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: 8
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=8-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: 90.793 | Val Loss: 0.004 | Val Acc: 78.356
Training:
Accuracies by groups:
0, 0  acc:  5722 / 14423 =  39.673
0, 1  acc:  2844 /  6548 =  43.433
1, 0  acc: 131014 / 133148 =  98.397
1, 1  acc:  8204 /  8651 =  94.833
--------------------------------------
Average acc: 147784 / 162770 =  90.793
Robust  acc:  5722 / 14423 =  39.673
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6079 /  8535 =  71.224
0, 1  acc:  6470 /  8276 =  78.178
1, 0  acc:  2847 /  2874 =  99.061
1, 1  acc:   171 /   182 =  93.956
------------------------------------
Average acc: 15567 / 19867 =  78.356
Robust  acc:  6079 /  8535 =  71.224
------------------------------------
New max robust acc: 71.22437024018747
debias model - Saving best checkpoint at epoch 0
replace: True
-> Updating checkpoint debias-wga-best_seed8.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed8.pt
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 80.363
Robust Acc: 76.902 | Best Acc: 98.871
-------------------------------------
Training, Epoch 0:
Accuracies by groups:
0, 0  acc:  7511 /  9767 =  76.902
0, 1  acc:  5906 /  7535 =  78.381
1, 0  acc:  2452 /  2480 =  98.871
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 16042 / 19962 =  80.363
Robust  acc:  7511 /  9767 =  76.902
------------------------------------
Accuracies by groups:
0, 0  acc:  7511 /  9767 =  76.902
0, 1  acc:  5906 /  7535 =  78.381
1, 0  acc:  2452 /  2480 =  98.871
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 16042 / 19962 =  80.363
Robust  acc:  7511 /  9767 =  76.902
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7511 /  9767 =  76.902
0, 1  acc:  5906 /  7535 =  78.381
1, 0  acc:  2452 /  2480 =  98.871
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 16042 / 19962 =  80.363
Robust  acc:  7511 /  9767 =  76.902
------------------------------------
Epoch:   2 | Train Loss: 0.001 | Train Acc: 94.595 | Val Loss: 0.003 | Val Acc: 83.898
Training:
Accuracies by groups:
0, 0  acc:  9293 / 14461 =  64.262
0, 1  acc:  4755 /  6550 =  72.595
1, 0  acc: 131843 / 133161 =  99.010
1, 1  acc:  8081 /  8598 =  93.987
--------------------------------------
Average acc: 153972 / 162770 =  94.595
Robust  acc:  9293 / 14461 =  64.262
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6459 /  8535 =  75.677
0, 1  acc:  7192 /  8276 =  86.902
1, 0  acc:  2844 /  2874 =  98.956
1, 1  acc:   173 /   182 =  95.055
------------------------------------
Average acc: 16668 / 19867 =  83.898
Robust  acc:  6459 /  8535 =  75.677
------------------------------------
New max robust acc: 75.67662565905097
debias model - Saving best checkpoint at epoch 1
replace: True
-> Updating checkpoint debias-wga-best_seed8.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed8.pt
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 85.412
Robust Acc: 80.762 | Best Acc: 98.790
-------------------------------------
Training, Epoch 1:
Accuracies by groups:
0, 0  acc:  7888 /  9767 =  80.762
0, 1  acc:  6545 /  7535 =  86.861
1, 0  acc:  2450 /  2480 =  98.790
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 17050 / 19962 =  85.412
Robust  acc:  7888 /  9767 =  80.762
------------------------------------
Accuracies by groups:
0, 0  acc:  7888 /  9767 =  80.762
0, 1  acc:  6545 /  7535 =  86.861
1, 0  acc:  2450 /  2480 =  98.790
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 17050 / 19962 =  85.412
Robust  acc:  7888 /  9767 =  80.762
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7888 /  9767 =  80.762
0, 1  acc:  6545 /  7535 =  86.861
1, 0  acc:  2450 /  2480 =  98.790
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 17050 / 19962 =  85.412
Robust  acc:  7888 /  9767 =  80.762
------------------------------------
Epoch:   3 | Train Loss: 0.001 | Train Acc: 95.255 | Val Loss: 0.002 | Val Acc: 87.406
Training:
Accuracies by groups:
0, 0  acc:  9779 / 14510 =  67.395
0, 1  acc:  5286 /  6614 =  79.921
1, 0  acc: 131793 / 132996 =  99.095
1, 1  acc:  8188 /  8650 =  94.659
--------------------------------------
Average acc: 155046 / 162770 =  95.255
Robust  acc:  9779 / 14510 =  67.395
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6841 /  8535 =  80.152
0, 1  acc:  7523 /  8276 =  90.901
1, 0  acc:  2829 /  2874 =  98.434
1, 1  acc:   172 /   182 =  94.505
------------------------------------
Average acc: 17365 / 19867 =  87.406
Robust  acc:  6841 /  8535 =  80.152
------------------------------------
New max robust acc: 80.15231400117165
debias model - Saving best checkpoint at epoch 2
replace: True
-> Updating checkpoint debias-wga-best_seed8.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed8.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 88.333
Robust Acc: 84.284 | Best Acc: 98.065
-------------------------------------
Training, Epoch 2:
Accuracies by groups:
0, 0  acc:  8232 /  9767 =  84.284
0, 1  acc:  6808 /  7535 =  90.352
1, 0  acc:  2432 /  2480 =  98.065
1, 1  acc:   161 /   180 =  89.444
------------------------------------
Average acc: 17633 / 19962 =  88.333
Robust  acc:  8232 /  9767 =  84.284
------------------------------------
Accuracies by groups:
0, 0  acc:  8232 /  9767 =  84.284
0, 1  acc:  6808 /  7535 =  90.352
1, 0  acc:  2432 /  2480 =  98.065
1, 1  acc:   161 /   180 =  89.444
------------------------------------
Average acc: 17633 / 19962 =  88.333
Robust  acc:  8232 /  9767 =  84.284
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8232 /  9767 =  84.284
0, 1  acc:  6808 /  7535 =  90.352
1, 0  acc:  2432 /  2480 =  98.065
1, 1  acc:   161 /   180 =  89.444
------------------------------------
Average acc: 17633 / 19962 =  88.333
Robust  acc:  8232 /  9767 =  84.284
------------------------------------
Epoch:   4 | Train Loss: 0.001 | Train Acc: 95.766 | Val Loss: 0.002 | Val Acc: 87.371
Training:
Accuracies by groups:
0, 0  acc: 10102 / 14521 =  69.568
0, 1  acc:  5499 /  6562 =  83.801
1, 0  acc: 132092 / 133100 =  99.243
1, 1  acc:  8185 /  8587 =  95.319
--------------------------------------
Average acc: 155878 / 162770 =  95.766
Robust  acc: 10102 / 14521 =  69.568
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6809 /  8535 =  79.777
0, 1  acc:  7544 /  8276 =  91.155
1, 0  acc:  2833 /  2874 =  98.573
1, 1  acc:   172 /   182 =  94.505
------------------------------------
Average acc: 17358 / 19867 =  87.371
Robust  acc:  6809 /  8535 =  79.777
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 88.303
Robust Acc: 83.915 | Best Acc: 98.105
-------------------------------------
Training, Epoch 3:
Accuracies by groups:
0, 0  acc:  8196 /  9767 =  83.915
0, 1  acc:  6839 /  7535 =  90.763
1, 0  acc:  2433 /  2480 =  98.105
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17627 / 19962 =  88.303
Robust  acc:  8196 /  9767 =  83.915
------------------------------------
Accuracies by groups:
0, 0  acc:  8196 /  9767 =  83.915
0, 1  acc:  6839 /  7535 =  90.763
1, 0  acc:  2433 /  2480 =  98.105
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17627 / 19962 =  88.303
Robust  acc:  8196 /  9767 =  83.915
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8196 /  9767 =  83.915
0, 1  acc:  6839 /  7535 =  90.763
1, 0  acc:  2433 /  2480 =  98.105
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 17627 / 19962 =  88.303
Robust  acc:  8196 /  9767 =  83.915
------------------------------------
Epoch:   5 | Train Loss: 0.001 | Train Acc: 96.371 | Val Loss: 0.002 | Val Acc: 88.302
Training:
Accuracies by groups:
0, 0  acc: 10517 / 14451 =  72.777
0, 1  acc:  5644 /  6534 =  86.379
1, 0  acc: 132261 / 133025 =  99.426
1, 1  acc:  8441 /  8760 =  96.358
--------------------------------------
Average acc: 156863 / 162770 =  96.371
Robust  acc: 10517 / 14451 =  72.777
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6907 /  8535 =  80.926
0, 1  acc:  7641 /  8276 =  92.327
1, 0  acc:  2825 /  2874 =  98.295
1, 1  acc:   170 /   182 =  93.407
------------------------------------
Average acc: 17543 / 19867 =  88.302
Robust  acc:  6907 /  8535 =  80.926
------------------------------------
New max robust acc: 80.92560046865846
debias model - Saving best checkpoint at epoch 4
replace: True
-> Updating checkpoint debias-wga-best_seed8.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed8.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 89.114
Robust Acc: 84.837 | Best Acc: 97.823
-------------------------------------
Training, Epoch 4:
Accuracies by groups:
0, 0  acc:  8286 /  9767 =  84.837
0, 1  acc:  6920 /  7535 =  91.838
1, 0  acc:  2426 /  2480 =  97.823
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 17789 / 19962 =  89.114
Robust  acc:  8286 /  9767 =  84.837
------------------------------------
Accuracies by groups:
0, 0  acc:  8286 /  9767 =  84.837
0, 1  acc:  6920 /  7535 =  91.838
1, 0  acc:  2426 /  2480 =  97.823
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 17789 / 19962 =  89.114
Robust  acc:  8286 /  9767 =  84.837
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8286 /  9767 =  84.837
0, 1  acc:  6920 /  7535 =  91.838
1, 0  acc:  2426 /  2480 =  97.823
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 17789 / 19962 =  89.114
Robust  acc:  8286 /  9767 =  84.837
------------------------------------
Epoch:   6 | Train Loss: 0.001 | Train Acc: 96.751 | Val Loss: 0.002 | Val Acc: 90.311
Training:
Accuracies by groups:
0, 0  acc: 10881 / 14607 =  74.492
0, 1  acc:  5851 /  6654 =  87.932
1, 0  acc: 132334 / 132897 =  99.576
1, 1  acc:  8415 /  8612 =  97.712
--------------------------------------
Average acc: 157481 / 162770 =  96.751
Robust  acc: 10881 / 14607 =  74.492
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7147 /  8535 =  83.738
0, 1  acc:  7822 /  8276 =  94.514
1, 0  acc:  2808 /  2874 =  97.704
1, 1  acc:   165 /   182 =  90.659
------------------------------------
Average acc: 17942 / 19867 =  90.311
Robust  acc:  7147 /  8535 =  83.738
------------------------------------
New max robust acc: 83.73755125951962
debias model - Saving best checkpoint at epoch 5
replace: True
-> Updating checkpoint debias-wga-best_seed8.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed8.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 90.893
Robust Acc: 83.889 | Best Acc: 97.500
-------------------------------------
Training, Epoch 5:
Accuracies by groups:
0, 0  acc:  8489 /  9767 =  86.915
0, 1  acc:  7086 /  7535 =  94.041
1, 0  acc:  2418 /  2480 =  97.500
1, 1  acc:   151 /   180 =  83.889
------------------------------------
Average acc: 18144 / 19962 =  90.893
Robust  acc:   151 /   180 =  83.889
------------------------------------
Accuracies by groups:
0, 0  acc:  8489 /  9767 =  86.915
0, 1  acc:  7086 /  7535 =  94.041
1, 0  acc:  2418 /  2480 =  97.500
1, 1  acc:   151 /   180 =  83.889
------------------------------------
Average acc: 18144 / 19962 =  90.893
Robust  acc:   151 /   180 =  83.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8489 /  9767 =  86.915
0, 1  acc:  7086 /  7535 =  94.041
1, 0  acc:  2418 /  2480 =  97.500
1, 1  acc:   151 /   180 =  83.889
------------------------------------
Average acc: 18144 / 19962 =  90.893
Robust  acc:   151 /   180 =  83.889
------------------------------------
Epoch:   7 | Train Loss: 0.001 | Train Acc: 97.374 | Val Loss: 0.002 | Val Acc: 90.935
Training:
Accuracies by groups:
0, 0  acc: 11084 / 14320 =  77.402
0, 1  acc:  5902 /  6532 =  90.355
1, 0  acc: 133023 / 133339 =  99.763
1, 1  acc:  8487 /  8579 =  98.928
--------------------------------------
Average acc: 158496 / 162770 =  97.374
Robust  acc: 11084 / 14320 =  77.402
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7256 /  8535 =  85.015
0, 1  acc:  7845 /  8276 =  94.792
1, 0  acc:  2801 /  2874 =  97.460
1, 1  acc:   164 /   182 =  90.110
------------------------------------
Average acc: 18066 / 19867 =  90.935
Robust  acc:  7256 /  8535 =  85.015
------------------------------------
New max robust acc: 85.01464557703574
debias model - Saving best checkpoint at epoch 6
replace: True
-> Updating checkpoint debias-wga-best_seed8.pt...
