acc:  7664 /  9767 =  78.468
0, 1  acc:  6662 /  7535 =  88.414
1, 0  acc:  2459 /  2480 =  99.153
1, 1  acc:   163 /   180 =  90.556
------------------------------------
Average acc: 16948 / 19962 =  84.901
Robust  acc:  7664 /  9767 =  78.468
------------------------------------
Accuracies by groups:
0, 0  acc:  7664 /  9767 =  78.468
0, 1  acc:  6662 /  7535 =  88.414
1, 0  acc:  2459 /  2480 =  99.153
1, 1  acc:   163 /   180 =  90.556
------------------------------------
Average acc: 16948 / 19962 =  84.901
Robust  acc:  7664 /  9767 =  78.468
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7664 /  9767 =  78.468
0, 1  acc:  6662 /  7535 =  88.414
1, 0  acc:  2459 /  2480 =  99.153
1, 1  acc:   163 /   180 =  90.556
------------------------------------
Average acc: 16948 / 19962 =  84.901
Robust  acc:  7664 /  9767 =  78.468
------------------------------------
Epoch:  16 | Train Loss: 0.001 | Train Acc: 96.464 | Val Loss: 0.002 | Val Acc: 94.438
Training:
Accuracies by groups:
0, 0  acc: 10685 / 14316 =  74.637
0, 1  acc:  5802 /  6579 =  88.190
1, 0  acc: 132255 / 133276 =  99.234
1, 1  acc:  8273 /  8599 =  96.209
--------------------------------------
Average acc: 157015 / 162770 =  96.464
Robust  acc: 10685 / 14316 =  74.637
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7975 /  8535 =  93.439
0, 1  acc:  8100 /  8276 =  97.873
1, 0  acc:  2552 /  2874 =  88.796
1, 1  acc:   135 /   182 =  74.176
------------------------------------
Average acc: 18762 / 19867 =  94.438
Robust  acc:   135 /   182 =  74.176
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 94.700
Robust Acc: 65.000 | Best Acc: 97.664
-------------------------------------
Training, Epoch 15:
Accuracies by groups:
0, 0  acc:  9224 /  9767 =  94.440
0, 1  acc:  7359 /  7535 =  97.664
1, 0  acc:  2204 /  2480 =  88.871
1, 1  acc:   117 /   180 =  65.000
------------------------------------
Average acc: 18904 / 19962 =  94.700
Robust  acc:   117 /   180 =  65.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9224 /  9767 =  94.440
0, 1  acc:  7359 /  7535 =  97.664
1, 0  acc:  2204 /  2480 =  88.871
1, 1  acc:   117 /   180 =  65.000
------------------------------------
Average acc: 18904 / 19962 =  94.700
Robust  acc:   117 /   180 =  65.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9224 /  9767 =  94.440
0, 1  acc:  7359 /  7535 =  97.664
1, 0  acc:  2204 /  2480 =  88.871
1, 1  acc:   117 /   180 =  65.000
------------------------------------
Average acc: 18904 / 19962 =  94.700
Robust  acc:   117 /   180 =  65.000
------------------------------------
Epoch:  17 | Train Loss: 0.001 | Train Acc: 96.026 | Val Loss: 0.003 | Val Acc: 84.235
Training:
Accuracies by groups:
0, 0  acc: 10616 / 14638 =  72.524
0, 1  acc:  5700 /  6657 =  85.624
1, 0  acc: 131549 / 132671 =  99.154
1, 1  acc:  8436 /  8804 =  95.820
--------------------------------------
Average acc: 156301 / 162770 =  96.026
Robust  acc: 10616 / 14638 =  72.524
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6315 /  8535 =  73.989
0, 1  acc:  7404 /  8276 =  89.464
1, 0  acc:  2846 /  2874 =  99.026
1, 1  acc:   170 /   182 =  93.407
------------------------------------
Average acc: 16735 / 19867 =  84.235
Robust  acc:  6315 /  8535 =  73.989
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 85.307
Robust Acc: 78.274 | Best Acc: 98.750
-------------------------------------
Training, Epoch 16:
Accuracies by groups:
0, 0  acc:  7645 /  9767 =  78.274
0, 1  acc:  6776 /  7535 =  89.927
1, 0  acc:  2449 /  2480 =  98.750
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17029 / 19962 =  85.307
Robust  acc:  7645 /  9767 =  78.274
------------------------------------
Accuracies by groups:
0, 0  acc:  7645 /  9767 =  78.274
0, 1  acc:  6776 /  7535 =  89.927
1, 0  acc:  2449 /  2480 =  98.750
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17029 / 19962 =  85.307
Robust  acc:  7645 /  9767 =  78.274
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7645 /  9767 =  78.274
0, 1  acc:  6776 /  7535 =  89.927
1, 0  acc:  2449 /  2480 =  98.750
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17029 / 19962 =  85.307
Robust  acc:  7645 /  9767 =  78.274
------------------------------------
Epoch:  18 | Train Loss: 0.001 | Train Acc: 95.628 | Val Loss: 0.003 | Val Acc: 87.718
Training:
Accuracies by groups:
0, 0  acc: 10269 / 14608 =  70.297
0, 1  acc:  5583 /  6632 =  84.183
1, 0  acc: 131675 / 132969 =  99.027
1, 1  acc:  8127 /  8561 =  94.930
--------------------------------------
Average acc: 155654 / 162770 =  95.628
Robust  acc: 10269 / 14608 =  70.297
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6670 /  8535 =  78.149
0, 1  acc:  7753 /  8276 =  93.681
1, 0  acc:  2836 /  2874 =  98.678
1, 1  acc:   168 /   182 =  92.308
------------------------------------
Average acc: 17427 / 19867 =  87.718
Robust  acc:  6670 /  8535 =  78.149
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 88.578
Robust Acc: 82.359 | Best Acc: 98.548
-------------------------------------
Training, Epoch 17:
Accuracies by groups:
0, 0  acc:  8044 /  9767 =  82.359
0, 1  acc:  7045 /  7535 =  93.497
1, 0  acc:  2444 /  2480 =  98.548
1, 1  acc:   149 /   180 =  82.778
------------------------------------
Average acc: 17682 / 19962 =  88.578
Robust  acc:  8044 /  9767 =  82.359
------------------------------------
Accuracies by groups:
0, 0  acc:  8044 /  9767 =  82.359
0, 1  acc:  7045 /  7535 =  93.497
1, 0  acc:  2444 /  2480 =  98.548
1, 1  acc:   149 /   180 =  82.778
------------------------------------
Average acc: 17682 / 19962 =  88.578
Robust  acc:  8044 /  9767 =  82.359
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8044 /  9767 =  82.359
0, 1  acc:  7045 /  7535 =  93.497
1, 0  acc:  2444 /  2480 =  98.548
1, 1  acc:   149 /   180 =  82.778
------------------------------------
Average acc: 17682 / 19962 =  88.578
Robust  acc:  8044 /  9767 =  82.359
------------------------------------
Epoch:  19 | Train Loss: 0.001 | Train Acc: 95.177 | Val Loss: 0.002 | Val Acc: 90.804
Training:
Accuracies by groups:
0, 0  acc:  9916 / 14683 =  67.534
0, 1  acc:  5134 /  6319 =  81.247
1, 0  acc: 131761 / 133187 =  98.929
1, 1  acc:  8109 /  8581 =  94.499
--------------------------------------
Average acc: 154920 / 162770 =  95.177
Robust  acc:  9916 / 14683 =  67.534
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7353 /  8535 =  86.151
0, 1  acc:  7745 /  8276 =  93.584
1, 0  acc:  2783 /  2874 =  96.834
1, 1  acc:   159 /   182 =  87.363
------------------------------------
Average acc: 18040 / 19867 =  90.804
Robust  acc:  7353 /  8535 =  86.151
------------------------------------
New max robust acc: 86.15114235500879
debias model - Saving best checkpoint at epoch 18
replace: True
-> Updating checkpoint debias-wga-best_seed11.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed11.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.905
Robust Acc: 83.889 | Best Acc: 96.331
-------------------------------------
Training, Epoch 18:
Accuracies by groups:
0, 0  acc:  8737 /  9767 =  89.454
0, 1  acc:  7069 /  7535 =  93.816
1, 0  acc:  2389 /  2480 =  96.331
1, 1  acc:   151 /   180 =  83.889
------------------------------------
Average acc: 18346 / 19962 =  91.905
Robust  acc:   151 /   180 =  83.889
------------------------------------
Accuracies by groups:
0, 0  acc:  8737 /  9767 =  89.454
0, 1  acc:  7069 /  7535 =  93.816
1, 0  acc:  2389 /  2480 =  96.331
1, 1  acc:   151 /   180 =  83.889
------------------------------------
