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: 32
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: 
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: 0
Replicate: 0
No cuda: False
Resume: False
New slice: False
Num workers: 32
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: False
Cosine: False
Exp: None
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=32-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=0-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
Pretrained model loaded from 
Epoch:   1 | Train Loss: 0.000 | Train Acc: 85.076 | Val Loss: 0.003 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71607 / 71629 =  99.969
0, 1  acc: 66866 / 66874 =  99.988
1, 0  acc:     5 / 22880 =   0.022
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138478 / 162770 =  85.076
Robust  acc:     0 /  1387 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 0:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:   2 | Train Loss: 0.001 | Train Acc: 87.066 | Val Loss: 0.002 | Val Acc: 91.629
Training:
Accuracies by groups:
0, 0  acc: 71303 / 71629 =  99.545
0, 1  acc: 66869 / 66874 =  99.993
1, 0  acc:  3537 / 22880 =  15.459
1, 1  acc:     9 /  1387 =   0.649
--------------------------------------
Average acc: 141718 / 162770 =  87.066
Robust  acc:     9 /  1387 =   0.649
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8274 /  8535 =  96.942
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:  1648 /  2874 =  57.342
1, 1  acc:     6 /   182 =   3.297
------------------------------------
Average acc: 18204 / 19867 =  91.629
Robust  acc:     6 /   182 =   3.297
------------------------------------
New max robust acc: 3.296703296703297
- Saving best checkpoint at epoch 1
replace: True
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=1-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 92.260
Robust Acc: 6.111 | Best Acc: 100.000
-------------------------------------
Training, Epoch 1:
Accuracies by groups:
0, 0  acc:  9592 /  9767 =  98.208
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:  1279 /  2480 =  51.573
1, 1  acc:    11 /   180 =   6.111
------------------------------------
Average acc: 18417 / 19962 =  92.260
Robust  acc:    11 /   180 =   6.111
------------------------------------
Accuracies by groups:
0, 0  acc:  9592 /  9767 =  98.208
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:  1279 /  2480 =  51.573
1, 1  acc:    11 /   180 =   6.111
------------------------------------
Average acc: 18417 / 19962 =  92.260
Robust  acc:    11 /   180 =   6.111
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9592 /  9767 =  98.208
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:  1279 /  2480 =  51.573
1, 1  acc:    11 /   180 =   6.111
------------------------------------
Average acc: 18417 / 19962 =  92.260
Robust  acc:    11 /   180 =   6.111
------------------------------------
Epoch:   3 | Train Loss: 0.000 | Train Acc: 93.090 | Val Loss: 0.001 | Val Acc: 93.960
Training:
Accuracies by groups:
0, 0  acc: 69336 / 71629 =  96.799
0, 1  acc: 66765 / 66874 =  99.837
1, 0  acc: 15238 / 22880 =  66.600
1, 1  acc:   183 /  1387 =  13.194
--------------------------------------
Average acc: 151522 / 162770 =  93.090
Robust  acc:   183 /  1387 =  13.194
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8210 /  8535 =  96.192
0, 1  acc:  8260 /  8276 =  99.807
1, 0  acc:  2175 /  2874 =  75.678
1, 1  acc:    22 /   182 =  12.088
------------------------------------
Average acc: 18667 / 19867 =  93.960
Robust  acc:    22 /   182 =  12.088
------------------------------------
New max robust acc: 12.087912087912088
- Saving best checkpoint at epoch 2
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=1-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=2-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.224
Robust Acc: 17.222 | Best Acc: 99.934
-------------------------------------
Training, Epoch 2:
Accuracies by groups:
0, 0  acc:  9503 /  9767 =  97.297
0, 1  acc:  7530 /  7535 =  99.934
1, 0  acc:  1745 /  2480 =  70.363
1, 1  acc:    31 /   180 =  17.222
------------------------------------
Average acc: 18809 / 19962 =  94.224
Robust  acc:    31 /   180 =  17.222
------------------------------------