Checkpoint saved at ./model/celebA/config/debias-wga-best_seed8.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.354
Robust Acc: 81.111 | Best Acc: 96.976
-------------------------------------
Training, Epoch 6:
Accuracies by groups:
0, 0  acc:  8580 /  9767 =  87.847
0, 1  acc:  7105 /  7535 =  94.293
1, 0  acc:  2405 /  2480 =  96.976
1, 1  acc:   146 /   180 =  81.111
------------------------------------
Average acc: 18236 / 19962 =  91.354
Robust  acc:   146 /   180 =  81.111
------------------------------------
Accuracies by groups:
0, 0  acc:  8580 /  9767 =  87.847
0, 1  acc:  7105 /  7535 =  94.293
1, 0  acc:  2405 /  2480 =  96.976
1, 1  acc:   146 /   180 =  81.111
------------------------------------
Average acc: 18236 / 19962 =  91.354
Robust  acc:   146 /   180 =  81.111
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8580 /  9767 =  87.847
0, 1  acc:  7105 /  7535 =  94.293
1, 0  acc:  2405 /  2480 =  96.976
1, 1  acc:   146 /   180 =  81.111
------------------------------------
Average acc: 18236 / 19962 =  91.354
Robust  acc:   146 /   180 =  81.111
------------------------------------
Epoch:   8 | Train Loss: 0.001 | Train Acc: 97.796 | Val Loss: 0.002 | Val Acc: 92.354
Training:
Accuracies by groups:
0, 0  acc: 11875 / 14682 =  80.881
0, 1  acc:  6014 /  6575 =  91.468
1, 0  acc: 132795 / 132982 =  99.859
1, 1  acc:  8498 /  8531 =  99.613
--------------------------------------
Average acc: 159182 / 162770 =  97.796
Robust  acc: 11875 / 14682 =  80.881
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7437 /  8535 =  87.135
0, 1  acc:  7986 /  8276 =  96.496
1, 0  acc:  2773 /  2874 =  96.486
1, 1  acc:   152 /   182 =  83.516
------------------------------------
Average acc: 18348 / 19867 =  92.354
Robust  acc:   152 /   182 =  83.516
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.536
Robust Acc: 77.222 | Best Acc: 95.806
-------------------------------------
Training, Epoch 7:
Accuracies by groups:
0, 0  acc:  8740 /  9767 =  89.485
0, 1  acc:  7219 /  7535 =  95.806
1, 0  acc:  2374 /  2480 =  95.726
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 18472 / 19962 =  92.536
Robust  acc:   139 /   180 =  77.222
------------------------------------
Accuracies by groups:
0, 0  acc:  8740 /  9767 =  89.485
0, 1  acc:  7219 /  7535 =  95.806
1, 0  acc:  2374 /  2480 =  95.726
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 18472 / 19962 =  92.536
Robust  acc:   139 /   180 =  77.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8740 /  9767 =  89.485
0, 1  acc:  7219 /  7535 =  95.806
1, 0  acc:  2374 /  2480 =  95.726
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 18472 / 19962 =  92.536
Robust  acc:   139 /   180 =  77.222
------------------------------------
Epoch:   9 | Train Loss: 0.001 | Train Acc: 98.206 | Val Loss: 0.001 | Val Acc: 93.039
Training:
Accuracies by groups:
0, 0  acc: 12078 / 14442 =  83.631
0, 1  acc:  5908 /  6353 =  92.995
1, 0  acc: 133293 / 133378 =  99.936
1, 1  acc:  8571 /  8597 =  99.698
--------------------------------------
Average acc: 159850 / 162770 =  98.206
Robust  acc: 12078 / 14442 =  83.631
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7570 /  8535 =  88.694
0, 1  acc:  8022 /  8276 =  96.931
1, 0  acc:  2749 /  2874 =  95.651
1, 1  acc:   143 /   182 =  78.571
------------------------------------
Average acc: 18484 / 19867 =  93.039
Robust  acc:   143 /   182 =  78.571
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 93.202
Robust Acc: 73.889 | Best Acc: 96.576
-------------------------------------
Training, Epoch 8:
Accuracies by groups:
0, 0  acc:  8871 /  9767 =  90.826
0, 1  acc:  7277 /  7535 =  96.576
1, 0  acc:  2324 /  2480 =  93.710
1, 1  acc:   133 /   180 =  73.889
------------------------------------
Average acc: 18605 / 19962 =  93.202
Robust  acc:   133 /   180 =  73.889
------------------------------------
Accuracies by groups:
0, 0  acc:  8871 /  9767 =  90.826
0, 1  acc:  7277 /  7535 =  96.576
1, 0  acc:  2324 /  2480 =  93.710
1, 1  acc:   133 /   180 =  73.889
------------------------------------
Average acc: 18605 / 19962 =  93.202
Robust  acc:   133 /   180 =  73.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8871 /  9767 =  90.826
0, 1  acc:  7277 /  7535 =  96.576
1, 0  acc:  2324 /  2480 =  93.710
1, 1  acc:   133 /   180 =  73.889
------------------------------------
Average acc: 18605 / 19962 =  93.202
Robust  acc:   133 /   180 =  73.889
------------------------------------
Epoch:  10 | Train Loss: 0.001 | Train Acc: 98.475 | Val Loss: 0.002 | Val Acc: 92.576
Training:
Accuracies by groups:
0, 0  acc: 12485 / 14491 =  86.157
0, 1  acc:  6206 /  6550 =  94.748
1, 0  acc: 132972 / 133063 =  99.932
1, 1  acc:  8625 /  8666 =  99.527
--------------------------------------
Average acc: 160288 / 162770 =  98.475
Robust  acc: 12485 / 14491 =  86.157
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7503 /  8535 =  87.909
0, 1  acc:  7988 /  8276 =  96.520
1, 0  acc:  2752 /  2874 =  95.755
1, 1  acc:   149 /   182 =  81.868
------------------------------------
Average acc: 18392 / 19867 =  92.576
Robust  acc:   149 /   182 =  81.868
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.861
Robust Acc: 77.778 | Best Acc: 95.939
-------------------------------------
Training, Epoch 9:
Accuracies by groups:
0, 0  acc:  8827 /  9767 =  90.376
0, 1  acc:  7229 /  7535 =  95.939
1, 0  acc:  2341 /  2480 =  94.395
1, 1  acc:   140 /   180 =  77.778
------------------------------------
Average acc: 18537 / 19962 =  92.861
Robust  acc:   140 /   180 =  77.778
------------------------------------
Accuracies by groups:
0, 0  acc:  8827 /  9767 =  90.376
0, 1  acc:  7229 /  7535 =  95.939
1, 0  acc:  2341 /  2480 =  94.395
1, 1  acc:   140 /   180 =  77.778
------------------------------------
Average acc: 18537 / 19962 =  92.861
Robust  acc:   140 /   180 =  77.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8827 /  9767 =  90.376
0, 1  acc:  7229 /  7535 =  95.939
1, 0  acc:  2341 /  2480 =  94.395
1, 1  acc:   140 /   180 =  77.778
------------------------------------
Average acc: 18537 / 19962 =  92.861
Robust  acc:   140 /   180 =  77.778
------------------------------------
Epoch:  11 | Train Loss: 0.001 | Train Acc: 98.493 | Val Loss: 0.001 | Val Acc: 93.351
Training:
Accuracies by groups:
0, 0  acc: 12845 / 14760 =  87.026
0, 1  acc:  6141 /  6461 =  95.047
1, 0  acc: 132774 / 132942 =  99.874
1, 1  acc:  8557 /  8607 =  99.419
--------------------------------------
Average acc: 160317 / 162770 =  98.493
Robust  acc: 12845 / 14760 =  87.026
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7681 /  8535 =  89.994
0, 1  acc:  8052 /  8276 =  97.293
1, 0  acc:  2672 /  2874 =  92.971
1, 1  acc:   141 /   182 =  77.473
------------------------------------
Average acc: 18546 / 19867 =  93.351
Robust  acc:   141 /   182 =  77.473
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 93.653
Robust Acc: 71.111 | Best Acc: 96.815
-------------------------------------
Training, Epoch 10:
Accuracies by groups:
0, 0  acc:  8988 /  9767 =  92.024
0, 1  acc:  7295 /  7535 =  96.815
1, 0  acc:  2284 /  2480 =  92.097
1, 1  acc:   128 /   180 =  71.111
------------------------------------
Average acc: 18695 / 19962 =  93.653
Robust  acc:   128 /   180 =  71.111
------------------------------------
Accuracies by groups:
0, 0  acc:  8988 /  9767 =  92.024
0, 1  acc:  7295 /  7535 =  96.815
1, 0  acc:  2284 /  2480 =  92.097
1, 1  acc:   128 /   180 =  71.111
------------------------------------
Average acc: 18695 / 19962 =  93.653
Robust  acc:   128 /   180 =  71.111
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8988 /  9767 =  92.024
0, 1  acc:  7295 /  7535 =  96.815
1, 0  acc:  2284 /  2480 =  92.097
1, 1  acc:   128 /   180 =  71.111
------------------------------------
Average acc: 18695 / 19962 =  93.653
Robust  acc:   128 /   180 =  71.111
------------------------------------
Epoch:  12 | Train Loss: 0.001 | Train Acc: 98.276 | Val Loss: 0.002 | Val Acc: 91.906
Training:
Accuracies by groups:
0, 0  acc: 12443 / 14500 =  85.814
0, 1  acc:  6203 /  6593 =  94.085
1, 0  acc: 132696 / 132974 =  99.791
1, 1  acc:  8622 /  8703 =  99.069
--------------------------------------
Average acc: 159964 / 162770 =  98.276
Robust  acc: 12443 / 14500 =  85.814
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7497 /  8535 =  87.838
0, 1  acc:  7912 /  8276 =  95.602
1, 0  acc:  2697 /  2874 =  93.841
1, 1  acc:   153 /   182 =  84.066
------------------------------------
Average acc: 18259 / 19867 =  91.906
Robust  acc:   153 /   182 =  84.066
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.295
Robust Acc: 80.556 | Best Acc: 94.877
-------------------------------------
Training, Epoch 11:
Accuracies by groups:
0, 0  acc:  8814 /  9767 =  90.243
0, 1  acc:  7149 /  7535 =  94.877
1, 0  acc:  2316 /  2480 =  93.387
1, 1  acc:   145 /   180 =  80.556
------------------------------------
Average acc: 18424 / 19962 =  92.295
Robust  acc:   145 /   180 =  80.556
------------------------------------
Accuracies by groups:
0, 0  acc:  8814 /  9767 =  90.243
0, 1  acc:  7149 /  7535 =  94.877
1, 0  acc:  2316 /  2480 =  93.387
1, 1  acc:   145 /   180 =  80.556
------------------------------------
Average acc: 18424 / 19962 =  92.295
Robust  acc:   145 /   180 =  80.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8814 /  9767 =  90.243
0, 1  acc:  7149 /  7535 =  94.877
1, 0  acc:  2316 /  2480 =  93.387
1, 1  acc:   145 /   180 =  80.556
------------------------------------
Average acc: 18424 / 19962 =  92.295
Robust  acc:   145 /   180 =  80.556
------------------------------------
Epoch:  13 | Train Loss: 0.001 | Train Acc: 97.882 | Val Loss: 0.002 | Val Acc: 90.542
Training:
Accuracies by groups:
0, 0  acc: 12085 / 14499 =  83.351
0, 1  acc:  6045 /  6506 =  92.914
1, 0  acc: 132838 / 133284 =  99.665
1, 1  acc:  8354 /  8481 =  98.503
--------------------------------------
Average acc: 159322 / 162770 =  97.882
Robust  acc: 12085 / 14499 =  83.351
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7132 /  8535 =  83.562
0, 1  acc:  7922 /  8276 =  95.723
1, 0  acc:  2789 /  2874 =  97.042
1, 1  acc:   145 /   182 =  79.670
------------------------------------
Average acc: 17988 / 19867 =  90.542
Robust  acc:   145 /   182 =  79.670
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.078
Robust Acc: 72.222 | Best Acc: 96.290
-------------------------------------
Training, Epoch 12:
Accuracies by groups:
0, 0  acc:  8456 /  9767 =  86.577
0, 1  acc:  7207 /  7535 =  95.647
1, 0  acc:  2388 /  2480 =  96.290
1, 1  acc:   130 /   180 =  72.222