Average acc: 18346 / 19962 =  91.905
Robust  acc:   151 /   180 =  83.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8737 /  9767 =  89.454
0, 1  acc:  7069 /  7535 =  93.816
1, 0  acc:  2389 /  2480 =  96.331
1, 1  acc:   151 /   180 =  83.889
------------------------------------
Average acc: 18346 / 19962 =  91.905
Robust  acc:   151 /   180 =  83.889
------------------------------------
Epoch:  20 | Train Loss: 0.001 | Train Acc: 95.046 | Val Loss: 0.004 | Val Acc: 80.083
Training:
Accuracies by groups:
0, 0  acc:  9708 / 14602 =  66.484
0, 1  acc:  5301 /  6665 =  79.535
1, 0  acc: 131472 / 132811 =  98.992
1, 1  acc:  8226 /  8692 =  94.639
--------------------------------------
Average acc: 154707 / 162770 =  95.046
Robust  acc:  9708 / 14602 =  66.484
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6055 /  8535 =  70.943
0, 1  acc:  6818 /  8276 =  82.383
1, 0  acc:  2857 /  2874 =  99.408
1, 1  acc:   180 /   182 =  98.901
------------------------------------
Average acc: 15910 / 19867 =  80.083
Robust  acc:  6055 /  8535 =  70.943
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 81.936
Robust Acc: 76.124 | Best Acc: 99.395
-------------------------------------
Training, Epoch 19:
Accuracies by groups:
0, 0  acc:  7435 /  9767 =  76.124
0, 1  acc:  6285 /  7535 =  83.411
1, 0  acc:  2465 /  2480 =  99.395
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 16356 / 19962 =  81.936
Robust  acc:  7435 /  9767 =  76.124
------------------------------------
Accuracies by groups:
0, 0  acc:  7435 /  9767 =  76.124
0, 1  acc:  6285 /  7535 =  83.411
1, 0  acc:  2465 /  2480 =  99.395
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 16356 / 19962 =  81.936
Robust  acc:  7435 /  9767 =  76.124
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7435 /  9767 =  76.124
0, 1  acc:  6285 /  7535 =  83.411
1, 0  acc:  2465 /  2480 =  99.395
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 16356 / 19962 =  81.936
Robust  acc:  7435 /  9767 =  76.124
------------------------------------
Epoch:  21 | Train Loss: 0.001 | Train Acc: 94.949 | Val Loss: 0.002 | Val Acc: 92.274
Training:
Accuracies by groups:
0, 0  acc:  9444 / 14415 =  65.515
0, 1  acc:  4994 /  6490 =  76.949
1, 0  acc: 131935 / 133295 =  98.980
1, 1  acc:  8176 /  8570 =  95.403
--------------------------------------
Average acc: 154549 / 162770 =  94.949
Robust  acc:  9444 / 14415 =  65.515
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7611 /  8535 =  89.174
0, 1  acc:  7903 /  8276 =  95.493
1, 0  acc:  2668 /  2874 =  92.832
1, 1  acc:   150 /   182 =  82.418
------------------------------------
Average acc: 18332 / 19867 =  92.274
Robust  acc:   150 /   182 =  82.418
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.846
Robust Acc: 72.222 | Best Acc: 95.275
-------------------------------------
Training, Epoch 20:
Accuracies by groups:
0, 0  acc:  8940 /  9767 =  91.533
0, 1  acc:  7179 /  7535 =  95.275
1, 0  acc:  2285 /  2480 =  92.137
1, 1  acc:   130 /   180 =  72.222
------------------------------------
Average acc: 18534 / 19962 =  92.846
Robust  acc:   130 /   180 =  72.222
------------------------------------
Accuracies by groups:
0, 0  acc:  8940 /  9767 =  91.533
0, 1  acc:  7179 /  7535 =  95.275
1, 0  acc:  2285 /  2480 =  92.137
1, 1  acc:   130 /   180 =  72.222
------------------------------------
Average acc: 18534 / 19962 =  92.846
Robust  acc:   130 /   180 =  72.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8940 /  9767 =  91.533
0, 1  acc:  7179 /  7535 =  95.275
1, 0  acc:  2285 /  2480 =  92.137
1, 1  acc:   130 /   180 =  72.222
------------------------------------
Average acc: 18534 / 19962 =  92.846
Robust  acc:   130 /   180 =  72.222
------------------------------------
Epoch:  22 | Train Loss: 0.002 | Train Acc: 94.681 | Val Loss: 0.002 | Val Acc: 94.201
Training:
Accuracies by groups:
0, 0  acc:  9451 / 14732 =  64.153
0, 1  acc:  4963 /  6590 =  75.311
1, 0  acc: 131654 / 132976 =  99.006
1, 1  acc:  8044 /  8472 =  94.948
--------------------------------------
Average acc: 154112 / 162770 =  94.681
Robust  acc:  9451 / 14732 =  64.153
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8266 /  8535 =  96.848
0, 1  acc:  8189 /  8276 =  98.949
1, 0  acc:  2179 /  2874 =  75.818
1, 1  acc:    81 /   182 =  44.505
------------------------------------
Average acc: 18715 / 19867 =  94.201
Robust  acc:    81 /   182 =  44.505
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 94.515
Robust Acc: 38.889 | Best Acc: 98.952
-------------------------------------
Training, Epoch 21:
Accuracies by groups:
0, 0  acc:  9534 /  9767 =  97.614
0, 1  acc:  7456 /  7535 =  98.952
1, 0  acc:  1807 /  2480 =  72.863
1, 1  acc:    70 /   180 =  38.889
------------------------------------
Average acc: 18867 / 19962 =  94.515
Robust  acc:    70 /   180 =  38.889
------------------------------------
Accuracies by groups:
0, 0  acc:  9534 /  9767 =  97.614
0, 1  acc:  7456 /  7535 =  98.952
1, 0  acc:  1807 /  2480 =  72.863
1, 1  acc:    70 /   180 =  38.889
------------------------------------
Average acc: 18867 / 19962 =  94.515
Robust  acc:    70 /   180 =  38.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9534 /  9767 =  97.614
0, 1  acc:  7456 /  7535 =  98.952
1, 0  acc:  1807 /  2480 =  72.863
1, 1  acc:    70 /   180 =  38.889
------------------------------------
Average acc: 18867 / 19962 =  94.515
Robust  acc:    70 /   180 =  38.889
------------------------------------
Epoch:  23 | Train Loss: 0.002 | Train Acc: 94.501 | Val Loss: 0.004 | Val Acc: 84.663
Training:
Accuracies by groups:
0, 0  acc:  9127 / 14604 =  62.497
0, 1  acc:  4816 /  6571 =  73.292
1, 0  acc: 131594 / 132928 =  98.996
1, 1  acc:  8282 /  8667 =  95.558
--------------------------------------
Average acc: 153819 / 162770 =  94.501
Robust  acc:  9127 / 14604 =  62.497
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6589 /  8535 =  77.200
0, 1  acc:  7235 /  8276 =  87.421
1, 0  acc:  2822 /  2874 =  98.191
1, 1  acc:   174 /   182 =  95.604
------------------------------------
Average acc: 16820 / 19867 =  84.663
Robust  acc:  6589 /  8535 =  77.200
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 85.993
Robust Acc: 82.257 | Best Acc: 97.742
-------------------------------------
Training, Epoch 22:
Accuracies by groups:
0, 0  acc:  8034 /  9767 =  82.257
0, 1  acc:  6542 /  7535 =  86.821
1, 0  acc:  2424 /  2480 =  97.742
1, 1  acc:   166 /   180 =  92.222
------------------------------------
Average acc: 17166 / 19962 =  85.993
Robust  acc:  8034 /  9767 =  82.257
------------------------------------
Accuracies by groups:
0, 0  acc:  8034 /  9767 =  82.257
0, 1  acc:  6542 /  7535 =  86.821
1, 0  acc:  2424 /  2480 =  97.742
1, 1  acc:   166 /   180 =  92.222
------------------------------------
Average acc: 17166 / 19962 =  85.993
Robust  acc:  8034 /  9767 =  82.257
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8034 /  9767 =  82.257
0, 1  acc:  6542 /  7535 =  86.821
1, 0  acc:  2424 /  2480 =  97.742
1, 1  acc:   166 /   180 =  92.222
------------------------------------
Average acc: 17166 / 19962 =  85.993
Robust  acc:  8034 /  9767 =  82.257
------------------------------------
Epoch:  24 | Train Loss: 0.002 | Train Acc: 94.532 | Val Loss: 0.004 | Val Acc: 81.547
Training:
Accuracies by groups:
0, 0  acc:  9012 / 14561 =  61.891
0, 1  acc:  4938 /  6670 =  74.033
1, 0  acc: 131741 / 132979 =  99.069
1, 1  acc:  8179 /  8560 =  95.549
--------------------------------------
Average acc: 153870 / 162770 =  94.532