Accuracies by groups:
0, 0  acc:  9503 /  9767 =  97.297
0, 1  acc:  7530 /  7535 =  99.934
1, 0  acc:  1745 /  2480 =  70.363
1, 1  acc:    31 /   180 =  17.222
------------------------------------
Average acc: 18809 / 19962 =  94.224
Robust  acc:    31 /   180 =  17.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9503 /  9767 =  97.297
0, 1  acc:  7530 /  7535 =  99.934
1, 0  acc:  1745 /  2480 =  70.363
1, 1  acc:    31 /   180 =  17.222
------------------------------------
Average acc: 18809 / 19962 =  94.224
Robust  acc:    31 /   180 =  17.222
------------------------------------
Epoch:   4 | Train Loss: 0.000 | Train Acc: 94.177 | Val Loss: 0.001 | Val Acc: 94.488
Training:
Accuracies by groups:
0, 0  acc: 68950 / 71629 =  96.260
0, 1  acc: 66621 / 66874 =  99.622
1, 0  acc: 17411 / 22880 =  76.097
1, 1  acc:   310 /  1387 =  22.350
--------------------------------------
Average acc: 153292 / 162770 =  94.177
Robust  acc:   310 /  1387 =  22.350
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8131 /  8535 =  95.267
0, 1  acc:  8252 /  8276 =  99.710
1, 0  acc:  2355 /  2874 =  81.942
1, 1  acc:    34 /   182 =  18.681
------------------------------------
Average acc: 18772 / 19867 =  94.488
Robust  acc:    34 /   182 =  18.681
------------------------------------
New max robust acc: 18.681318681318682
- Saving best checkpoint at epoch 3
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=2-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=3-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.915
Robust Acc: 25.556 | Best Acc: 99.748
-------------------------------------
Training, Epoch 3:
Accuracies by groups:
0, 0  acc:  9448 /  9767 =  96.734
0, 1  acc:  7516 /  7535 =  99.748
1, 0  acc:  1937 /  2480 =  78.105
1, 1  acc:    46 /   180 =  25.556
------------------------------------
Average acc: 18947 / 19962 =  94.915
Robust  acc:    46 /   180 =  25.556
------------------------------------
Accuracies by groups:
0, 0  acc:  9448 /  9767 =  96.734
0, 1  acc:  7516 /  7535 =  99.748
1, 0  acc:  1937 /  2480 =  78.105
1, 1  acc:    46 /   180 =  25.556
------------------------------------
Average acc: 18947 / 19962 =  94.915
Robust  acc:    46 /   180 =  25.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9448 /  9767 =  96.734
0, 1  acc:  7516 /  7535 =  99.748
1, 0  acc:  1937 /  2480 =  78.105
1, 1  acc:    46 /   180 =  25.556
------------------------------------
Average acc: 18947 / 19962 =  94.915
Robust  acc:    46 /   180 =  25.556
------------------------------------
Epoch:   5 | Train Loss: 0.000 | Train Acc: 94.594 | Val Loss: 0.001 | Val Acc: 94.775
Training:
Accuracies by groups:
0, 0  acc: 68810 / 71629 =  96.064
0, 1  acc: 66571 / 66874 =  99.547
1, 0  acc: 18194 / 22880 =  79.519
1, 1  acc:   395 /  1387 =  28.479
--------------------------------------
Average acc: 153970 / 162770 =  94.594
Robust  acc:   395 /  1387 =  28.479
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8201 /  8535 =  96.087
0, 1  acc:  8256 /  8276 =  99.758
1, 0  acc:  2336 /  2874 =  81.280
1, 1  acc:    36 /   182 =  19.780
------------------------------------
Average acc: 18829 / 19867 =  94.775
Robust  acc:    36 /   182 =  19.780
------------------------------------
New max robust acc: 19.78021978021978
- Saving best checkpoint at epoch 4
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=3-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=4-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.161
Robust Acc: 28.889 | Best Acc: 99.748
-------------------------------------
Training, Epoch 4:
Accuracies by groups:
0, 0  acc:  9492 /  9767 =  97.184
0, 1  acc:  7516 /  7535 =  99.748
1, 0  acc:  1936 /  2480 =  78.065
1, 1  acc:    52 /   180 =  28.889
------------------------------------
Average acc: 18996 / 19962 =  95.161
Robust  acc:    52 /   180 =  28.889
------------------------------------
Accuracies by groups:
0, 0  acc:  9492 /  9767 =  97.184
0, 1  acc:  7516 /  7535 =  99.748
1, 0  acc:  1936 /  2480 =  78.065
1, 1  acc:    52 /   180 =  28.889
------------------------------------
Average acc: 18996 / 19962 =  95.161
Robust  acc:    52 /   180 =  28.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9492 /  9767 =  97.184
0, 1  acc:  7516 /  7535 =  99.748