------------------------------------
Average acc: 18181 / 19962 =  91.078
Robust  acc:   130 /   180 =  72.222
------------------------------------
Accuracies by groups:
0, 0  acc:  8456 /  9767 =  86.577
0, 1  acc:  7207 /  7535 =  95.647
1, 0  acc:  2388 /  2480 =  96.290
1, 1  acc:   130 /   180 =  72.222
------------------------------------
Average acc: 18181 / 19962 =  91.078
Robust  acc:   130 /   180 =  72.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8456 /  9767 =  86.577
0, 1  acc:  7207 /  7535 =  95.647
1, 0  acc:  2388 /  2480 =  96.290
1, 1  acc:   130 /   180 =  72.222
------------------------------------
Average acc: 18181 / 19962 =  91.078
Robust  acc:   130 /   180 =  72.222
------------------------------------
Epoch:  14 | Train Loss: 0.001 | Train Acc: 97.349 | Val Loss: 0.003 | Val Acc: 83.973
Training:
Accuracies by groups:
0, 0  acc: 11539 / 14432 =  79.954
0, 1  acc:  5900 /  6469 =  91.204
1, 0  acc: 132599 / 133258 =  99.505
1, 1  acc:  8417 /  8611 =  97.747
--------------------------------------
Average acc: 158455 / 162770 =  97.349
Robust  acc: 11539 / 14432 =  79.954
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6220 /  8535 =  72.876
0, 1  acc:  7438 /  8276 =  89.874
1, 0  acc:  2852 /  2874 =  99.235
1, 1  acc:   173 /   182 =  95.055
------------------------------------
Average acc: 16683 / 19867 =  83.973
Robust  acc:  6220 /  8535 =  72.876
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 84.981
Robust Acc: 77.885 | Best Acc: 98.992
-------------------------------------
Training, Epoch 13:
Accuracies by groups:
0, 0  acc:  7607 /  9767 =  77.885
0, 1  acc:  6736 /  7535 =  89.396
1, 0  acc:  2455 /  2480 =  98.992
1, 1  acc:   166 /   180 =  92.222
------------------------------------
Average acc: 16964 / 19962 =  84.981
Robust  acc:  7607 /  9767 =  77.885
------------------------------------
Accuracies by groups:
0, 0  acc:  7607 /  9767 =  77.885
0, 1  acc:  6736 /  7535 =  89.396
1, 0  acc:  2455 /  2480 =  98.992
1, 1  acc:   166 /   180 =  92.222
------------------------------------
Average acc: 16964 / 19962 =  84.981
Robust  acc:  7607 /  9767 =  77.885
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7607 /  9767 =  77.885
0, 1  acc:  6736 /  7535 =  89.396
1, 0  acc:  2455 /  2480 =  98.992
1, 1  acc:   166 /   180 =  92.222
------------------------------------
Average acc: 16964 / 19962 =  84.981
Robust  acc:  7607 /  9767 =  77.885
------------------------------------
Epoch:  15 | Train Loss: 0.001 | Train Acc: 96.958 | Val Loss: 0.003 | Val Acc: 87.431
Training:
Accuracies by groups:
0, 0  acc: 11390 / 14671 =  77.636
0, 1  acc:  5879 /  6559 =  89.633
1, 0  acc: 132115 / 132870 =  99.432
1, 1  acc:  8435 /  8670 =  97.290
--------------------------------------
Average acc: 157819 / 162770 =  96.958
Robust  acc: 11390 / 14671 =  77.636
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6775 /  8535 =  79.379
0, 1  acc:  7587 /  8276 =  91.675
1, 0  acc:  2840 /  2874 =  98.817
1, 1  acc:   168 /   182 =  92.308
------------------------------------
Average acc: 17370 / 19867 =  87.431
Robust  acc:  6775 /  8535 =  79.379
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 88.238
Robust Acc: 82.943 | Best Acc: 98.347
-------------------------------------
Training, Epoch 14:
Accuracies by groups:
0, 0  acc:  8101 /  9767 =  82.943
0, 1  acc:  6917 /  7535 =  91.798
1, 0  acc:  2439 /  2480 =  98.347
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 17614 / 19962 =  88.238
Robust  acc:  8101 /  9767 =  82.943
------------------------------------
Accuracies by groups:
0, 0  acc:  8101 /  9767 =  82.943
0, 1  acc:  6917 /  7535 =  91.798
1, 0  acc:  2439 /  2480 =  98.347
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 17614 / 19962 =  88.238
Robust  acc:  8101 /  9767 =  82.943
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8101 /  9767 =  82.943
0, 1  acc:  6917 /  7535 =  91.798
1, 0  acc:  2439 /  2480 =  98.347
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 17614 / 19962 =  88.238
Robust  acc:  8101 /  9767 =  82.943
------------------------------------
Epoch:  16 | Train Loss: 0.001 | Train Acc: 96.357 | Val Loss: 0.005 | Val Acc: 74.133
Training:
Accuracies by groups:
0, 0  acc: 10977 / 14787 =  74.234
0, 1  acc:  5524 /  6370 =  86.719
1, 0  acc: 131905 / 132877 =  99.268
1, 1  acc:  8435 /  8736 =  96.554
--------------------------------------
Average acc: 156841 / 162770 =  96.357
Robust  acc: 10977 / 14787 =  74.234
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4967 /  8535 =  58.196
0, 1  acc:  6720 /  8276 =  81.199
1, 0  acc:  2865 /  2874 =  99.687
1, 1  acc:   176 /   182 =  96.703
------------------------------------
Average acc: 14728 / 19867 =  74.133
Robust  acc:  4967 /  8535 =  58.196
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 74.842
Robust Acc: 62.732 | Best Acc: 99.597
-------------------------------------
Training, Epoch 15:
Accuracies by groups:
0, 0  acc:  6127 /  9767 =  62.732
0, 1  acc:  6174 /  7535 =  81.938
1, 0  acc:  2470 /  2480 =  99.597
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 14940 / 19962 =  74.842
Robust  acc:  6127 /  9767 =  62.732
------------------------------------
Accuracies by groups:
0, 0  acc:  6127 /  9767 =  62.732
0, 1  acc:  6174 /  7535 =  81.938
1, 0  acc:  2470 /  2480 =  99.597
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 14940 / 19962 =  74.842
Robust  acc:  6127 /  9767 =  62.732
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6127 /  9767 =  62.732
0, 1  acc:  6174 /  7535 =  81.938
1, 0  acc:  2470 /  2480 =  99.597
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 14940 / 19962 =  74.842
Robust  acc:  6127 /  9767 =  62.732
------------------------------------
Epoch:  17 | Train Loss: 0.001 | Train Acc: 96.051 | Val Loss: 0.003 | Val Acc: 81.834
Training:
Accuracies by groups:
0, 0  acc: 10376 / 14375 =  72.181
0, 1  acc:  5618 /  6579 =  85.393
1, 0  acc: 132046 / 133137 =  99.181
1, 1  acc:  8302 /  8679 =  95.656
--------------------------------------
Average acc: 156342 / 162770 =  96.051
Robust  acc: 10376 / 14375 =  72.181
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6214 /  8535 =  72.806
0, 1  acc:  7013 /  8276 =  84.739
1, 0  acc:  2856 /  2874 =  99.374
1, 1  acc:   175 /   182 =  96.154
------------------------------------
Average acc: 16258 / 19867 =  81.834
Robust  acc:  6214 /  8535 =  72.806
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 83.383
Robust Acc: 77.588 | Best Acc: 98.992
-------------------------------------
Training, Epoch 16:
Accuracies by groups:
0, 0  acc:  7578 /  9767 =  77.588
0, 1  acc:  6448 /  7535 =  85.574
1, 0  acc:  2455 /  2480 =  98.992
1, 1  acc:   164 /   180 =  91.111
------------------------------------
Average acc: 16645 / 19962 =  83.383
Robust  acc:  7578 /  9767 =  77.588
------------------------------------
Accuracies by groups:
0, 0  acc:  7578 /  9767 =  77.588
0, 1  acc:  6448 /  7535 =  85.574
1, 0  acc:  2455 /  2480 =  98.992
1, 1  acc:   164 /   180 =  91.111
------------------------------------
Average acc: 16645 / 19962 =  83.383
Robust  acc:  7578 /  9767 =  77.588
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7578 /  9767 =  77.588
0, 1  acc:  6448 /  7535 =  85.574
1, 0  acc:  2455 /  2480 =  98.992
1, 1  acc:   164 /   180 =  91.111
------------------------------------
Average acc: 16645 / 19962 =  83.383
Robust  acc:  7578 /  9767 =  77.588
------------------------------------
Epoch:  18 | Train Loss: 0.001 | Train Acc: 95.623 | Val Loss: 0.003 | Val Acc: 83.983
Training:
Accuracies by groups:
0, 0  acc: 10110 / 14395 =  70.233
0, 1  acc:  5609 /  6697 =  83.754
1, 0  acc: 131853 / 133146 =  99.029
1, 1  acc:  8074 /  8532 =  94.632
--------------------------------------
Average acc: 155646 / 162770 =  95.623
Robust  acc: 10110 / 14395 =  70.233
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6575 /  8535 =  77.036
0, 1  acc:  7088 /  8276 =  85.645
1, 0  acc:  2846 /  2874 =  99.026
1, 1  acc:   176 /   182 =  96.703
------------------------------------
Average acc: 16685 / 19867 =  83.983
Robust  acc:  6575 /  8535 =  77.036
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 85.472
Robust Acc: 81.448 | Best Acc: 98.911
-------------------------------------
Training, Epoch 17:
Accuracies by groups:
0, 0  acc:  7955 /  9767 =  81.448
0, 1  acc:  6485 /  7535 =  86.065
1, 0  acc:  2453 /  2480 =  98.911
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 17062 / 19962 =  85.472
Robust  acc:  7955 /  9767 =  81.448
------------------------------------
Accuracies by groups:
0, 0  acc:  7955 /  9767 =  81.448
0, 1  acc:  6485 /  7535 =  86.065
1, 0  acc:  2453 /  2480 =  98.911
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 17062 / 19962 =  85.472
Robust  acc:  7955 /  9767 =  81.448
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7955 /  9767 =  81.448
0, 1  acc:  6485 /  7535 =  86.065
1, 0  acc:  2453 /  2480 =  98.911
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 17062 / 19962 =  85.472
Robust  acc:  7955 /  9767 =  81.448
------------------------------------
Epoch:  19 | Train Loss: 0.001 | Train Acc: 95.328 | Val Loss: 0.004 | Val Acc: 78.492
Training:
Accuracies by groups:
0, 0  acc:  9787 / 14408 =  67.928
0, 1  acc:  5334 /  6507 =  81.973
1, 0  acc: 131815 / 133196 =  98.963
1, 1  acc:  8229 /  8659 =  95.034
--------------------------------------
Average acc: 155165 / 162770 =  95.328
Robust  acc:  9787 / 14408 =  67.928
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5662 /  8535 =  66.339
0, 1  acc:  6894 /  8276 =  83.301
1, 0  acc:  2861 /  2874 =  99.548
1, 1  acc:   177 /   182 =  97.253
------------------------------------
Average acc: 15594 / 19867 =  78.492
Robust  acc:  5662 /  8535 =  66.339
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 80.257
Robust Acc: 72.182 | Best Acc: 99.476
-------------------------------------
Training, Epoch 18:
Accuracies by groups:
0, 0  acc:  7050 /  9767 =  72.182
0, 1  acc:  6333 /  7535 =  84.048
1, 0  acc:  2467 /  2480 =  99.476
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 16021 / 19962 =  80.257
Robust  acc:  7050 /  9767 =  72.182
------------------------------------
Accuracies by groups:
0, 0  acc:  7050 /  9767 =  72.182
0, 1  acc:  6333 /  7535 =  84.048
1, 0  acc:  2467 /  2480 =  99.476
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 16021 / 19962 =  80.257
Robust  acc:  7050 /  9767 =  72.182
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7050 /  9767 =  72.182
0, 1  acc:  6333 /  7535 =  84.048
1, 0  acc:  2467 /  2480 =  99.476