Robust  acc:  9012 / 14561 =  61.891
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6238 /  8535 =  73.087
0, 1  acc:  6943 /  8276 =  83.893
1, 0  acc:  2845 /  2874 =  98.991
1, 1  acc:   175 /   182 =  96.154
------------------------------------
Average acc: 16201 / 19867 =  81.547
Robust  acc:  6238 /  8535 =  73.087
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 83.328
Robust Acc: 78.827 | Best Acc: 98.710
-------------------------------------
Training, Epoch 23:
Accuracies by groups:
0, 0  acc:  7699 /  9767 =  78.827
0, 1  acc:  6320 /  7535 =  83.875
1, 0  acc:  2448 /  2480 =  98.710
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16634 / 19962 =  83.328
Robust  acc:  7699 /  9767 =  78.827
------------------------------------
Accuracies by groups:
0, 0  acc:  7699 /  9767 =  78.827
0, 1  acc:  6320 /  7535 =  83.875
1, 0  acc:  2448 /  2480 =  98.710
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16634 / 19962 =  83.328
Robust  acc:  7699 /  9767 =  78.827
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7699 /  9767 =  78.827
0, 1  acc:  6320 /  7535 =  83.875
1, 0  acc:  2448 /  2480 =  98.710
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16634 / 19962 =  83.328
Robust  acc:  7699 /  9767 =  78.827
------------------------------------
Epoch:  25 | Train Loss: 0.002 | Train Acc: 94.349 | Val Loss: 0.003 | Val Acc: 91.217
Training:
Accuracies by groups:
0, 0  acc:  8878 / 14488 =  61.278
0, 1  acc:  4629 /  6514 =  71.062
1, 0  acc: 131878 / 133200 =  99.008
1, 1  acc:  8187 /  8568 =  95.553
--------------------------------------
Average acc: 153572 / 162770 =  94.349
Robust  acc:  8878 / 14488 =  61.278
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7500 /  8535 =  87.873
0, 1  acc:  7745 /  8276 =  93.584
1, 0  acc:  2724 /  2874 =  94.781
1, 1  acc:   153 /   182 =  84.066
------------------------------------
Average acc: 18122 / 19867 =  91.217
Robust  acc:   153 /   182 =  84.066
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 92.355
Robust Acc: 81.667 | Best Acc: 94.032
-------------------------------------
Training, Epoch 24:
Accuracies by groups:
0, 0  acc:  8875 /  9767 =  90.867
0, 1  acc:  7082 /  7535 =  93.988
1, 0  acc:  2332 /  2480 =  94.032
1, 1  acc:   147 /   180 =  81.667
------------------------------------
Average acc: 18436 / 19962 =  92.355
Robust  acc:   147 /   180 =  81.667
------------------------------------
Accuracies by groups:
0, 0  acc:  8875 /  9767 =  90.867
0, 1  acc:  7082 /  7535 =  93.988
1, 0  acc:  2332 /  2480 =  94.032
1, 1  acc:   147 /   180 =  81.667
------------------------------------
Average acc: 18436 / 19962 =  92.355
Robust  acc:   147 /   180 =  81.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8875 /  9767 =  90.867
0, 1  acc:  7082 /  7535 =  93.988
1, 0  acc:  2332 /  2480 =  94.032
1, 1  acc:   147 /   180 =  81.667
------------------------------------
Average acc: 18436 / 19962 =  92.355
Robust  acc:   147 /   180 =  81.667
------------------------------------
Epoch:  26 | Train Loss: 0.002 | Train Acc: 94.324 | Val Loss: 0.005 | Val Acc: 77.203
Training:
Accuracies by groups:
0, 0  acc:  9044 / 14700 =  61.524
0, 1  acc:  4483 /  6433 =  69.688
1, 0  acc: 131731 / 133022 =  99.029
1, 1  acc:  8273 /  8615 =  96.030
--------------------------------------
Average acc: 153531 / 162770 =  94.324
Robust  acc:  9044 / 14700 =  61.524
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6088 /  8535 =  71.330
0, 1  acc:  6232 /  8276 =  75.302
1, 0  acc:  2838 /  2874 =  98.747
1, 1  acc:   180 /   182 =  98.901
------------------------------------
Average acc: 15338 / 19867 =  77.203
Robust  acc:  6088 /  8535 =  71.330
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 79.261
Robust Acc: 76.216 | Best Acc: 98.710
-------------------------------------
Training, Epoch 25:
Accuracies by groups:
0, 0  acc:  7444 /  9767 =  76.216
0, 1  acc:  5761 /  7535 =  76.457
1, 0  acc:  2448 /  2480 =  98.710
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 15822 / 19962 =  79.261
Robust  acc:  7444 /  9767 =  76.216
------------------------------------
Accuracies by groups:
0, 0  acc:  7444 /  9767 =  76.216
0, 1  acc:  5761 /  7535 =  76.457
1, 0  acc:  2448 /  2480 =  98.710
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 15822 / 19962 =  79.261
Robust  acc:  7444 /  9767 =  76.216
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7444 /  9767 =  76.216
0, 1  acc:  5761 /  7535 =  76.457
1, 0  acc:  2448 /  2480 =  98.710
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 15822 / 19962 =  79.261
Robust  acc:  7444 /  9767 =  76.216
------------------------------------
Epoch:  27 | Train Loss: 0.002 | Train Acc: 94.124 | Val Loss: 0.005 | Val Acc: 72.905
Training:
Accuracies by groups:
0, 0  acc:  8691 / 14673 =  59.231
0, 1  acc:  4449 /  6564 =  67.779
1, 0  acc: 131719 / 132868 =  99.135
1, 1  acc:  8347 /  8665 =  96.330
--------------------------------------
Average acc: 153206 / 162770 =  94.124
Robust  acc:  8691 / 14673 =  59.231
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5506 /  8535 =  64.511
0, 1  acc:  5946 /  8276 =  71.846
1, 0  acc:  2853 /  2874 =  99.269
1, 1  acc:   179 /   182 =  98.352
------------------------------------
Average acc: 14484 / 19867 =  72.905
Robust  acc:  5506 /  8535 =  64.511
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 74.492
Robust Acc: 69.899 | Best Acc: 99.234
-------------------------------------
Training, Epoch 26:
Accuracies by groups:
0, 0  acc:  6827 /  9767 =  69.899
0, 1  acc:  5406 /  7535 =  71.745
1, 0  acc:  2461 /  2480 =  99.234
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 14870 / 19962 =  74.492
Robust  acc:  6827 /  9767 =  69.899
------------------------------------
Accuracies by groups:
0, 0  acc:  6827 /  9767 =  69.899
0, 1  acc:  5406 /  7535 =  71.745
1, 0  acc:  2461 /  2480 =  99.234
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 14870 / 19962 =  74.492
Robust  acc:  6827 /  9767 =  69.899
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6827 /  9767 =  69.899
0, 1  acc:  5406 /  7535 =  71.745
1, 0  acc:  2461 /  2480 =  99.234
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 14870 / 19962 =  74.492
Robust  acc:  6827 /  9767 =  69.899
------------------------------------
Epoch:  28 | Train Loss: 0.002 | Train Acc: 94.076 | Val Loss: 0.004 | Val Acc: 86.898
Training:
Accuracies by groups:
0, 0  acc:  8369 / 14347 =  58.333
0, 1  acc:  4307 /  6556 =  65.696
1, 0  acc: 132270 / 133355 =  99.186
1, 1  acc:  8182 /  8512 =  96.123
--------------------------------------
Average acc: 153128 / 162770 =  94.076
Robust  acc:  8369 / 14347 =  58.333
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7119 /  8535 =  83.409
0, 1  acc:  7222 /  8276 =  87.264
1, 0  acc:  2754 /  2874 =  95.825
1, 1  acc:   169 /   182 =  92.857
------------------------------------
Average acc: 17264 / 19867 =  86.898
Robust  acc:  7119 /  8535 =  83.409
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 87.917
Robust Acc: 86.729 | Best Acc: 94.960
-------------------------------------
Training, Epoch 27:
Accuracies by groups:
0, 0  acc:  8501 /  9767 =  87.038
0, 1  acc:  6535 /  7535 =  86.729
1, 0  acc:  2355 /  2480 =  94.960
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17550 / 19962 =  87.917
Robust  acc:  6535 /  7535 =  86.729
------------------------------------
Accuracies by groups:
0, 0  acc:  8501 /  9767 =  87.038
0, 1  acc:  6535 /  7535 =  86.729
1, 0  acc:  2355 /  2480 =  94.960
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17550 / 19962 =  87.917
Robust  acc:  6535 /  7535 =  86.729
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8501 /  9767 =  87.038
0, 1  acc:  6535 /  7535 =  86.729
1, 0  acc:  2355 /  2480 =  94.960
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17550 / 19962 =  87.917
Robust  acc:  6535 /  7535 =  86.729
------------------------------------
Epoch:  29 | Train Loss: 0.002 | Train Acc: 94.019 | Val Loss: 0.006 | Val Acc: 70.856
Training:
Accuracies by groups:
0, 0  acc:  8480 / 14583 =  58.150
0, 1  acc:  4288 /  6526 =  65.706
1, 0  acc: 132008 / 133066 =  99.205
1, 1  acc:  8258 /  8595 =  96.079
--------------------------------------
Average acc: 153034 / 162770 =  94.019
Robust  acc:  8480 / 14583 =  58.150
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5019 /  8535 =  58.805
0, 1  acc:  6017 /  8276 =  72.704
1, 0  acc:  2862 /  2874 =  99.582
1, 1  acc:   179 /   182 =  98.352
------------------------------------
Average acc: 14077 / 19867 =  70.856
Robust  acc:  5019 /  8535 =  58.805
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 72.808
Robust Acc: 65.035 | Best Acc: 99.435
-------------------------------------
Training, Epoch 28:
Accuracies by groups:
0, 0  acc:  6352 /  9767 =  65.035
0, 1  acc:  5541 /  7535 =  73.537
1, 0  acc:  2466 /  2480 =  99.435
1, 1  acc:   175 /   180 =  97.222
------------------------------------
Average acc: 14534 / 19962 =  72.808
Robust  acc:  6352 /  9767 =  65.035
------------------------------------
Accuracies by groups:
0, 0  acc:  6352 /  9767 =  65.035
0, 1  acc:  5541 /  7535 =  73.537
1, 0  acc:  2466 /  2480 =  99.435
1, 1  acc:   175 /   180 =  97.222
------------------------------------
Average acc: 14534 / 19962 =  72.808
Robust  acc:  6352 /  9767 =  65.035
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6352 /  9767 =  65.035
0, 1  acc:  5541 /  7535 =  73.537
1, 0  acc:  2466 /  2480 =  99.435
1, 1  acc:   175 /   180 =  97.222
------------------------------------
Average acc: 14534 / 19962 =  72.808
Robust  acc:  6352 /  9767 =  65.035
------------------------------------
Epoch:  30 | Train Loss: 0.002 | Train Acc: 93.964 | Val Loss: 0.006 | Val Acc: 62.184
Training:
Accuracies by groups:
0, 0  acc:  8287 / 14463 =  57.298
0, 1  acc:  4305 /  6614 =  65.089
1, 0  acc: 132132 / 133154 =  99.232
1, 1  acc:  8221 /  8539 =  96.276
--------------------------------------
Average acc: 152945 / 162770 =  93.964
Robust  acc:  8287 / 14463 =  57.298
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4546 /  8535 =  53.263
0, 1  acc:  4759 /  8276 =  57.504
1, 0  acc:  2867 /  2874 =  99.756
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc: 12354 / 19867 =  62.184
Robust  acc:  4546 /  8535 =  53.263
------------------------------------
-------------------------------------------
Avg Test Loss: 0.006 | Avg Test Acc: 62.769
Robust Acc: 55.647 | Best Acc: 99.718
-------------------------------------
Training, Epoch 29:
Accuracies by groups:
0, 0  acc:  5687 /  9767 =  58.227
0, 1  acc:  4193 /  7535 =  55.647
1, 0  acc:  2473 /  2480 =  99.718
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 12530 / 19962 =  62.769
Robust  acc:  4193 /  7535 =  55.647
------------------------------------
Accuracies by groups:
0, 0  acc:  5687 /  9767 =  58.227
0, 1  acc:  4193 /  7535 =  55.647
1, 0  acc:  2473 /  2480 =  99.718
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 12530 / 19962 =  62.769
Robust  acc:  4193 /  7535 =  55.647
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  5687 /  9767 =  58.227
0, 1  acc:  4193 /  7535 =  55.647
1, 0  acc:  2473 /  2480 =  99.718
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 12530 / 19962 =  62.769
Robust  acc:  4193 /  7535 =  55.647
------------------------------------
Epoch:  31 | Train Loss: 0.002 | Train Acc: 93.824 | Val Loss: 0.004 | Val Acc: 83.636
Training:
Accuracies by groups:
0, 0  acc:  8322 / 14669 =  56.732
0, 1  acc:  4328 /  6688 =  64.713
1, 0  acc: 131735 / 132753 =  99.233
1, 1  acc:  8333 /  8660 =  96.224
--------------------------------------
Average acc: 152718 / 162770 =  93.824
Robust  acc:  8322 / 14669 =  56.732
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6533 /  8535 =  76.544
0, 1  acc:  7067 /  8276 =  85.391
1, 0  acc:  2837 /  2874 =  98.713
1, 1  acc:   179 /   182 =  98.352
------------------------------------
Average acc: 16616 / 19867 =  83.636
Robust  acc:  6533 /  8535 =  76.544
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 85.037
Robust Acc: 81.458 | Best Acc: 98.065
-------------------------------------
Training, Epoch 30:
Accuracies by groups:
0, 0  acc:  7956 /  9767 =  81.458
0, 1  acc:  6418 /  7535 =  85.176
1, 0  acc:  2432 /  2480 =  98.065
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 16975 / 19962 =  85.037
Robust  acc:  7956 /  9767 =  81.458
------------------------------------
Accuracies by groups:
0, 0  acc:  7956 /  9767 =  81.458
0, 1  acc:  6418 /  7535 =  85.176
1, 0  acc:  2432 /  2480 =  98.065
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 16975 / 19962 =  85.037
Robust  acc:  7956 /  9767 =  81.458
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7956 /  9767 =  81.458
0, 1  acc:  6418 /  7535 =  85.176
1, 0  acc:  2432 /  2480 =  98.065
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 16975 / 19962 =  85.037
Robust  acc:  7956 /  9767 =  81.458
------------------------------------
Epoch:  32 | Train Loss: 0.002 | Train Acc: 93.894 | Val Loss: 0.005 | Val Acc: 76.841
Training:
Accuracies by groups:
0, 0  acc:  8360 / 14670 =  56.987
0, 1  acc:  4269 /  6591 =  64.770
1, 0  acc: 132012 / 132979 =  99.273
1, 1  acc:  8191 /  8530 =  96.026
--------------------------------------
Average acc: 152832 / 162770 =  93.894
Robust  acc:  8360 / 14670 =  56.987
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5904 /  8535 =  69.174
0, 1  acc:  6346 /  8276 =  76.680
1, 0  acc:  2841 /  2874 =  98.852
1, 1  acc:   175 /   182 =  96.154
------------------------------------
Average acc: 15266 / 19867 =  76.841
Robust  acc:  5904 /  8535 =  69.174
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 78.534
Robust Acc: 74.752 | Best Acc: 98.710
-------------------------------------
Training, Epoch 31:
Accuracies by groups:
0, 0  acc:  7301 /  9767 =  74.752
0, 1  acc:  5759 /  7535 =  76.430
1, 0  acc:  2448 /  2480 =  98.710
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 15677 / 19962 =  78.534
Robust  acc:  7301 /  9767 =  74.752
------------------------------------
Accuracies by groups:
0, 0  acc:  7301 /  9767 =  74.752
0, 1  acc:  5759 /  7535 =  76.430
1, 0  acc:  2448 /  2480 =  98.710
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 15677 / 19962 =  78.534
Robust  acc:  7301 /  9767 =  74.752
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7301 /  9767 =  74.752
0, 1  acc:  5759 /  7535 =  76.430
1, 0  acc:  2448 /  2480 =  98.710
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 15677 / 19962 =  78.534
Robust  acc:  7301 /  9767 =  74.752
------------------------------------
Epoch:  33 | Train Loss: 0.002 | Train Acc: 93.980 | Val Loss: 0.004 | Val Acc: 88.916
Training:
Accuracies by groups:
0, 0  acc:  8360 / 14520 =  57.576
0, 1  acc:  4360 /  6698 =  65.094
1, 0  acc: 131950 / 132921 =  99.269
1, 1  acc:  8302 /  8631 =  96.188
--------------------------------------
Average acc: 152972 / 162770 =  93.980
Robust  acc:  8360 / 14520 =  57.576
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7123 /  8535 =  83.456