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 16021 / 19962 =  80.257
Robust  acc:  7050 /  9767 =  72.182
------------------------------------
Epoch:  20 | Train Loss: 0.001 | Train Acc: 95.058 | Val Loss: 0.003 | Val Acc: 86.571
Training:
Accuracies by groups:
0, 0  acc:  9563 / 14396 =  66.428
0, 1  acc:  5385 /  6747 =  79.813
1, 0  acc: 131703 / 133098 =  98.952
1, 1  acc:  8075 /  8529 =  94.677
--------------------------------------
Average acc: 154726 / 162770 =  95.058
Robust  acc:  9563 / 14396 =  66.428
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6756 /  8535 =  79.156
0, 1  acc:  7455 /  8276 =  90.080
1, 0  acc:  2822 /  2874 =  98.191
1, 1  acc:   166 /   182 =  91.209
------------------------------------
Average acc: 17199 / 19867 =  86.571
Robust  acc:  6756 /  8535 =  79.156
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 87.752
Robust Acc: 83.035 | Best Acc: 97.702
-------------------------------------
Training, Epoch 19:
Accuracies by groups:
0, 0  acc:  8110 /  9767 =  83.035
0, 1  acc:  6831 /  7535 =  90.657
1, 0  acc:  2423 /  2480 =  97.702
1, 1  acc:   153 /   180 =  85.000
------------------------------------
Average acc: 17517 / 19962 =  87.752
Robust  acc:  8110 /  9767 =  83.035
------------------------------------
Accuracies by groups:
0, 0  acc:  8110 /  9767 =  83.035
0, 1  acc:  6831 /  7535 =  90.657
1, 0  acc:  2423 /  2480 =  97.702
1, 1  acc:   153 /   180 =  85.000
------------------------------------
Average acc: 17517 / 19962 =  87.752
Robust  acc:  8110 /  9767 =  83.035
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8110 /  9767 =  83.035
0, 1  acc:  6831 /  7535 =  90.657
1, 0  acc:  2423 /  2480 =  97.702
1, 1  acc:   153 /   180 =  85.000
------------------------------------
Average acc: 17517 / 19962 =  87.752
Robust  acc:  8110 /  9767 =  83.035
------------------------------------
Epoch:  21 | Train Loss: 0.001 | Train Acc: 94.918 | Val Loss: 0.009 | Val Acc: 42.160
Training:
Accuracies by groups:
0, 0  acc:  9385 / 14401 =  65.169
0, 1  acc:  5037 /  6513 =  77.338
1, 0  acc: 131802 / 133144 =  98.992
1, 1  acc:  8274 /  8712 =  94.972
--------------------------------------
Average acc: 154498 / 162770 =  94.918
Robust  acc:  9385 / 14401 =  65.169
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  2447 /  8535 =  28.670
0, 1  acc:  2877 /  8276 =  34.763
1, 0  acc:  2871 /  2874 =  99.896
1, 1  acc:   181 /   182 =  99.451
------------------------------------
Average acc:  8376 / 19867 =  42.160
Robust  acc:  2447 /  8535 =  28.670
------------------------------------
-------------------------------------------
Avg Test Loss: 0.009 | Avg Test Acc: 42.025
Robust Acc: 32.262 | Best Acc: 99.960
-------------------------------------
Training, Epoch 20:
Accuracies by groups:
0, 0  acc:  3151 /  9767 =  32.262
0, 1  acc:  2581 /  7535 =  34.253
1, 0  acc:  2479 /  2480 =  99.960
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  8389 / 19962 =  42.025
Robust  acc:  3151 /  9767 =  32.262
------------------------------------
Accuracies by groups:
0, 0  acc:  3151 /  9767 =  32.262
0, 1  acc:  2581 /  7535 =  34.253
1, 0  acc:  2479 /  2480 =  99.960
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  8389 / 19962 =  42.025
Robust  acc:  3151 /  9767 =  32.262
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  3151 /  9767 =  32.262
0, 1  acc:  2581 /  7535 =  34.253
1, 0  acc:  2479 /  2480 =  99.960
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  8389 / 19962 =  42.025
Robust  acc:  3151 /  9767 =  32.262
------------------------------------
Epoch:  22 | Train Loss: 0.002 | Train Acc: 94.814 | Val Loss: 0.003 | Val Acc: 86.027
Training:
Accuracies by groups:
0, 0  acc:  9156 / 14335 =  63.872
0, 1  acc:  4922 /  6507 =  75.642
1, 0  acc: 132061 / 133315 =  99.059
1, 1  acc:  8189 /  8613 =  95.077
--------------------------------------
Average acc: 154328 / 162770 =  94.814
Robust  acc:  9156 / 14335 =  63.872
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6777 /  8535 =  79.402
0, 1  acc:  7320 /  8276 =  88.449
1, 0  acc:  2819 /  2874 =  98.086
1, 1  acc:   175 /   182 =  96.154
------------------------------------
Average acc: 17091 / 19867 =  86.027
Robust  acc:  6777 /  8535 =  79.402
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 87.336
Robust Acc: 84.059 | Best Acc: 97.782
-------------------------------------
Training, Epoch 21:
Accuracies by groups:
0, 0  acc:  8210 /  9767 =  84.059
0, 1  acc:  6634 /  7535 =  88.042
1, 0  acc:  2425 /  2480 =  97.782
1, 1  acc:   165 /   180 =  91.667
------------------------------------
Average acc: 17434 / 19962 =  87.336
Robust  acc:  8210 /  9767 =  84.059
------------------------------------
Accuracies by groups:
0, 0  acc:  8210 /  9767 =  84.059
0, 1  acc:  6634 /  7535 =  88.042
1, 0  acc:  2425 /  2480 =  97.782
1, 1  acc:   165 /   180 =  91.667
------------------------------------
Average acc: 17434 / 19962 =  87.336
Robust  acc:  8210 /  9767 =  84.059
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8210 /  9767 =  84.059
0, 1  acc:  6634 /  7535 =  88.042
1, 0  acc:  2425 /  2480 =  97.782
1, 1  acc:   165 /   180 =  91.667
------------------------------------
Average acc: 17434 / 19962 =  87.336
Robust  acc:  8210 /  9767 =  84.059
------------------------------------
Epoch:  23 | Train Loss: 0.002 | Train Acc: 94.501 | Val Loss: 0.003 | Val Acc: 91.800
Training:
Accuracies by groups:
0, 0  acc:  9112 / 14584 =  62.479
0, 1  acc:  4901 /  6717 =  72.964
1, 0  acc: 131634 / 132902 =  99.046
1, 1  acc:  8172 /  8567 =  95.389
--------------------------------------
Average acc: 153819 / 162770 =  94.501
Robust  acc:  9112 / 14584 =  62.479
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7485 /  8535 =  87.698
0, 1  acc:  7877 /  8276 =  95.179
1, 0  acc:  2735 /  2874 =  95.164
1, 1  acc:   141 /   182 =  77.473
------------------------------------
Average acc: 18238 / 19867 =  91.800
Robust  acc:   141 /   182 =  77.473
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 92.696
Robust Acc: 75.000 | Best Acc: 95.488
-------------------------------------
Training, Epoch 22:
Accuracies by groups:
0, 0  acc:  8836 /  9767 =  90.468
0, 1  acc:  7195 /  7535 =  95.488
1, 0  acc:  2338 /  2480 =  94.274
1, 1  acc:   135 /   180 =  75.000
------------------------------------
Average acc: 18504 / 19962 =  92.696
Robust  acc:   135 /   180 =  75.000
------------------------------------
Accuracies by groups:
0, 0  acc:  8836 /  9767 =  90.468
0, 1  acc:  7195 /  7535 =  95.488
1, 0  acc:  2338 /  2480 =  94.274
1, 1  acc:   135 /   180 =  75.000
------------------------------------
Average acc: 18504 / 19962 =  92.696
Robust  acc:   135 /   180 =  75.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8836 /  9767 =  90.468
0, 1  acc:  7195 /  7535 =  95.488
1, 0  acc:  2338 /  2480 =  94.274
1, 1  acc:   135 /   180 =  75.000
------------------------------------
Average acc: 18504 / 19962 =  92.696
Robust  acc:   135 /   180 =  75.000
------------------------------------
Epoch:  24 | Train Loss: 0.002 | Train Acc: 94.524 | Val Loss: 0.006 | Val Acc: 62.299
Training:
Accuracies by groups:
0, 0  acc:  8897 / 14524 =  61.257
0, 1  acc:  4746 /  6511 =  72.892
1, 0  acc: 132026 / 133176 =  99.136
1, 1  acc:  8188 /  8559 =  95.665
--------------------------------------
Average acc: 153857 / 162770 =  94.524
Robust  acc:  8897 / 14524 =  61.257
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4609 /  8535 =  54.001
0, 1  acc:  4721 /  8276 =  57.044
1, 0  acc:  2866 /  2874 =  99.722
1, 1  acc:   181 /   182 =  99.451
------------------------------------
Average acc: 12377 / 19867 =  62.299
Robust  acc:  4609 /  8535 =  54.001
------------------------------------
-------------------------------------------
Avg Test Loss: 0.006 | Avg Test Acc: 64.252
Robust Acc: 57.823 | Best Acc: 99.758
-------------------------------------
Training, Epoch 23:
Accuracies by groups:
0, 0  acc:  5817 /  9767 =  59.558
0, 1  acc:  4357 /  7535 =  57.823
1, 0  acc:  2474 /  2480 =  99.758
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 12826 / 19962 =  64.252
Robust  acc:  4357 /  7535 =  57.823
------------------------------------
Accuracies by groups:
0, 0  acc:  5817 /  9767 =  59.558
0, 1  acc:  4357 /  7535 =  57.823
1, 0  acc:  2474 /  2480 =  99.758
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 12826 / 19962 =  64.252
Robust  acc:  4357 /  7535 =  57.823
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  5817 /  9767 =  59.558
0, 1  acc:  4357 /  7535 =  57.823
1, 0  acc:  2474 /  2480 =  99.758
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 12826 / 19962 =  64.252
Robust  acc:  4357 /  7535 =  57.823
------------------------------------
Epoch:  25 | Train Loss: 0.002 | Train Acc: 94.450 | Val Loss: 0.003 | Val Acc: 93.079
Training:
Accuracies by groups:
0, 0  acc:  9008 / 14586 =  61.758
0, 1  acc:  4780 /  6618 =  72.227
1, 0  acc: 131768 / 132988 =  99.083
1, 1  acc:  8180 /  8578 =  95.360
--------------------------------------
Average acc: 153736 / 162770 =  94.450
Robust  acc:  9008 / 14586 =  61.758
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7814 /  8535 =  91.552
0, 1  acc:  7952 /  8276 =  96.085
1, 0  acc:  2599 /  2874 =  90.431
1, 1  acc:   127 /   182 =  69.780
------------------------------------
Average acc: 18492 / 19867 =  93.079
Robust  acc:   127 /   182 =  69.780
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 93.969
Robust Acc: 69.444 | Best Acc: 96.138
-------------------------------------
Training, Epoch 24:
Accuracies by groups:
0, 0  acc:  9173 /  9767 =  93.918
0, 1  acc:  7244 /  7535 =  96.138
1, 0  acc:  2216 /  2480 =  89.355
1, 1  acc:   125 /   180 =  69.444
------------------------------------
Average acc: 18758 / 19962 =  93.969
Robust  acc:   125 /   180 =  69.444
------------------------------------
Accuracies by groups:
0, 0  acc:  9173 /  9767 =  93.918
0, 1  acc:  7244 /  7535 =  96.138
1, 0  acc:  2216 /  2480 =  89.355
1, 1  acc:   125 /   180 =  69.444
------------------------------------
Average acc: 18758 / 19962 =  93.969
Robust  acc:   125 /   180 =  69.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9173 /  9767 =  93.918
0, 1  acc:  7244 /  7535 =  96.138
1, 0  acc:  2216 /  2480 =  89.355
1, 1  acc:   125 /   180 =  69.444
------------------------------------
Average acc: 18758 / 19962 =  93.969
Robust  acc:   125 /   180 =  69.444
------------------------------------
Epoch:  26 | Train Loss: 0.002 | Train Acc: 94.326 | Val Loss: 0.006 | Val Acc: 66.603
Training:
Accuracies by groups:
0, 0  acc:  8775 / 14469 =  60.647
0, 1  acc:  4660 /  6620 =  70.393
1, 0  acc: 131715 / 132909 =  99.102
1, 1  acc:  8385 /  8772 =  95.588
--------------------------------------
Average acc: 153535 / 162770 =  94.326
Robust  acc:  8775 / 14469 =  60.647
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4857 /  8535 =  56.907
0, 1  acc:  5333 /  8276 =  64.439
1, 0  acc:  2863 /  2874 =  99.617
1, 1  acc:   179 /   182 =  98.352
------------------------------------
Average acc: 13232 / 19867 =  66.603
Robust  acc:  4857 /  8535 =  56.907
------------------------------------
-------------------------------------------
Avg Test Loss: 0.006 | Avg Test Acc: 67.799
Robust Acc: 61.759 | Best Acc: 99.718
-------------------------------------
Training, Epoch 25:
Accuracies by groups:
0, 0  acc:  6032 /  9767 =  61.759
0, 1  acc:  4852 /  7535 =  64.393
1, 0  acc:  2473 /  2480 =  99.718
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 13534 / 19962 =  67.799
Robust  acc:  6032 /  9767 =  61.759
------------------------------------
Accuracies by groups:
0, 0  acc:  6032 /  9767 =  61.759
0, 1  acc:  4852 /  7535 =  64.393
1, 0  acc:  2473 /  2480 =  99.718
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 13534 / 19962 =  67.799
Robust  acc:  6032 /  9767 =  61.759
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6032 /  9767 =  61.759
0, 1  acc:  4852 /  7535 =  64.393
1, 0  acc:  2473 /  2480 =  99.718
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 13534 / 19962 =  67.799
Robust  acc:  6032 /  9767 =  61.759
------------------------------------
Epoch:  27 | Train Loss: 0.002 | Train Acc: 94.177 | Val Loss: 0.007 | Val Acc: 56.898
Training:
Accuracies by groups:
0, 0  acc:  8587 / 14528 =  59.107
0, 1  acc:  4346 /  6412 =  67.779
1, 0  acc: 132111 / 133225 =  99.164
1, 1  acc:  8248 /  8605 =  95.851
--------------------------------------
Average acc: 153292 / 162770 =  94.177
Robust  acc:  8587 / 14528 =  59.107
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  3204 /  8535 =  37.540
0, 1  acc:  5047 /  8276 =  60.984
1, 0  acc:  2871 /  2874 =  99.896
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc: 11304 / 19867 =  56.898
Robust  acc:  3204 /  8535 =  37.540
------------------------------------
-------------------------------------------
Avg Test Loss: 0.006 | Avg Test Acc: 56.588
Robust Acc: 42.316 | Best Acc: 99.919
-------------------------------------
Training, Epoch 26:
Accuracies by groups:
0, 0  acc:  4133 /  9767 =  42.316
0, 1  acc:  4509 /  7535 =  59.841
1, 0  acc:  2478 /  2480 =  99.919
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 11296 / 19962 =  56.588
Robust  acc:  4133 /  9767 =  42.316
------------------------------------
Accuracies by groups:
0, 0  acc:  4133 /  9767 =  42.316
0, 1  acc:  4509 /  7535 =  59.841
1, 0  acc:  2478 /  2480 =  99.919
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 11296 / 19962 =  56.588
Robust  acc:  4133 /  9767 =  42.316
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  4133 /  9767 =  42.316
0, 1  acc:  4509 /  7535 =  59.841
1, 0  acc:  2478 /  2480 =  99.919
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 11296 / 19962 =  56.588
Robust  acc:  4133 /  9767 =  42.316
------------------------------------
Epoch:  28 | Train Loss: 0.002 | Train Acc: 94.079 | Val Loss: 0.005 | Val Acc: 71.969
Training:
Accuracies by groups:
0, 0  acc:  8546 / 14651 =  58.330
0, 1  acc:  4482 /  6664 =  67.257
1, 0  acc: 131779 / 132805 =  99.227
1, 1  acc:  8325 /  8650 =  96.243
--------------------------------------
Average acc: 153132 / 162770 =  94.079
Robust  acc:  8546 / 14651 =  58.330
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5487 /  8535 =  64.288
0, 1  acc:  5791 /  8276 =  69.973
1, 0  acc:  2843 /  2874 =  98.921
1, 1  acc:   177 /   182 =  97.253
------------------------------------
Average acc: 14298 / 19867 =  71.969
Robust  acc:  5487 /  8535 =  64.288
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 73.425
Robust Acc: 69.171 | Best Acc: 98.790
-------------------------------------
Training, Epoch 27:
Accuracies by groups:
0, 0  acc:  6824 /  9767 =  69.868
0, 1  acc:  5212 /  7535 =  69.171
1, 0  acc:  2450 /  2480 =  98.790
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 14657 / 19962 =  73.425
Robust  acc:  5212 /  7535 =  69.171
------------------------------------
Accuracies by groups:
0, 0  acc:  6824 /  9767 =  69.868
0, 1  acc:  5212 /  7535 =  69.171
1, 0  acc:  2450 /  2480 =  98.790
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 14657 / 19962 =  73.425
Robust  acc:  5212 /  7535 =  69.171
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6824 /  9767 =  69.868
0, 1  acc:  5212 /  7535 =  69.171
1, 0  acc:  2450 /  2480 =  98.790
1, 1  acc:   171 /   180 =  95.000
------------------------------------
Average acc: 14657 / 19962 =  73.425
Robust  acc:  5212 /  7535 =  69.171
------------------------------------
Epoch:  29 | Train Loss: 0.002 | Train Acc: 94.044 | Val Loss: 0.005 | Val Acc: 76.871
Training:
Accuracies by groups:
0, 0  acc:  8490 / 14655 =  57.932
0, 1  acc:  4236 /  6419 =  65.992
1, 0  acc: 132015 / 133019 =  99.245
1, 1  acc:  8335 /  8677 =  96.059
--------------------------------------
Average acc: 153076 / 162770 =  94.044
Robust  acc:  8490 / 14655 =  57.932
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5721 /  8535 =  67.030
0, 1  acc:  6517 /  8276 =  78.746
1, 0  acc:  2854 /  2874 =  99.304
1, 1  acc:   180 /   182 =  98.901
------------------------------------
Average acc: 15272 / 19867 =  76.871
Robust  acc:  5721 /  8535 =  67.030
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 78.449
Robust Acc: 72.356 | Best Acc: 98.952
-------------------------------------
Training, Epoch 28:
Accuracies by groups:
0, 0  acc:  7067 /  9767 =  72.356
0, 1  acc:  5969 /  7535 =  79.217
1, 0  acc:  2454 /  2480 =  98.952
1, 1  acc:   170 /   180 =  94.444
------------------------------------
Average acc: 15660 / 19962 =  78.449
Robust  acc:  7067 /  9767 =  72.356
------------------------------------
Accuracies by groups:
0, 0  acc:  7067 /  9767 =  72.356
0, 1  acc:  5969 /  7535 =  79.217
1, 0  acc:  2454 /  2480 =  98.952
1, 1  acc:   170 /   180 =  94.444
------------------------------------
Average acc: 15660 / 19962 =  78.449
Robust  acc:  7067 /  9767 =  72.356
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7067 /  9767 =  72.356
0, 1  acc:  5969 /  7535 =  79.217
1, 0  acc:  2454 /  2480 =  98.952
1, 1  acc:   170 /   180 =  94.444
------------------------------------
Average acc: 15660 / 19962 =  78.449
Robust  acc:  7067 /  9767 =  72.356
------------------------------------
Epoch:  30 | Train Loss: 0.002 | Train Acc: 93.941 | Val Loss: 0.005 | Val Acc: 74.294
Training:
Accuracies by groups:
0, 0  acc:  8238 / 14514 =  56.759
0, 1  acc:  4244 /  6479 =  65.504
1, 0  acc: 132093 / 133109 =  99.237
1, 1  acc:  8332 /  8668 =  96.124
--------------------------------------
Average acc: 152907 / 162770 =  93.941
Robust  acc:  8238 / 14514 =  56.759
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5668 /  8535 =  66.409
0, 1  acc:  6057 /  8276 =  73.188
1, 0  acc:  2857 /  2874 =  99.408
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 14760 / 19867 =  74.294
Robust  acc:  5668 /  8535 =  66.409
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 75.594
Robust Acc: 71.772 | Best Acc: 99.153
-------------------------------------
Training, Epoch 29:
Accuracies by groups:
0, 0  acc:  7010 /  9767 =  71.772
0, 1  acc:  5445 /  7535 =  72.263
1, 0  acc:  2459 /  2480 =  99.153
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 15090 / 19962 =  75.594
Robust  acc:  7010 /  9767 =  71.772
------------------------------------
Accuracies by groups:
0, 0  acc:  7010 /  9767 =  71.772
0, 1  acc:  5445 /  7535 =  72.263
1, 0  acc:  2459 /  2480 =  99.153
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 15090 / 19962 =  75.594
Robust  acc:  7010 /  9767 =  71.772
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7010 /  9767 =  71.772
0, 1  acc:  5445 /  7535 =  72.263
1, 0  acc:  2459 /  2480 =  99.153
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 15090 / 19962 =  75.594
Robust  acc:  7010 /  9767 =  71.772
------------------------------------
Epoch:  31 | Train Loss: 0.002 | Train Acc: 93.955 | Val Loss: 0.013 | Val Acc: 18.815
Training:
Accuracies by groups:
0, 0  acc:  8080 / 14378 =  56.197
0, 1  acc:  4248 /  6486 =  65.495
1, 0  acc: 132349 / 133340 =  99.257
1, 1  acc:  8254 /  8566 =  96.358
--------------------------------------
Average acc: 152931 / 162770 =  93.955
Robust  acc:  8080 / 14378 =  56.197
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:    81 /  8535 =   0.949
0, 1  acc:   601 /  8276 =   7.262
1, 0  acc:  2874 /  2874 = 100.000
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc:  3738 / 19867 =  18.815
Robust  acc:    81 /  8535 =   0.949
------------------------------------
-------------------------------------------
Avg Test Loss: 0.013 | Avg Test Acc: 16.346
Robust Acc: 1.300 | Best Acc: 100.000
-------------------------------------
Training, Epoch 30:
Accuracies by groups:
0, 0  acc:   127 /  9767 =   1.300
0, 1  acc:   477 /  7535 =   6.330
1, 0  acc:  2480 /  2480 = 100.000
1, 1  acc:   179 /   180 =  99.444
------------------------------------
Average acc:  3263 / 19962 =  16.346
Robust  acc:   127 /  9767 =   1.300
------------------------------------
Accuracies by groups:
0, 0  acc:   127 /  9767 =   1.300
0, 1  acc:   477 /  7535 =   6.330
1, 0  acc:  2480 /  2480 = 100.000
1, 1  acc:   179 /   180 =  99.444
------------------------------------
Average acc:  3263 / 19962 =  16.346
Robust  acc:   127 /  9767 =   1.300
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:   127 /  9767 =   1.300
0, 1  acc:   477 /  7535 =   6.330
1, 0  acc:  2480 /  2480 = 100.000
1, 1  acc:   179 /   180 =  99.444
------------------------------------
Average acc:  3263 / 19962 =  16.346
Robust  acc:   127 /  9767 =   1.300
------------------------------------
Epoch:  32 | Train Loss: 0.002 | Train Acc: 93.826 | Val Loss: 0.006 | Val Acc: 66.950
Training:
Accuracies by groups:
0, 0  acc:  8194 / 14586 =  56.177
0, 1  acc:  4262 /  6605 =  64.527
1, 0  acc: 131966 / 132974 =  99.242
1, 1  acc:  8298 /  8605 =  96.432
--------------------------------------
Average acc: 152720 / 162770 =  93.826
Robust  acc:  8194 / 14586 =  56.177
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5139 /  8535 =  60.211
0, 1  acc:  5120 /  8276 =  61.866
1, 0  acc:  2861 /  2874 =  99.548
1, 1  acc:   181 /   182 =  99.451
------------------------------------
Average acc: 13301 / 19867 =  66.950