0, 1  acc:  7597 /  8276 =  91.796
1, 0  acc:  2780 /  2874 =  96.729
1, 1  acc:   165 /   182 =  90.659
------------------------------------
Average acc: 17665 / 19867 =  88.916
Robust  acc:  7123 /  8535 =  83.456
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 90.046
Robust Acc: 86.667 | Best Acc: 96.331
-------------------------------------
Training, Epoch 32:
Accuracies by groups:
0, 0  acc:  8494 /  9767 =  86.966
0, 1  acc:  6936 /  7535 =  92.050
1, 0  acc:  2389 /  2480 =  96.331
1, 1  acc:   156 /   180 =  86.667
------------------------------------
Average acc: 17975 / 19962 =  90.046
Robust  acc:   156 /   180 =  86.667
------------------------------------
Accuracies by groups:
0, 0  acc:  8494 /  9767 =  86.966
0, 1  acc:  6936 /  7535 =  92.050
1, 0  acc:  2389 /  2480 =  96.331
1, 1  acc:   156 /   180 =  86.667
------------------------------------
Average acc: 17975 / 19962 =  90.046
Robust  acc:   156 /   180 =  86.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8494 /  9767 =  86.966
0, 1  acc:  6936 /  7535 =  92.050
1, 0  acc:  2389 /  2480 =  96.331
1, 1  acc:   156 /   180 =  86.667
------------------------------------
Average acc: 17975 / 19962 =  90.046
Robust  acc:   156 /   180 =  86.667
------------------------------------
Epoch:  34 | Train Loss: 0.002 | Train Acc: 94.100 | Val Loss: 0.006 | Val Acc: 66.326
Training:
Accuracies by groups:
0, 0  acc:  8252 / 14391 =  57.341
0, 1  acc:  4357 /  6619 =  65.826
1, 0  acc: 132161 / 133072 =  99.315
1, 1  acc:  8397 /  8688 =  96.651
--------------------------------------
Average acc: 153167 / 162770 =  94.100
Robust  acc:  8252 / 14391 =  57.341
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4750 /  8535 =  55.653
0, 1  acc:  5378 /  8276 =  64.983
1, 0  acc:  2867 /  2874 =  99.756
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc: 13177 / 19867 =  66.326
Robust  acc:  4750 /  8535 =  55.653
------------------------------------
-------------------------------------------
Avg Test Loss: 0.006 | Avg Test Acc: 67.819
Robust Acc: 61.564 | Best Acc: 99.677
-------------------------------------
Training, Epoch 33:
Accuracies by groups:
0, 0  acc:  6013 /  9767 =  61.564
0, 1  acc:  4875 /  7535 =  64.698
1, 0  acc:  2472 /  2480 =  99.677
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 13538 / 19962 =  67.819
Robust  acc:  6013 /  9767 =  61.564
------------------------------------
Accuracies by groups:
0, 0  acc:  6013 /  9767 =  61.564
0, 1  acc:  4875 /  7535 =  64.698
1, 0  acc:  2472 /  2480 =  99.677
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 13538 / 19962 =  67.819
Robust  acc:  6013 /  9767 =  61.564
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6013 /  9767 =  61.564
0, 1  acc:  4875 /  7535 =  64.698
1, 0  acc:  2472 /  2480 =  99.677
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 13538 / 19962 =  67.819
Robust  acc:  6013 /  9767 =  61.564
------------------------------------
Epoch:  35 | Train Loss: 0.002 | Train Acc: 93.979 | Val Loss: 0.007 | Val Acc: 58.167
Training:
Accuracies by groups:
0, 0  acc:  8132 / 14296 =  56.883
0, 1  acc:  4318 /  6665 =  64.786
1, 0  acc: 132025 / 133023 =  99.250
1, 1  acc:  8495 /  8786 =  96.688
--------------------------------------
Average acc: 152970 / 162770 =  93.979
Robust  acc:  8132 / 14296 =  56.883
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4287 /  8535 =  50.228
0, 1  acc:  4219 /  8276 =  50.979
1, 0  acc:  2869 /  2874 =  99.826
1, 1  acc:   181 /   182 =  99.451
------------------------------------
Average acc: 11556 / 19867 =  58.167
Robust  acc:  4287 /  8535 =  50.228
------------------------------------
-------------------------------------------
Avg Test Loss: 0.007 | Avg Test Acc: 59.733
Robust Acc: 50.936 | Best Acc: 99.758
-------------------------------------
Training, Epoch 34:
Accuracies by groups:
0, 0  acc:  5434 /  9767 =  55.636
0, 1  acc:  3838 /  7535 =  50.936
1, 0  acc:  2474 /  2480 =  99.758
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 11924 / 19962 =  59.733
Robust  acc:  3838 /  7535 =  50.936
------------------------------------
Accuracies by groups:
0, 0  acc:  5434 /  9767 =  55.636
0, 1  acc:  3838 /  7535 =  50.936
1, 0  acc:  2474 /  2480 =  99.758
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 11924 / 19962 =  59.733
Robust  acc:  3838 /  7535 =  50.936
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  5434 /  9767 =  55.636
0, 1  acc:  3838 /  7535 =  50.936
1, 0  acc:  2474 /  2480 =  99.758
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 11924 / 19962 =  59.733
Robust  acc:  3838 /  7535 =  50.936
------------------------------------
Epoch:  36 | Train Loss: 0.002 | Train Acc: 93.918 | Val Loss: 0.005 | Val Acc: 78.210
Training:
Accuracies by groups:
0, 0  acc:  8365 / 14566 =  57.428
0, 1  acc:  4262 /  6602 =  64.556
1, 0  acc: 132027 / 133105 =  99.190
1, 1  acc:  8216 /  8497 =  96.693
--------------------------------------
Average acc: 152870 / 162770 =  93.918
Robust  acc:  8365 / 14566 =  57.428
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6065 /  8535 =  71.060
0, 1  acc:  6449 /  8276 =  77.924
1, 0  acc:  2845 /  2874 =  98.991
1, 1  acc:   179 /   182 =  98.352
------------------------------------
Average acc: 15538 / 19867 =  78.210
Robust  acc:  6065 /  8535 =  71.060
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 79.867
Robust Acc: 76.554 | Best Acc: 98.508
-------------------------------------
Training, Epoch 35:
Accuracies by groups:
0, 0  acc:  7477 /  9767 =  76.554
0, 1  acc:  5850 /  7535 =  77.638
1, 0  acc:  2443 /  2480 =  98.508
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15943 / 19962 =  79.867
Robust  acc:  7477 /  9767 =  76.554
------------------------------------
Accuracies by groups:
0, 0  acc:  7477 /  9767 =  76.554
0, 1  acc:  5850 /  7535 =  77.638
1, 0  acc:  2443 /  2480 =  98.508
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15943 / 19962 =  79.867
Robust  acc:  7477 /  9767 =  76.554
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7477 /  9767 =  76.554
0, 1  acc:  5850 /  7535 =  77.638
1, 0  acc:  2443 /  2480 =  98.508
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15943 / 19962 =  79.867
Robust  acc:  7477 /  9767 =  76.554
------------------------------------
Epoch:  37 | Train Loss: 0.002 | Train Acc: 94.080 | Val Loss: 0.004 | Val Acc: 80.505
Training:
Accuracies by groups:
0, 0  acc:  8390 / 14472 =  57.974
0, 1  acc:  4312 /  6522 =  66.115
1, 0  acc: 132095 / 133140 =  99.215
1, 1  acc:  8337 /  8636 =  96.538
--------------------------------------
Average acc: 153134 / 162770 =  94.080
Robust  acc:  8390 / 14472 =  57.974
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6380 /  8535 =  74.751
0, 1  acc:  6635 /  8276 =  80.172
1, 0  acc:  2814 /  2874 =  97.912
1, 1  acc:   165 /   182 =  90.659
------------------------------------
Average acc: 15994 / 19867 =  80.505
Robust  acc:  6380 /  8535 =  74.751
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 81.520
Robust Acc: 78.872 | Best Acc: 97.379
-------------------------------------
Training, Epoch 36:
Accuracies by groups:
0, 0  acc:  7752 /  9767 =  79.369
0, 1  acc:  5943 /  7535 =  78.872
1, 0  acc:  2415 /  2480 =  97.379
1, 1  acc:   163 /   180 =  90.556
------------------------------------
Average acc: 16273 / 19962 =  81.520
Robust  acc:  5943 /  7535 =  78.872
------------------------------------
Accuracies by groups:
0, 0  acc:  7752 /  9767 =  79.369
0, 1  acc:  5943 /  7535 =  78.872