Robust  acc:  5139 /  8535 =  60.211
------------------------------------
-------------------------------------------
Avg Test Loss: 0.006 | Avg Test Acc: 68.019
Robust Acc: 60.836 | Best Acc: 99.194
-------------------------------------
Training, Epoch 31:
Accuracies by groups:
0, 0  acc:  6356 /  9767 =  65.076
0, 1  acc:  4584 /  7535 =  60.836
1, 0  acc:  2460 /  2480 =  99.194
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 13578 / 19962 =  68.019
Robust  acc:  4584 /  7535 =  60.836
------------------------------------
Accuracies by groups:
0, 0  acc:  6356 /  9767 =  65.076
0, 1  acc:  4584 /  7535 =  60.836
1, 0  acc:  2460 /  2480 =  99.194
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 13578 / 19962 =  68.019
Robust  acc:  4584 /  7535 =  60.836
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6356 /  9767 =  65.076
0, 1  acc:  4584 /  7535 =  60.836
1, 0  acc:  2460 /  2480 =  99.194
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 13578 / 19962 =  68.019
Robust  acc:  4584 /  7535 =  60.836
------------------------------------
Epoch:  33 | Train Loss: 0.002 | Train Acc: 93.961 | Val Loss: 0.003 | Val Acc: 90.044
Training:
Accuracies by groups:
0, 0  acc:  8250 / 14431 =  57.169
0, 1  acc:  4150 /  6492 =  63.925
1, 0  acc: 132265 / 133256 =  99.256
1, 1  acc:  8276 /  8591 =  96.333
--------------------------------------
Average acc: 152941 / 162770 =  93.961
Robust  acc:  8250 / 14431 =  57.169
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7220 /  8535 =  84.593
0, 1  acc:  7789 /  8276 =  94.116
1, 0  acc:  2733 /  2874 =  95.094
1, 1  acc:   147 /   182 =  80.769
------------------------------------
Average acc: 17889 / 19867 =  90.044
Robust  acc:   147 /   182 =  80.769
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 91.173
Robust Acc: 78.333 | Best Acc: 95.081
-------------------------------------
Training, Epoch 32:
Accuracies by groups:
0, 0  acc:  8619 /  9767 =  88.246
0, 1  acc:  7082 /  7535 =  93.988
1, 0  acc:  2358 /  2480 =  95.081
1, 1  acc:   141 /   180 =  78.333
------------------------------------
Average acc: 18200 / 19962 =  91.173
Robust  acc:   141 /   180 =  78.333
------------------------------------
Accuracies by groups:
0, 0  acc:  8619 /  9767 =  88.246
0, 1  acc:  7082 /  7535 =  93.988
1, 0  acc:  2358 /  2480 =  95.081
1, 1  acc:   141 /   180 =  78.333
------------------------------------
Average acc: 18200 / 19962 =  91.173
Robust  acc:   141 /   180 =  78.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8619 /  9767 =  88.246
0, 1  acc:  7082 /  7535 =  93.988
1, 0  acc:  2358 /  2480 =  95.081
1, 1  acc:   141 /   180 =  78.333
------------------------------------
Average acc: 18200 / 19962 =  91.173
Robust  acc:   141 /   180 =  78.333
------------------------------------
Epoch:  34 | Train Loss: 0.002 | Train Acc: 93.936 | Val Loss: 0.007 | Val Acc: 62.601
Training:
Accuracies by groups:
0, 0  acc:  8200 / 14523 =  56.462
0, 1  acc:  4281 /  6577 =  65.090
1, 0  acc: 132159 / 133093 =  99.298
1, 1  acc:  8259 /  8577 =  96.292
--------------------------------------
Average acc: 152899 / 162770 =  93.936
Robust  acc:  8200 / 14523 =  56.462
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4430 /  8535 =  51.904
0, 1  acc:  4960 /  8276 =  59.932
1, 0  acc:  2866 /  2874 =  99.722
1, 1  acc:   181 /   182 =  99.451
------------------------------------
Average acc: 12437 / 19867 =  62.601
Robust  acc:  4430 /  8535 =  51.904
------------------------------------
-------------------------------------------
Avg Test Loss: 0.006 | Avg Test Acc: 63.601
Robust Acc: 57.172 | Best Acc: 99.516
-------------------------------------
Training, Epoch 33:
Accuracies by groups:
0, 0  acc:  5584 /  9767 =  57.172
0, 1  acc:  4467 /  7535 =  59.283
1, 0  acc:  2468 /  2480 =  99.516
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 12696 / 19962 =  63.601
Robust  acc:  5584 /  9767 =  57.172
------------------------------------
Accuracies by groups:
0, 0  acc:  5584 /  9767 =  57.172
0, 1  acc:  4467 /  7535 =  59.283
1, 0  acc:  2468 /  2480 =  99.516
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 12696 / 19962 =  63.601
Robust  acc:  5584 /  9767 =  57.172
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  5584 /  9767 =  57.172
0, 1  acc:  4467 /  7535 =  59.283
1, 0  acc:  2468 /  2480 =  99.516
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 12696 / 19962 =  63.601
Robust  acc:  5584 /  9767 =  57.172
------------------------------------
Epoch:  35 | Train Loss: 0.002 | Train Acc: 93.982 | Val Loss: 0.004 | Val Acc: 83.037
Training:
Accuracies by groups:
0, 0  acc:  8350 / 14577 =  57.282
0, 1  acc:  4152 /  6431 =  64.562
1, 0  acc: 132090 / 133036 =  99.289
1, 1  acc:  8382 /  8726 =  96.058
--------------------------------------
Average acc: 152974 / 162770 =  93.982
Robust  acc:  8350 / 14577 =  57.282
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6608 /  8535 =  77.422
0, 1  acc:  6877 /  8276 =  83.096
1, 0  acc:  2835 /  2874 =  98.643
1, 1  acc:   177 /   182 =  97.253
------------------------------------
Average acc: 16497 / 19867 =  83.037
Robust  acc:  6608 /  8535 =  77.422
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 84.425
Robust Acc: 81.386 | Best Acc: 97.984
-------------------------------------
Training, Epoch 34:
Accuracies by groups:
0, 0  acc:  7949 /  9767 =  81.386
0, 1  acc:  6307 /  7535 =  83.703
1, 0  acc:  2430 /  2480 =  97.984
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16853 / 19962 =  84.425
Robust  acc:  7949 /  9767 =  81.386
------------------------------------
Accuracies by groups:
0, 0  acc:  7949 /  9767 =  81.386
0, 1  acc:  6307 /  7535 =  83.703
1, 0  acc:  2430 /  2480 =  97.984
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16853 / 19962 =  84.425
Robust  acc:  7949 /  9767 =  81.386
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7949 /  9767 =  81.386
0, 1  acc:  6307 /  7535 =  83.703
1, 0  acc:  2430 /  2480 =  97.984
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 16853 / 19962 =  84.425
Robust  acc:  7949 /  9767 =  81.386
------------------------------------
Epoch:  36 | Train Loss: 0.002 | Train Acc: 93.976 | Val Loss: 0.005 | Val Acc: 74.938
Training:
Accuracies by groups:
0, 0  acc:  8349 / 14568 =  57.311
0, 1  acc:  4164 /  6474 =  64.319
1, 0  acc: 132125 / 133081 =  99.282
1, 1  acc:  8326 /  8647 =  96.288
--------------------------------------
Average acc: 152964 / 162770 =  93.976
Robust  acc:  8349 / 14568 =  57.311
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5735 /  8535 =  67.194
0, 1  acc:  6119 /  8276 =  73.937
1, 0  acc:  2855 /  2874 =  99.339
1, 1  acc:   179 /   182 =  98.352
------------------------------------
Average acc: 14888 / 19867 =  74.938
Robust  acc:  5735 /  8535 =  67.194
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 76.891
Robust Acc: 73.390 | Best Acc: 98.952
-------------------------------------
Training, Epoch 35:
Accuracies by groups:
0, 0  acc:  7168 /  9767 =  73.390
0, 1  acc:  5551 /  7535 =  73.670
1, 0  acc:  2454 /  2480 =  98.952
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 15349 / 19962 =  76.891
Robust  acc:  7168 /  9767 =  73.390
------------------------------------
Accuracies by groups:
0, 0  acc:  7168 /  9767 =  73.390
0, 1  acc:  5551 /  7535 =  73.670
1, 0  acc:  2454 /  2480 =  98.952
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 15349 / 19962 =  76.891
Robust  acc:  7168 /  9767 =  73.390
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7168 /  9767 =  73.390
0, 1  acc:  5551 /  7535 =  73.670
1, 0  acc:  2454 /  2480 =  98.952
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 15349 / 19962 =  76.891
Robust  acc:  7168 /  9767 =  73.390
------------------------------------
Epoch:  37 | Train Loss: 0.002 | Train Acc: 94.000 | Val Loss: 0.009 | Val Acc: 39.327
Training:
Accuracies by groups:
0, 0  acc:  8263 / 14481 =  57.061
0, 1  acc:  4184 /  6507 =  64.300
1, 0  acc: 132241 / 133162 =  99.308
1, 1  acc:  8316 /  8620 =  96.473
--------------------------------------
Average acc: 153004 / 162770 =  94.000
Robust  acc:  8263 / 14481 =  57.061
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  2281 /  8535 =  26.725
0, 1  acc:  2476 /  8276 =  29.918
1, 0  acc:  2874 /  2874 = 100.000
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc:  7813 / 19867 =  39.327
Robust  acc:  2281 /  8535 =  26.725
------------------------------------
-------------------------------------------
Avg Test Loss: 0.009 | Avg Test Acc: 38.418
Robust Acc: 28.149 | Best Acc: 99.960
-------------------------------------
Training, Epoch 36:
Accuracies by groups:
0, 0  acc:  2891 /  9767 =  29.600
0, 1  acc:  2121 /  7535 =  28.149
1, 0  acc:  2479 /  2480 =  99.960
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  7669 / 19962 =  38.418
Robust  acc:  2121 /  7535 =  28.149
------------------------------------
Accuracies by groups:
0, 0  acc:  2891 /  9767 =  29.600
0, 1  acc:  2121 /  7535 =  28.149
1, 0  acc:  2479 /  2480 =  99.960
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  7669 / 19962 =  38.418
Robust  acc:  2121 /  7535 =  28.149
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  2891 /  9767 =  29.600
0, 1  acc:  2121 /  7535 =  28.149
1, 0  acc:  2479 /  2480 =  99.960
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc:  7669 / 19962 =  38.418
Robust  acc:  2121 /  7535 =  28.149
------------------------------------
Epoch:  38 | Train Loss: 0.002 | Train Acc: 93.893 | Val Loss: 0.005 | Val Acc: 78.885
Training:
Accuracies by groups:
0, 0  acc:  8351 / 14620 =  57.120
0, 1  acc:  4152 /  6515 =  63.730
1, 0  acc: 131981 / 132979 =  99.250
1, 1  acc:  8346 /  8656 =  96.419
--------------------------------------
Average acc: 152830 / 162770 =  93.893
Robust  acc:  8351 / 14620 =  57.120
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6346 /  8535 =  74.353
0, 1  acc:  6320 /  8276 =  76.365
1, 0  acc:  2829 /  2874 =  98.434
1, 1  acc:   177 /   182 =  97.253
------------------------------------
Average acc: 15672 / 19867 =  78.885
Robust  acc:  6346 /  8535 =  74.353
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 80.408
Robust Acc: 76.045 | Best Acc: 97.823
-------------------------------------
Training, Epoch 37:
Accuracies by groups:
0, 0  acc:  7726 /  9767 =  79.103
0, 1  acc:  5730 /  7535 =  76.045
1, 0  acc:  2426 /  2480 =  97.823
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 16051 / 19962 =  80.408
Robust  acc:  5730 /  7535 =  76.045
------------------------------------
Accuracies by groups:
0, 0  acc:  7726 /  9767 =  79.103
0, 1  acc:  5730 /  7535 =  76.045