1, 0  acc:  2415 /  2480 =  97.379
1, 1  acc:   163 /   180 =  90.556
------------------------------------
Average acc: 16273 / 19962 =  81.520
Robust  acc:  5943 /  7535 =  78.872
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7752 /  9767 =  79.369
0, 1  acc:  5943 /  7535 =  78.872
1, 0  acc:  2415 /  2480 =  97.379
1, 1  acc:   163 /   180 =  90.556
------------------------------------
Average acc: 16273 / 19962 =  81.520
Robust  acc:  5943 /  7535 =  78.872
------------------------------------
Epoch:  38 | Train Loss: 0.002 | Train Acc: 94.130 | Val Loss: 0.004 | Val Acc: 84.331
Training:
Accuracies by groups:
0, 0  acc:  8459 / 14514 =  58.282
0, 1  acc:  4319 /  6522 =  66.222
1, 0  acc: 132023 / 133043 =  99.233
1, 1  acc:  8414 /  8691 =  96.813
--------------------------------------
Average acc: 153215 / 162770 =  94.130
Robust  acc:  8459 / 14514 =  58.282
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6665 /  8535 =  78.090
0, 1  acc:  7085 /  8276 =  85.609
1, 0  acc:  2828 /  2874 =  98.399
1, 1  acc:   176 /   182 =  96.703
------------------------------------
Average acc: 16754 / 19867 =  84.331
Robust  acc:  6665 /  8535 =  78.090
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 85.362
Robust Acc: 82.369 | Best Acc: 97.540
-------------------------------------
Training, Epoch 37:
Accuracies by groups:
0, 0  acc:  8045 /  9767 =  82.369
0, 1  acc:  6407 /  7535 =  85.030
1, 0  acc:  2419 /  2480 =  97.540
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 17040 / 19962 =  85.362
Robust  acc:  8045 /  9767 =  82.369
------------------------------------
Accuracies by groups:
0, 0  acc:  8045 /  9767 =  82.369
0, 1  acc:  6407 /  7535 =  85.030
1, 0  acc:  2419 /  2480 =  97.540
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 17040 / 19962 =  85.362
Robust  acc:  8045 /  9767 =  82.369
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8045 /  9767 =  82.369
0, 1  acc:  6407 /  7535 =  85.030
1, 0  acc:  2419 /  2480 =  97.540
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 17040 / 19962 =  85.362
Robust  acc:  8045 /  9767 =  82.369
------------------------------------
Epoch:  39 | Train Loss: 0.002 | Train Acc: 94.138 | Val Loss: 0.005 | Val Acc: 81.673
Training:
Accuracies by groups:
0, 0  acc:  8423 / 14344 =  58.721
0, 1  acc:  4373 /  6607 =  66.187
1, 0  acc: 132269 / 133335 =  99.201
1, 1  acc:  8163 /  8484 =  96.216
--------------------------------------
Average acc: 153228 / 162770 =  94.138
Robust  acc:  8423 / 14344 =  58.721
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6349 /  8535 =  74.388
0, 1  acc:  6865 /  8276 =  82.951
1, 0  acc:  2837 /  2874 =  98.713
1, 1  acc:   175 /   182 =  96.154
------------------------------------
Average acc: 16226 / 19867 =  81.673
Robust  acc:  6349 /  8535 =  74.388
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 83.173
Robust Acc: 79.687 | Best Acc: 98.266
-------------------------------------
Training, Epoch 38:
Accuracies by groups:
0, 0  acc:  7783 /  9767 =  79.687
0, 1  acc:  6216 /  7535 =  82.495
1, 0  acc:  2437 /  2480 =  98.266
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16603 / 19962 =  83.173
Robust  acc:  7783 /  9767 =  79.687
------------------------------------
Accuracies by groups:
0, 0  acc:  7783 /  9767 =  79.687
0, 1  acc:  6216 /  7535 =  82.495
1, 0  acc:  2437 /  2480 =  98.266
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16603 / 19962 =  83.173
Robust  acc:  7783 /  9767 =  79.687
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7783 /  9767 =  79.687
0, 1  acc:  6216 /  7535 =  82.495
1, 0  acc:  2437 /  2480 =  98.266
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16603 / 19962 =  83.173
Robust  acc:  7783 /  9767 =  79.687
------------------------------------
Epoch:  40 | Train Loss: 0.002 | Train Acc: 94.113 | Val Loss: 0.004 | Val Acc: 86.093
Training:
Accuracies by groups:
0, 0  acc:  8395 / 14319 =  58.628
0, 1  acc:  4453 /  6587 =  67.603
1, 0  acc: 131876 / 133046 =  99.121
1, 1  acc:  8464 /  8818 =  95.985
--------------------------------------
Average acc: 153188 / 162770 =  94.113
Robust  acc:  8395 / 14319 =  58.628
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6928 /  8535 =  81.172
0, 1  acc:  7227 /  8276 =  87.325
1, 0  acc:  2785 /  2874 =  96.903
1, 1  acc:   164 /   182 =  90.110
------------------------------------
Average acc: 17104 / 19867 =  86.093
Robust  acc:  6928 /  8535 =  81.172
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 87.196
Robust Acc: 85.226 | Best Acc: 95.887
-------------------------------------
Training, Epoch 39:
Accuracies by groups:
0, 0  acc:  8324 /  9767 =  85.226
0, 1  acc:  6540 /  7535 =  86.795
1, 0  acc:  2378 /  2480 =  95.887
1, 1  acc:   164 /   180 =  91.111
------------------------------------
Average acc: 17406 / 19962 =  87.196
Robust  acc:  8324 /  9767 =  85.226
------------------------------------
Accuracies by groups:
0, 0  acc:  8324 /  9767 =  85.226
0, 1  acc:  6540 /  7535 =  86.795
1, 0  acc:  2378 /  2480 =  95.887
1, 1  acc:   164 /   180 =  91.111
------------------------------------
Average acc: 17406 / 19962 =  87.196
Robust  acc:  8324 /  9767 =  85.226
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8324 /  9767 =  85.226
0, 1  acc:  6540 /  7535 =  86.795
1, 0  acc:  2378 /  2480 =  95.887
1, 1  acc:   164 /   180 =  91.111
------------------------------------
Average acc: 17406 / 19962 =  87.196
Robust  acc:  8324 /  9767 =  85.226
------------------------------------
Epoch:  41 | Train Loss: 0.002 | Train Acc: 94.062 | Val Loss: 0.004 | Val Acc: 89.299
Training:
Accuracies by groups:
0, 0  acc:  8679 / 14635 =  59.303
0, 1  acc:  4397 /  6556 =  67.068
1, 0  acc: 131840 / 133018 =  99.114
1, 1  acc:  8189 /  8561 =  95.655
--------------------------------------
Average acc: 153105 / 162770 =  94.062
Robust  acc:  8679 / 14635 =  59.303
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7252 /  8535 =  84.968
0, 1  acc:  7581 /  8276 =  91.602
1, 0  acc:  2746 /  2874 =  95.546
1, 1  acc:   162 /   182 =  89.011
------------------------------------
Average acc: 17741 / 19867 =  89.299
Robust  acc:  7252 /  8535 =  84.968
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 90.267
Robust Acc: 83.889 | Best Acc: 94.073
-------------------------------------
Training, Epoch 40:
Accuracies by groups:
0, 0  acc:  8595 /  9767 =  88.000
0, 1  acc:  6940 /  7535 =  92.104
1, 0  acc:  2333 /  2480 =  94.073
1, 1  acc:   151 /   180 =  83.889
------------------------------------
Average acc: 18019 / 19962 =  90.267
Robust  acc:   151 /   180 =  83.889
------------------------------------
Accuracies by groups:
0, 0  acc:  8595 /  9767 =  88.000
0, 1  acc:  6940 /  7535 =  92.104
1, 0  acc:  2333 /  2480 =  94.073
1, 1  acc:   151 /   180 =  83.889
------------------------------------
Average acc: 18019 / 19962 =  90.267
Robust  acc:   151 /   180 =  83.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8595 /  9767 =  88.000
0, 1  acc:  6940 /  7535 =  92.104
1, 0  acc:  2333 /  2480 =  94.073
1, 1  acc:   151 /   180 =  83.889
------------------------------------
Average acc: 18019 / 19962 =  90.267
Robust  acc:   151 /   180 =  83.889
------------------------------------
Epoch:  42 | Train Loss: 0.002 | Train Acc: 94.219 | Val Loss: 0.004 | Val Acc: 86.037
Training:
Accuracies by groups:
0, 0  acc:  8457 / 14311 =  59.094
0, 1  acc:  4320 /  6456 =  66.914
1, 0  acc: 132221 / 133310 =  99.183
1, 1  acc:  8362 /  8693 =  96.192
--------------------------------------
Average acc: 153360 / 162770 =  94.219
Robust  acc:  8457 / 14311 =  59.094
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6651 /  8535 =  77.926
0, 1  acc:  7456 /  8276 =  90.092
1, 0  acc:  2816 /  2874 =  97.982
1, 1  acc:   170 /   182 =  93.407
------------------------------------
Average acc: 17093 / 19867 =  86.037
Robust  acc:  6651 /  8535 =  77.926
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 87.206
Robust Acc: 82.123 | Best Acc: 97.540
-------------------------------------
Training, Epoch 41:
Accuracies by groups:
0, 0  acc:  8021 /  9767 =  82.123
0, 1  acc:  6815 /  7535 =  90.445
1, 0  acc:  2419 /  2480 =  97.540
1, 1  acc:   153 /   180 =  85.000
------------------------------------
Average acc: 17408 / 19962 =  87.206
Robust  acc:  8021 /  9767 =  82.123
------------------------------------
Accuracies by groups:
0, 0  acc:  8021 /  9767 =  82.123
0, 1  acc:  6815 /  7535 =  90.445
1, 0  acc:  2419 /  2480 =  97.540
1, 1  acc:   153 /   180 =  85.000
------------------------------------
Average acc: 17408 / 19962 =  87.206
Robust  acc:  8021 /  9767 =  82.123
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8021 /  9767 =  82.123
0, 1  acc:  6815 /  7535 =  90.445
1, 0  acc:  2419 /  2480 =  97.540
1, 1  acc:   153 /   180 =  85.000
------------------------------------
Average acc: 17408 / 19962 =  87.206
Robust  acc:  8021 /  9767 =  82.123
------------------------------------
Epoch:  43 | Train Loss: 0.002 | Train Acc: 94.146 | Val Loss: 0.005 | Val Acc: 78.250
Training:
Accuracies by groups:
0, 0  acc:  8537 / 14388 =  59.334
0, 1  acc:  4425 /  6648 =  66.561
1, 0  acc: 131912 / 133031 =  99.159
1, 1  acc:  8367 /  8703 =  96.139
--------------------------------------
Average acc: 153241 / 162770 =  94.146
Robust  acc:  8537 / 14388 =  59.334
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5930 /  8535 =  69.479
0, 1  acc:  6583 /  8276 =  79.543
1, 0  acc:  2854 /  2874 =  99.304
1, 1  acc:   179 /   182 =  98.352
------------------------------------
Average acc: 15546 / 19867 =  78.250
Robust  acc:  5930 /  8535 =  69.479
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 80.192
Robust Acc: 75.161 | Best Acc: 98.911
-------------------------------------
Training, Epoch 42:
Accuracies by groups:
0, 0  acc:  7341 /  9767 =  75.161
0, 1  acc:  6042 /  7535 =  80.186
1, 0  acc:  2453 /  2480 =  98.911
1, 1  acc:   172 /   180 =  95.556
------------------------------------
Average acc: 16008 / 19962 =  80.192
Robust  acc:  7341 /  9767 =  75.161
------------------------------------
Accuracies by groups:
0, 0  acc:  7341 /  9767 =  75.161
0, 1  acc:  6042 /  7535 =  80.186
1, 0  acc:  2453 /  2480 =  98.911
1, 1  acc:   172 /   180 =  95.556
------------------------------------
Average acc: 16008 / 19962 =  80.192
Robust  acc:  7341 /  9767 =  75.161
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7341 /  9767 =  75.161
0, 1  acc:  6042 /  7535 =  80.186
1, 0  acc:  2453 /  2480 =  98.911
1, 1  acc:   172 /   180 =  95.556
------------------------------------
Average acc: 16008 / 19962 =  80.192
Robust  acc:  7341 /  9767 =  75.161
------------------------------------
Epoch:  44 | Train Loss: 0.002 | Train Acc: 94.176 | Val Loss: 0.005 | Val Acc: 74.984
Training:
Accuracies by groups:
0, 0  acc:  8654 / 14496 =  59.699
0, 1  acc:  4554 /  6682 =  68.153
1, 0  acc: 131659 / 132845 =  99.107
1, 1  acc:  8424 /  8747 =  96.307
--------------------------------------
Average acc: 153291 / 162770 =  94.176
Robust  acc:  8654 / 14496 =  59.699
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5648 /  8535 =  66.175
0, 1  acc:  6218 /  8276 =  75.133
1, 0  acc:  2853 /  2874 =  99.269
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 14897 / 19867 =  74.984
Robust  acc:  5648 /  8535 =  66.175
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 77.137
Robust Acc: 71.967 | Best Acc: 99.032
-------------------------------------
Training, Epoch 43:
Accuracies by groups:
0, 0  acc:  7029 /  9767 =  71.967
0, 1  acc:  5739 /  7535 =  76.165
1, 0  acc:  2456 /  2480 =  99.032
1, 1  acc:   174 /   180 =  96.667
------------------------------------
Average acc: 15398 / 19962 =  77.137
Robust  acc:  7029 /  9767 =  71.967
------------------------------------
Accuracies by groups:
0, 0  acc:  7029 /  9767 =  71.967
0, 1  acc:  5739 /  7535 =  76.165
1, 0  acc:  2456 /  2480 =  99.032
1, 1  acc:   174 /   180 =  96.667
------------------------------------
Average acc: 15398 / 19962 =  77.137
Robust  acc:  7029 /  9767 =  71.967
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7029 /  9767 =  71.967
0, 1  acc:  5739 /  7535 =  76.165
1, 0  acc:  2456 /  2480 =  99.032
1, 1  acc:   174 /   180 =  96.667
------------------------------------
Average acc: 15398 / 19962 =  77.137
Robust  acc:  7029 /  9767 =  71.967
------------------------------------
Epoch:  45 | Train Loss: 0.002 | Train Acc: 94.161 | Val Loss: 0.010 | Val Acc: 32.416
Training:
Accuracies by groups:
0, 0  acc:  8594 / 14491 =  59.306
0, 1  acc:  4465 /  6532 =  68.356
1, 0  acc: 132009 / 133181 =  99.120
1, 1  acc:  8198 /  8566 =  95.704
--------------------------------------
Average acc: 153266 / 162770 =  94.161
Robust  acc:  8594 / 14491 =  59.306
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  1345 /  8535 =  15.759
0, 1  acc:  2040 /  8276 =  24.650
1, 0  acc:  2873 /  2874 =  99.965
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc:  6440 / 19867 =  32.416
Robust  acc:  1345 /  8535 =  15.759
------------------------------------
-------------------------------------------
Avg Test Loss: 0.009 | Avg Test Acc: 30.543
Robust Acc: 17.600 | Best Acc: 99.919
-------------------------------------
Training, Epoch 44:
Accuracies by groups:
0, 0  acc:  1719 /  9767 =  17.600
0, 1  acc:  1722 /  7535 =  22.853
1, 0  acc:  2478 /  2480 =  99.919
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  6097 / 19962 =  30.543
Robust  acc:  1719 /  9767 =  17.600
------------------------------------
Accuracies by groups:
0, 0  acc:  1719 /  9767 =  17.600
0, 1  acc:  1722 /  7535 =  22.853
1, 0  acc:  2478 /  2480 =  99.919
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  6097 / 19962 =  30.543
Robust  acc:  1719 /  9767 =  17.600
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  1719 /  9767 =  17.600
0, 1  acc:  1722 /  7535 =  22.853
1, 0  acc:  2478 /  2480 =  99.919
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  6097 / 19962 =  30.543
Robust  acc:  1719 /  9767 =  17.600
------------------------------------
Epoch:  46 | Train Loss: 0.002 | Train Acc: 94.392 | Val Loss: 0.004 | Val Acc: 84.215
Training:
Accuracies by groups:
0, 0  acc:  8661 / 14330 =  60.440
0, 1  acc:  4427 /  6430 =  68.849
1, 0  acc: 132179 / 133294 =  99.164
1, 1  acc:  8375 /  8716 =  96.088
--------------------------------------
Average acc: 153642 / 162770 =  94.392
Robust  acc:  8661 / 14330 =  60.440
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6759 /  8535 =  79.192
0, 1  acc:  6979 /  8276 =  84.328
1, 0  acc:  2817 /  2874 =  98.017
1, 1  acc:   176 /   182 =  96.703
------------------------------------
Average acc: 16731 / 19867 =  84.215
Robust  acc:  6759 /  8535 =  79.192
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 85.808
Robust Acc: 83.762 | Best Acc: 97.137
-------------------------------------
Training, Epoch 45:
Accuracies by groups:
0, 0  acc:  8181 /  9767 =  83.762
0, 1  acc:  6371 /  7535 =  84.552
1, 0  acc:  2409 /  2480 =  97.137