1, 0  acc:  2426 /  2480 =  97.823
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 16051 / 19962 =  80.408
Robust  acc:  5730 /  7535 =  76.045
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7726 /  9767 =  79.103
0, 1  acc:  5730 /  7535 =  76.045
1, 0  acc:  2426 /  2480 =  97.823
1, 1  acc:   169 /   180 =  93.889
------------------------------------
Average acc: 16051 / 19962 =  80.408
Robust  acc:  5730 /  7535 =  76.045
------------------------------------
Epoch:  39 | Train Loss: 0.002 | Train Acc: 94.002 | Val Loss: 0.007 | Val Acc: 57.336
Training:
Accuracies by groups:
0, 0  acc:  8178 / 14391 =  56.827
0, 1  acc:  4255 /  6500 =  65.462
1, 0  acc: 132246 / 133225 =  99.265
1, 1  acc:  8328 /  8654 =  96.233
--------------------------------------
Average acc: 153007 / 162770 =  94.002
Robust  acc:  8178 / 14391 =  56.827
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  4616 /  8535 =  54.083
0, 1  acc:  3729 /  8276 =  45.058
1, 0  acc:  2865 /  2874 =  99.687
1, 1  acc:   181 /   182 =  99.451
------------------------------------
Average acc: 11391 / 19867 =  57.336
Robust  acc:  3729 /  8276 =  45.058
------------------------------------
-------------------------------------------
Avg Test Loss: 0.007 | Avg Test Acc: 59.147
Robust Acc: 44.393 | Best Acc: 99.476
-------------------------------------
Training, Epoch 38:
Accuracies by groups:
0, 0  acc:  5817 /  9767 =  59.558
0, 1  acc:  3345 /  7535 =  44.393
1, 0  acc:  2467 /  2480 =  99.476
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 11807 / 19962 =  59.147
Robust  acc:  3345 /  7535 =  44.393
------------------------------------
Accuracies by groups:
0, 0  acc:  5817 /  9767 =  59.558
0, 1  acc:  3345 /  7535 =  44.393
1, 0  acc:  2467 /  2480 =  99.476
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 11807 / 19962 =  59.147
Robust  acc:  3345 /  7535 =  44.393
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  5817 /  9767 =  59.558
0, 1  acc:  3345 /  7535 =  44.393
1, 0  acc:  2467 /  2480 =  99.476
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 11807 / 19962 =  59.147
Robust  acc:  3345 /  7535 =  44.393
------------------------------------
Epoch:  40 | Train Loss: 0.002 | Train Acc: 94.089 | Val Loss: 0.007 | Val Acc: 54.920
Training:
Accuracies by groups:
0, 0  acc:  8312 / 14391 =  57.758
0, 1  acc:  4292 /  6540 =  65.627
1, 0  acc: 132163 / 133127 =  99.276
1, 1  acc:  8382 /  8712 =  96.212
--------------------------------------
Average acc: 153149 / 162770 =  94.089
Robust  acc:  8312 / 14391 =  57.758
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  3846 /  8535 =  45.062
0, 1  acc:  4014 /  8276 =  48.502
1, 0  acc:  2870 /  2874 =  99.861
1, 1  acc:   181 /   182 =  99.451
------------------------------------
Average acc: 10911 / 19867 =  54.920
Robust  acc:  3846 /  8535 =  45.062
------------------------------------
-------------------------------------------
Avg Test Loss: 0.007 | Avg Test Acc: 55.265
Robust Acc: 47.113 | Best Acc: 99.758
-------------------------------------
Training, Epoch 39:
Accuracies by groups:
0, 0  acc:  4830 /  9767 =  49.452
0, 1  acc:  3550 /  7535 =  47.113
1, 0  acc:  2474 /  2480 =  99.758
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 11032 / 19962 =  55.265
Robust  acc:  3550 /  7535 =  47.113
------------------------------------
Accuracies by groups:
0, 0  acc:  4830 /  9767 =  49.452
0, 1  acc:  3550 /  7535 =  47.113
1, 0  acc:  2474 /  2480 =  99.758
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 11032 / 19962 =  55.265
Robust  acc:  3550 /  7535 =  47.113
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  4830 /  9767 =  49.452
0, 1  acc:  3550 /  7535 =  47.113
1, 0  acc:  2474 /  2480 =  99.758
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 11032 / 19962 =  55.265
Robust  acc:  3550 /  7535 =  47.113
------------------------------------
Epoch:  41 | Train Loss: 0.002 | Train Acc: 93.995 | Val Loss: 0.006 | Val Acc: 67.881
Training:
Accuracies by groups:
0, 0  acc:  8477 / 14618 =  57.990
0, 1  acc:  4278 /  6532 =  65.493
1, 0  acc: 131931 / 132972 =  99.217
1, 1  acc:  8310 /  8648 =  96.092
--------------------------------------
Average acc: 152996 / 162770 =  93.995
Robust  acc:  8477 / 14618 =  57.990
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5145 /  8535 =  60.281
0, 1  acc:  5303 /  8276 =  64.077
1, 0  acc:  2858 /  2874 =  99.443
1, 1  acc:   180 /   182 =  98.901
------------------------------------
Average acc: 13486 / 19867 =  67.881
Robust  acc:  5145 /  8535 =  60.281
------------------------------------
-------------------------------------------
Avg Test Loss: 0.006 | Avg Test Acc: 69.222
Robust Acc: 62.933 | Best Acc: 99.153
-------------------------------------
Training, Epoch 40:
Accuracies by groups:
0, 0  acc:  6441 /  9767 =  65.947
0, 1  acc:  4742 /  7535 =  62.933
1, 0  acc:  2459 /  2480 =  99.153
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 13818 / 19962 =  69.222
Robust  acc:  4742 /  7535 =  62.933
------------------------------------
Accuracies by groups:
0, 0  acc:  6441 /  9767 =  65.947
0, 1  acc:  4742 /  7535 =  62.933
1, 0  acc:  2459 /  2480 =  99.153
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 13818 / 19962 =  69.222
Robust  acc:  4742 /  7535 =  62.933
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6441 /  9767 =  65.947
0, 1  acc:  4742 /  7535 =  62.933
1, 0  acc:  2459 /  2480 =  99.153
1, 1  acc:   176 /   180 =  97.778
------------------------------------
Average acc: 13818 / 19962 =  69.222
Robust  acc:  4742 /  7535 =  62.933
------------------------------------
Epoch:  42 | Train Loss: 0.002 | Train Acc: 94.149 | Val Loss: 0.007 | Val Acc: 55.590
Training:
Accuracies by groups:
0, 0  acc:  8402 / 14299 =  58.759
0, 1  acc:  4357 /  6556 =  66.458
1, 0  acc: 132125 / 133217 =  99.180
1, 1  acc:  8363 /  8698 =  96.149
--------------------------------------
Average acc: 153247 / 162770 =  94.149
Robust  acc:  8402 / 14299 =  58.759
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  3938 /  8535 =  46.139
0, 1  acc:  4055 /  8276 =  48.997
1, 0  acc:  2869 /  2874 =  99.826
1, 1  acc:   182 /   182 = 100.000
------------------------------------
Average acc: 11044 / 19867 =  55.590
Robust  acc:  3938 /  8535 =  46.139
------------------------------------
-------------------------------------------
Avg Test Loss: 0.007 | Avg Test Acc: 56.608
Robust Acc: 47.206 | Best Acc: 99.879
-------------------------------------
Training, Epoch 41:
Accuracies by groups:
0, 0  acc:  5088 /  9767 =  52.094
0, 1  acc:  3557 /  7535 =  47.206
1, 0  acc:  2477 /  2480 =  99.879
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 11300 / 19962 =  56.608
Robust  acc:  3557 /  7535 =  47.206
------------------------------------
Accuracies by groups:
0, 0  acc:  5088 /  9767 =  52.094
0, 1  acc:  3557 /  7535 =  47.206
1, 0  acc:  2477 /  2480 =  99.879
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 11300 / 19962 =  56.608
Robust  acc:  3557 /  7535 =  47.206
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  5088 /  9767 =  52.094
0, 1  acc:  3557 /  7535 =  47.206
1, 0  acc:  2477 /  2480 =  99.879
1, 1  acc:   178 /   180 =  98.889
------------------------------------
Average acc: 11300 / 19962 =  56.608
Robust  acc:  3557 /  7535 =  47.206
------------------------------------
Epoch:  43 | Train Loss: 0.002 | Train Acc: 94.234 | Val Loss: 0.004 | Val Acc: 90.970
Training:
Accuracies by groups:
0, 0  acc:  8620 / 14504 =  59.432
0, 1  acc:  4354 /  6487 =  67.119
1, 0  acc: 132076 / 133124 =  99.213
1, 1  acc:  8335 /  8655 =  96.303
--------------------------------------
Average acc: 153385 / 162770 =  94.234
Robust  acc:  8620 / 14504 =  59.432
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7310 /  8535 =  85.647
0, 1  acc:  7887 /  8276 =  95.300
1, 0  acc:  2741 /  2874 =  95.372
1, 1  acc:   135 /   182 =  74.176
------------------------------------
Average acc: 18073 / 19867 =  90.970
Robust  acc:   135 /   182 =  74.176
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 92.035
Robust Acc: 68.889 | Best Acc: 95.780
-------------------------------------
Training, Epoch 42:
Accuracies by groups:
0, 0  acc:  8686 /  9767 =  88.932
0, 1  acc:  7217 /  7535 =  95.780
1, 0  acc:  2345 /  2480 =  94.556
1, 1  acc:   124 /   180 =  68.889
------------------------------------
Average acc: 18372 / 19962 =  92.035
Robust  acc:   124 /   180 =  68.889
------------------------------------
Accuracies by groups:
0, 0  acc:  8686 /  9767 =  88.932
0, 1  acc:  7217 /  7535 =  95.780
1, 0  acc:  2345 /  2480 =  94.556
1, 1  acc:   124 /   180 =  68.889
------------------------------------
Average acc: 18372 / 19962 =  92.035
Robust  acc:   124 /   180 =  68.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8686 /  9767 =  88.932
0, 1  acc:  7217 /  7535 =  95.780
1, 0  acc:  2345 /  2480 =  94.556
1, 1  acc:   124 /   180 =  68.889
------------------------------------
Average acc: 18372 / 19962 =  92.035
Robust  acc:   124 /   180 =  68.889
------------------------------------
Epoch:  44 | Train Loss: 0.002 | Train Acc: 94.175 | Val Loss: 0.005 | Val Acc: 74.284
Training:
Accuracies by groups:
0, 0  acc:  8546 / 14415 =  59.285
0, 1  acc:  4447 /  6588 =  67.502
1, 0  acc: 132092 / 133222 =  99.152
1, 1  acc:  8203 /  8545 =  95.998
--------------------------------------
Average acc: 153288 / 162770 =  94.175
Robust  acc:  8546 / 14415 =  59.285
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5628 /  8535 =  65.940
0, 1  acc:  6108 /  8276 =  73.804
1, 0  acc:  2846 /  2874 =  99.026
1, 1  acc:   176 /   182 =  96.703
------------------------------------
Average acc: 14758 / 19867 =  74.284
Robust  acc:  5628 /  8535 =  65.940
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 75.819
Robust Acc: 71.660 | Best Acc: 98.871
-------------------------------------
Training, Epoch 43:
Accuracies by groups:
0, 0  acc:  6999 /  9767 =  71.660
0, 1  acc:  5511 /  7535 =  73.139
1, 0  acc:  2452 /  2480 =  98.871
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15135 / 19962 =  75.819
Robust  acc:  6999 /  9767 =  71.660
------------------------------------
Accuracies by groups:
0, 0  acc:  6999 /  9767 =  71.660
0, 1  acc:  5511 /  7535 =  73.139
1, 0  acc:  2452 /  2480 =  98.871
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15135 / 19962 =  75.819
Robust  acc:  6999 /  9767 =  71.660
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6999 /  9767 =  71.660
0, 1  acc:  5511 /  7535 =  73.139
1, 0  acc:  2452 /  2480 =  98.871
1, 1  acc:   173 /   180 =  96.111
------------------------------------