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 17129 / 19962 =  85.808
Robust  acc:  8181 /  9767 =  83.762
------------------------------------
Accuracies by groups:
0, 0  acc:  8181 /  9767 =  83.762
0, 1  acc:  6371 /  7535 =  84.552
1, 0  acc:  2409 /  2480 =  97.137
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 17129 / 19962 =  85.808
Robust  acc:  8181 /  9767 =  83.762
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8181 /  9767 =  83.762
0, 1  acc:  6371 /  7535 =  84.552
1, 0  acc:  2409 /  2480 =  97.137
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 17129 / 19962 =  85.808
Robust  acc:  8181 /  9767 =  83.762
------------------------------------
Epoch:  47 | Train Loss: 0.002 | Train Acc: 94.232 | Val Loss: 0.004 | Val Acc: 81.497
Training:
Accuracies by groups:
0, 0  acc:  8520 / 14317 =  59.510
0, 1  acc:  4407 /  6465 =  68.167
1, 0  acc: 132267 / 133441 =  99.120
1, 1  acc:  8188 /  8547 =  95.800
--------------------------------------
Average acc: 153382 / 162770 =  94.232
Robust  acc:  8520 / 14317 =  59.510
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6350 /  8535 =  74.400
0, 1  acc:  6826 /  8276 =  82.479
1, 0  acc:  2838 /  2874 =  98.747
1, 1  acc:   177 /   182 =  97.253
------------------------------------
Average acc: 16191 / 19867 =  81.497
Robust  acc:  6350 /  8535 =  74.400
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 83.083
Robust Acc: 79.625 | Best Acc: 98.427
-------------------------------------
Training, Epoch 46:
Accuracies by groups:
0, 0  acc:  7777 /  9767 =  79.625
0, 1  acc:  6199 /  7535 =  82.269
1, 0  acc:  2441 /  2480 =  98.427
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 16585 / 19962 =  83.083
Robust  acc:  7777 /  9767 =  79.625
------------------------------------
Accuracies by groups:
0, 0  acc:  7777 /  9767 =  79.625
0, 1  acc:  6199 /  7535 =  82.269
1, 0  acc:  2441 /  2480 =  98.427
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 16585 / 19962 =  83.083
Robust  acc:  7777 /  9767 =  79.625
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7777 /  9767 =  79.625
0, 1  acc:  6199 /  7535 =  82.269
1, 0  acc:  2441 /  2480 =  98.427
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 16585 / 19962 =  83.083
Robust  acc:  7777 /  9767 =  79.625
------------------------------------
Epoch:  48 | Train Loss: 0.002 | Train Acc: 94.231 | Val Loss: 0.005 | Val Acc: 74.596
Training:
Accuracies by groups:
0, 0  acc:  8646 / 14498 =  59.636
0, 1  acc:  4467 /  6541 =  68.292
1, 0  acc: 131918 / 133047 =  99.151
1, 1  acc:  8348 /  8684 =  96.131
--------------------------------------
Average acc: 153379 / 162770 =  94.231
Robust  acc:  8646 / 14498 =  59.636
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5644 /  8535 =  66.128
0, 1  acc:  6145 /  8276 =  74.251
1, 0  acc:  2852 /  2874 =  99.235
1, 1  acc:   179 /   182 =  98.352
------------------------------------
Average acc: 14820 / 19867 =  74.596
Robust  acc:  5644 /  8535 =  66.128
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 76.375
Robust Acc: 71.465 | Best Acc: 99.274
-------------------------------------
Training, Epoch 47:
Accuracies by groups:
0, 0  acc:  6980 /  9767 =  71.465
0, 1  acc:  5631 /  7535 =  74.731
1, 0  acc:  2462 /  2480 =  99.274
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15246 / 19962 =  76.375
Robust  acc:  6980 /  9767 =  71.465
------------------------------------
Accuracies by groups:
0, 0  acc:  6980 /  9767 =  71.465
0, 1  acc:  5631 /  7535 =  74.731
1, 0  acc:  2462 /  2480 =  99.274
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15246 / 19962 =  76.375
Robust  acc:  6980 /  9767 =  71.465
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6980 /  9767 =  71.465
0, 1  acc:  5631 /  7535 =  74.731
1, 0  acc:  2462 /  2480 =  99.274
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15246 / 19962 =  76.375
Robust  acc:  6980 /  9767 =  71.465
------------------------------------
Epoch:  49 | Train Loss: 0.002 | Train Acc: 94.224 | Val Loss: 0.003 | Val Acc: 93.819
Training:
Accuracies by groups:
0, 0  acc:  8810 / 14603 =  60.330
0, 1  acc:  4641 /  6696 =  69.310
1, 0  acc: 131592 / 132792 =  99.096
1, 1  acc:  8325 /  8679 =  95.921
--------------------------------------
Average acc: 153368 / 162770 =  94.224
Robust  acc:  8810 / 14603 =  60.330
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8168 /  8535 =  95.700
0, 1  acc:  8137 /  8276 =  98.320
1, 0  acc:  2231 /  2874 =  77.627
1, 1  acc:   103 /   182 =  56.593
------------------------------------
Average acc: 18639 / 19867 =  93.819
Robust  acc:   103 /   182 =  56.593
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 94.134
Robust Acc: 48.333 | Best Acc: 98.208
-------------------------------------
Training, Epoch 48:
Accuracies by groups:
0, 0  acc:  9402 /  9767 =  96.263
0, 1  acc:  7400 /  7535 =  98.208
1, 0  acc:  1902 /  2480 =  76.694
1, 1  acc:    87 /   180 =  48.333
------------------------------------
Average acc: 18791 / 19962 =  94.134
Robust  acc:    87 /   180 =  48.333
------------------------------------
Accuracies by groups:
0, 0  acc:  9402 /  9767 =  96.263
0, 1  acc:  7400 /  7535 =  98.208
1, 0  acc:  1902 /  2480 =  76.694
1, 1  acc:    87 /   180 =  48.333
------------------------------------
Average acc: 18791 / 19962 =  94.134
Robust  acc:    87 /   180 =  48.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9402 /  9767 =  96.263
0, 1  acc:  7400 /  7535 =  98.208
1, 0  acc:  1902 /  2480 =  76.694
1, 1  acc:    87 /   180 =  48.333
------------------------------------
Average acc: 18791 / 19962 =  94.134
Robust  acc:    87 /   180 =  48.333
------------------------------------
Epoch:  50 | Train Loss: 0.002 | Train Acc: 94.355 | Val Loss: 0.004 | Val Acc: 85.886
Training:
Accuracies by groups:
0, 0  acc:  8771 / 14444 =  60.724
0, 1  acc:  4585 /  6609 =  69.375
1, 0  acc: 131861 / 133006 =  99.139
1, 1  acc:  8365 /  8711 =  96.028
--------------------------------------
Average acc: 153582 / 162770 =  94.355
Robust  acc:  8771 / 14444 =  60.724
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6788 /  8535 =  79.531
0, 1  acc:  7292 /  8276 =  88.110
1, 0  acc:  2819 /  2874 =  98.086
1, 1  acc:   164 /   182 =  90.110
------------------------------------
Average acc: 17063 / 19867 =  85.886
Robust  acc:  6788 /  8535 =  79.531
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 87.366
Robust Acc: 83.710 | Best Acc: 97.339
-------------------------------------
Training, Epoch 49:
Accuracies by groups:
0, 0  acc:  8176 /  9767 =  83.710
0, 1  acc:  6689 /  7535 =  88.772
1, 0  acc:  2414 /  2480 =  97.339
1, 1  acc:   161 /   180 =  89.444
------------------------------------
Average acc: 17440 / 19962 =  87.366
Robust  acc:  8176 /  9767 =  83.710
------------------------------------
Accuracies by groups:
0, 0  acc:  8176 /  9767 =  83.710
0, 1  acc:  6689 /  7535 =  88.772
1, 0  acc:  2414 /  2480 =  97.339
1, 1  acc:   161 /   180 =  89.444
------------------------------------
Average acc: 17440 / 19962 =  87.366
Robust  acc:  8176 /  9767 =  83.710
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8176 /  9767 =  83.710
0, 1  acc:  6689 /  7535 =  88.772
1, 0  acc:  2414 /  2480 =  97.339
1, 1  acc:   161 /   180 =  89.444
------------------------------------
Average acc: 17440 / 19962 =  87.366
Robust  acc:  8176 /  9767 =  83.710
------------------------------------
replace: True
-> Updating checkpoint debias-end_seed11.pt...
Checkpoint saved at ./model/celebA/config/debias-end_seed11.pt