Average acc: 15135 / 19962 =  75.819
Robust  acc:  6999 /  9767 =  71.660
------------------------------------
Epoch:  45 | Train Loss: 0.002 | Train Acc: 94.110 | Val Loss: 0.005 | Val Acc: 78.069
Training:
Accuracies by groups:
0, 0  acc:  8567 / 14463 =  59.234
0, 1  acc:  4367 /  6562 =  66.550
1, 0  acc: 131947 / 133129 =  99.112
1, 1  acc:  8302 /  8616 =  96.356
--------------------------------------
Average acc: 153183 / 162770 =  94.110
Robust  acc:  8567 / 14463 =  59.234
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5907 /  8535 =  69.209
0, 1  acc:  6573 /  8276 =  79.422
1, 0  acc:  2851 /  2874 =  99.200
1, 1  acc:   179 /   182 =  98.352
------------------------------------
Average acc: 15510 / 19867 =  78.069
Robust  acc:  5907 /  8535 =  69.209
------------------------------------
-------------------------------------------
Avg Test Loss: 0.005 | Avg Test Acc: 79.556
Robust Acc: 74.475 | Best Acc: 99.113
-------------------------------------
Training, Epoch 44:
Accuracies by groups:
0, 0  acc:  7274 /  9767 =  74.475
0, 1  acc:  5975 /  7535 =  79.297
1, 0  acc:  2458 /  2480 =  99.113
1, 1  acc:   174 /   180 =  96.667
------------------------------------
Average acc: 15881 / 19962 =  79.556
Robust  acc:  7274 /  9767 =  74.475
------------------------------------
Accuracies by groups:
0, 0  acc:  7274 /  9767 =  74.475
0, 1  acc:  5975 /  7535 =  79.297
1, 0  acc:  2458 /  2480 =  99.113
1, 1  acc:   174 /   180 =  96.667
------------------------------------
Average acc: 15881 / 19962 =  79.556
Robust  acc:  7274 /  9767 =  74.475
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7274 /  9767 =  74.475
0, 1  acc:  5975 /  7535 =  79.297
1, 0  acc:  2458 /  2480 =  99.113
1, 1  acc:   174 /   180 =  96.667
------------------------------------
Average acc: 15881 / 19962 =  79.556
Robust  acc:  7274 /  9767 =  74.475
------------------------------------
Epoch:  46 | Train Loss: 0.002 | Train Acc: 94.176 | Val Loss: 0.004 | Val Acc: 81.492
Training:
Accuracies by groups:
0, 0  acc:  8679 / 14522 =  59.764
0, 1  acc:  4391 /  6546 =  67.079
1, 0  acc: 131899 / 133036 =  99.145
1, 1  acc:  8321 /  8666 =  96.019
--------------------------------------
Average acc: 153290 / 162770 =  94.176
Robust  acc:  8679 / 14522 =  59.764
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6575 /  8535 =  77.036
0, 1  acc:  6641 /  8276 =  80.244
1, 0  acc:  2800 /  2874 =  97.425
1, 1  acc:   174 /   182 =  95.604
------------------------------------
Average acc: 16190 / 19867 =  81.492
Robust  acc:  6575 /  8535 =  77.036
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 83.008
Robust Acc: 80.027 | Best Acc: 97.137
-------------------------------------
Training, Epoch 45:
Accuracies by groups:
0, 0  acc:  7969 /  9767 =  81.591
0, 1  acc:  6030 /  7535 =  80.027
1, 0  acc:  2409 /  2480 =  97.137
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 16570 / 19962 =  83.008
Robust  acc:  6030 /  7535 =  80.027
------------------------------------
Accuracies by groups:
0, 0  acc:  7969 /  9767 =  81.591
0, 1  acc:  6030 /  7535 =  80.027
1, 0  acc:  2409 /  2480 =  97.137
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 16570 / 19962 =  83.008
Robust  acc:  6030 /  7535 =  80.027
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7969 /  9767 =  81.591
0, 1  acc:  6030 /  7535 =  80.027
1, 0  acc:  2409 /  2480 =  97.137
1, 1  acc:   162 /   180 =  90.000
------------------------------------
Average acc: 16570 / 19962 =  83.008
Robust  acc:  6030 /  7535 =  80.027
------------------------------------
Epoch:  47 | Train Loss: 0.002 | Train Acc: 94.192 | Val Loss: 0.007 | Val Acc: 56.773
Training:
Accuracies by groups:
0, 0  acc:  8586 / 14421 =  59.538
0, 1  acc:  4424 /  6520 =  67.853
1, 0  acc: 132019 / 133206 =  99.109
1, 1  acc:  8287 /  8623 =  96.103
--------------------------------------
Average acc: 153316 / 162770 =  94.192
Robust  acc:  8586 / 14421 =  59.538
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  3776 /  8535 =  44.241
0, 1  acc:  4452 /  8276 =  53.794
1, 0  acc:  2870 /  2874 =  99.861
1, 1  acc:   181 /   182 =  99.451
------------------------------------
Average acc: 11279 / 19867 =  56.773
Robust  acc:  3776 /  8535 =  44.241
------------------------------------
-------------------------------------------
Avg Test Loss: 0.007 | Avg Test Acc: 57.905
Robust Acc: 50.732 | Best Acc: 99.879
-------------------------------------
Training, Epoch 46:
Accuracies by groups:
0, 0  acc:  4955 /  9767 =  50.732
0, 1  acc:  3950 /  7535 =  52.422
1, 0  acc:  2477 /  2480 =  99.879
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 11559 / 19962 =  57.905
Robust  acc:  4955 /  9767 =  50.732
------------------------------------
Accuracies by groups:
0, 0  acc:  4955 /  9767 =  50.732
0, 1  acc:  3950 /  7535 =  52.422
1, 0  acc:  2477 /  2480 =  99.879
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 11559 / 19962 =  57.905
Robust  acc:  4955 /  9767 =  50.732
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  4955 /  9767 =  50.732
0, 1  acc:  3950 /  7535 =  52.422
1, 0  acc:  2477 /  2480 =  99.879
1, 1  acc:   177 /   180 =  98.333
------------------------------------
Average acc: 11559 / 19962 =  57.905
Robust  acc:  4955 /  9767 =  50.732
------------------------------------
Epoch:  48 | Train Loss: 0.002 | Train Acc: 94.242 | Val Loss: 0.003 | Val Acc: 91.896
Training:
Accuracies by groups:
0, 0  acc:  8596 / 14404 =  59.678
0, 1  acc:  4418 /  6480 =  68.179
1, 0  acc: 131980 / 133137 =  99.131
1, 1  acc:  8404 /  8749 =  96.057
--------------------------------------
Average acc: 153398 / 162770 =  94.242
Robust  acc:  8596 / 14404 =  59.678
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7696 /  8535 =  90.170
0, 1  acc:  7801 /  8276 =  94.261
1, 0  acc:  2614 /  2874 =  90.953
1, 1  acc:   146 /   182 =  80.220
------------------------------------
Average acc: 18257 / 19867 =  91.896
Robust  acc:   146 /   182 =  80.220
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 92.932
Robust Acc: 76.111 | Best Acc: 94.851
-------------------------------------
Training, Epoch 47:
Accuracies by groups:
0, 0  acc:  9034 /  9767 =  92.495
0, 1  acc:  7147 /  7535 =  94.851
1, 0  acc:  2233 /  2480 =  90.040
1, 1  acc:   137 /   180 =  76.111
------------------------------------
Average acc: 18551 / 19962 =  92.932
Robust  acc:   137 /   180 =  76.111
------------------------------------
Accuracies by groups:
0, 0  acc:  9034 /  9767 =  92.495
0, 1  acc:  7147 /  7535 =  94.851
1, 0  acc:  2233 /  2480 =  90.040
1, 1  acc:   137 /   180 =  76.111
------------------------------------
Average acc: 18551 / 19962 =  92.932
Robust  acc:   137 /   180 =  76.111
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9034 /  9767 =  92.495
0, 1  acc:  7147 /  7535 =  94.851
1, 0  acc:  2233 /  2480 =  90.040
1, 1  acc:   137 /   180 =  76.111
------------------------------------
Average acc: 18551 / 19962 =  92.932
Robust  acc:   137 /   180 =  76.111
------------------------------------
Epoch:  49 | Train Loss: 0.002 | Train Acc: 94.290 | Val Loss: 0.003 | Val Acc: 90.069
Training:
Accuracies by groups:
0, 0  acc:  8953 / 14663 =  61.058
0, 1  acc:  4392 /  6444 =  68.156
1, 0  acc: 131728 / 132927 =  99.098
1, 1  acc:  8403 /  8736 =  96.188
--------------------------------------
Average acc: 153476 / 162770 =  94.290
Robust  acc:  8953 / 14663 =  61.058
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7278 /  8535 =  85.272
0, 1  acc:  7756 /  8276 =  93.717
1, 0  acc:  2719 /  2874 =  94.607
1, 1  acc:   141 /   182 =  77.473
------------------------------------
Average acc: 17894 / 19867 =  90.069
Robust  acc:   141 /   182 =  77.473
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 91.193
Robust Acc: 75.000 | Best Acc: 94.280
-------------------------------------
Training, Epoch 48:
Accuracies by groups:
0, 0  acc:  8657 /  9767 =  88.635
0, 1  acc:  7104 /  7535 =  94.280
1, 0  acc:  2308 /  2480 =  93.065
1, 1  acc:   135 /   180 =  75.000
------------------------------------
Average acc: 18204 / 19962 =  91.193
Robust  acc:   135 /   180 =  75.000
------------------------------------
Accuracies by groups:
0, 0  acc:  8657 /  9767 =  88.635
0, 1  acc:  7104 /  7535 =  94.280
1, 0  acc:  2308 /  2480 =  93.065
1, 1  acc:   135 /   180 =  75.000
------------------------------------
Average acc: 18204 / 19962 =  91.193
Robust  acc:   135 /   180 =  75.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8657 /  9767 =  88.635
0, 1  acc:  7104 /  7535 =  94.280
1, 0  acc:  2308 /  2480 =  93.065
1, 1  acc:   135 /   180 =  75.000
------------------------------------
Average acc: 18204 / 19962 =  91.193
Robust  acc:   135 /   180 =  75.000
------------------------------------
Epoch:  50 | Train Loss: 0.002 | Train Acc: 94.073 | Val Loss: 0.003 | Val Acc: 89.636
Training:
Accuracies by groups:
0, 0  acc:  8610 / 14571 =  59.090
0, 1  acc:  4566 /  6641 =  68.755
1, 0  acc: 131672 / 132952 =  99.037
1, 1  acc:  8274 /  8606 =  96.142
--------------------------------------
Average acc: 153122 / 162770 =  94.073
Robust  acc:  8610 / 14571 =  59.090
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7598 /  8535 =  89.022
0, 1  acc:  7538 /  8276 =  91.083
1, 0  acc:  2536 /  2874 =  88.239
1, 1  acc:   136 /   182 =  74.725
------------------------------------
Average acc: 17808 / 19867 =  89.636
Robust  acc:   136 /   182 =  74.725
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 90.382
Robust Acc: 67.222 | Best Acc: 91.241
-------------------------------------
Training, Epoch 49:
Accuracies by groups:
0, 0  acc:  8899 /  9767 =  91.113
0, 1  acc:  6875 /  7535 =  91.241
1, 0  acc:  2147 /  2480 =  86.573
1, 1  acc:   121 /   180 =  67.222
------------------------------------
Average acc: 18042 / 19962 =  90.382
Robust  acc:   121 /   180 =  67.222
------------------------------------
Accuracies by groups:
0, 0  acc:  8899 /  9767 =  91.113
0, 1  acc:  6875 /  7535 =  91.241
1, 0  acc:  2147 /  2480 =  86.573
1, 1  acc:   121 /   180 =  67.222
------------------------------------
Average acc: 18042 / 19962 =  90.382
Robust  acc:   121 /   180 =  67.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8899 /  9767 =  91.113
0, 1  acc:  6875 /  7535 =  91.241
1, 0  acc:  2147 /  2480 =  86.573
1, 1  acc:   121 /   180 =  67.222
------------------------------------
Average acc: 18042 / 19962 =  90.382
Robust  acc:   121 /   180 =  67.222
------------------------------------
replace: True
-> Updating checkpoint debias-end_seed8.pt...
Checkpoint saved at ./model/celebA/config/debias-end_seed8.pt