1, 0  acc:  1936 /  2480 =  78.065
1, 1  acc:    52 /   180 =  28.889
------------------------------------
Average acc: 18996 / 19962 =  95.161
Robust  acc:    52 /   180 =  28.889
------------------------------------
Epoch:   6 | Train Loss: 0.000 | Train Acc: 94.879 | Val Loss: 0.001 | Val Acc: 94.982
Training:
Accuracies by groups:
0, 0  acc: 68792 / 71629 =  96.039
0, 1  acc: 66568 / 66874 =  99.542
1, 0  acc: 18617 / 22880 =  81.368
1, 1  acc:   458 /  1387 =  33.021
--------------------------------------
Average acc: 154435 / 162770 =  94.879
Robust  acc:   458 /  1387 =  33.021
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8167 /  8535 =  95.688
0, 1  acc:  8252 /  8276 =  99.710
1, 0  acc:  2407 /  2874 =  83.751
1, 1  acc:    44 /   182 =  24.176
------------------------------------
Average acc: 18870 / 19867 =  94.982
Robust  acc:    44 /   182 =  24.176
------------------------------------
New max robust acc: 24.175824175824175
- Saving best checkpoint at epoch 5
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=4-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=5-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.481
Robust Acc: 34.444 | Best Acc: 99.695
-------------------------------------
Training, Epoch 5:
Accuracies by groups:
0, 0  acc:  9456 /  9767 =  96.816
0, 1  acc:  7512 /  7535 =  99.695
1, 0  acc:  2030 /  2480 =  81.855
1, 1  acc:    62 /   180 =  34.444
------------------------------------
Average acc: 19060 / 19962 =  95.481
Robust  acc:    62 /   180 =  34.444
------------------------------------
Accuracies by groups:
0, 0  acc:  9456 /  9767 =  96.816
0, 1  acc:  7512 /  7535 =  99.695
1, 0  acc:  2030 /  2480 =  81.855
1, 1  acc:    62 /   180 =  34.444
------------------------------------
Average acc: 19060 / 19962 =  95.481
Robust  acc:    62 /   180 =  34.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9456 /  9767 =  96.816
0, 1  acc:  7512 /  7535 =  99.695
1, 0  acc:  2030 /  2480 =  81.855
1, 1  acc:    62 /   180 =  34.444
------------------------------------
Average acc: 19060 / 19962 =  95.481
Robust  acc:    62 /   180 =  34.444
------------------------------------
Epoch:   7 | Train Loss: 0.000 | Train Acc: 95.081 | Val Loss: 0.001 | Val Acc: 95.248
Training:
Accuracies by groups:
0, 0  acc: 68805 / 71629 =  96.057
0, 1  acc: 66553 / 66874 =  99.520
1, 0  acc: 18916 / 22880 =  82.675
1, 1  acc:   489 /  1387 =  35.256
--------------------------------------
Average acc: 154763 / 162770 =  95.081
Robust  acc:   489 /  1387 =  35.256
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8184 /  8535 =  95.888
0, 1  acc:  8251 /  8276 =  99.698
1, 0  acc:  2438 /  2874 =  84.830
1, 1  acc:    50 /   182 =  27.473
------------------------------------
Average acc: 18923 / 19867 =  95.248
Robust  acc:    50 /   182 =  27.473
------------------------------------
New max robust acc: 27.472527472527474
- Saving best checkpoint at epoch 6
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=5-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=6-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.612
Robust Acc: 33.889 | Best Acc: 99.668
-------------------------------------
Training, Epoch 6:
Accuracies by groups:
0, 0  acc:  9454 /  9767 =  96.795
0, 1  acc:  7510 /  7535 =  99.668
1, 0  acc:  2061 /  2480 =  83.105
1, 1  acc:    61 /   180 =  33.889
------------------------------------
Average acc: 19086 / 19962 =  95.612
Robust  acc:    61 /   180 =  33.889
------------------------------------
Accuracies by groups:
0, 0  acc:  9454 /  9767 =  96.795
0, 1  acc:  7510 /  7535 =  99.668
1, 0  acc:  2061 /  2480 =  83.105
1, 1  acc:    61 /   180 =  33.889
------------------------------------
Average acc: 19086 / 19962 =  95.612
Robust  acc:    61 /   180 =  33.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9454 /  9767 =  96.795
0, 1  acc:  7510 /  7535 =  99.668
1, 0  acc:  2061 /  2480 =  83.105
1, 1  acc:    61 /   180 =  33.889
------------------------------------
Average acc: 19086 / 19962 =  95.612
Robust  acc:    61 /   180 =  33.889
------------------------------------
Epoch:   8 | Train Loss: 0.000 | Train Acc: 95.276 | Val Loss: 0.001 | Val Acc: 95.274
Training:
Accuracies by groups:
0, 0  acc: 68791 / 71629 =  96.038
0, 1  acc: 66541 / 66874 =  99.502
1, 0  acc: 19234 / 22880 =  84.065
1, 1  acc:   515 /  1387 =  37.130
--------------------------------------
Average acc: 155081 / 162770 =  95.276
Robust  acc:   515 /  1387 =  37.130
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8137 /  8535 =  95.337
0, 1  acc:  8249 /  8276 =  99.674
1, 0  acc:  2488 /  2874 =  86.569
1, 1  acc:    54 /   182 =  29.670
------------------------------------
Average acc: 18928 / 19867 =  95.274
Robust  acc:    54 /   182 =  29.670
------------------------------------
New max robust acc: 29.67032967032967
- Saving best checkpoint at epoch 7
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=6-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=7-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.682
Robust Acc: 35.000 | Best Acc: 99.655
-------------------------------------
Training, Epoch 7:
Accuracies by groups:
0, 0  acc:  9411 /  9767 =  96.355
0, 1  acc:  7509 /  7535 =  99.655
1, 0  acc:  2117 /  2480 =  85.363
1, 1  acc:    63 /   180 =  35.000
------------------------------------
Average acc: 19100 / 19962 =  95.682
Robust  acc:    63 /   180 =  35.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9411 /  9767 =  96.355
0, 1  acc:  7509 /  7535 =  99.655
1, 0  acc:  2117 /  2480 =  85.363
1, 1  acc:    63 /   180 =  35.000
------------------------------------
Average acc: 19100 / 19962 =  95.682
Robust  acc:    63 /   180 =  35.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9411 /  9767 =  96.355
0, 1  acc:  7509 /  7535 =  99.655
1, 0  acc:  2117 /  2480 =  85.363
1, 1  acc:    63 /   180 =  35.000
------------------------------------
Average acc: 19100 / 19962 =  95.682
Robust  acc:    63 /   180 =  35.000
------------------------------------
Epoch:   9 | Train Loss: 0.000 | Train Acc: 95.407 | Val Loss: 0.001 | Val Acc: 95.344
Training:
Accuracies by groups:
0, 0  acc: 68690 / 71629 =  95.897
0, 1  acc: 66544 / 66874 =  99.507
1, 0  acc: 19526 / 22880 =  85.341
1, 1  acc:   534 /  1387 =  38.500
--------------------------------------
Average acc: 155294 / 162770 =  95.407
Robust  acc:   534 /  1387 =  38.500
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8141 /  8535 =  95.384
0, 1  acc:  8252 /  8276 =  99.710
1, 0  acc:  2498 /  2874 =  86.917
1, 1  acc:    51 /   182 =  28.022
------------------------------------
Average acc: 18942 / 19867 =  95.344
Robust  acc:    51 /   182 =  28.022
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.772
Robust Acc: 38.333 | Best Acc: 99.668
-------------------------------------
Training, Epoch 8:
Accuracies by groups:
0, 0  acc:  9408 /  9767 =  96.324
0, 1  acc:  7510 /  7535 =  99.668
1, 0  acc:  2131 /  2480 =  85.927
1, 1  acc:    69 /   180 =  38.333
------------------------------------
Average acc: 19118 / 19962 =  95.772
Robust  acc:    69 /   180 =  38.333
------------------------------------
Accuracies by groups:
0, 0  acc:  9408 /  9767 =  96.324
0, 1  acc:  7510 /  7535 =  99.668
1, 0  acc:  2131 /  2480 =  85.927
1, 1  acc:    69 /   180 =  38.333
------------------------------------
Average acc: 19118 / 19962 =  95.772
Robust  acc:    69 /   180 =  38.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9408 /  9767 =  96.324
0, 1  acc:  7510 /  7535 =  99.668
1, 0  acc:  2131 /  2480 =  85.927
1, 1  acc:    69 /   180 =  38.333
------------------------------------
Average acc: 19118 / 19962 =  95.772
Robust  acc:    69 /   180 =  38.333
------------------------------------
Epoch:  10 | Train Loss: 0.000 | Train Acc: 95.466 | Val Loss: 0.001 | Val Acc: 95.455
Training:
Accuracies by groups:
0, 0  acc: 68591 / 71629 =  95.759
0, 1  acc: 66515 / 66874 =  99.463
1, 0  acc: 19741 / 22880 =  86.281
1, 1  acc:   543 /  1387 =  39.149
--------------------------------------
Average acc: 155390 / 162770 =  95.466
Robust  acc:   543 /  1387 =  39.149
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8218 /  8535 =  96.286
0, 1  acc:  8259 /  8276 =  99.795
1, 0  acc:  2441 /  2874 =  84.934
1, 1  acc:    46 /   182 =  25.275
------------------------------------
Average acc: 18964 / 19867 =  95.455
Robust  acc:    46 /   182 =  25.275
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.752
Robust Acc: 32.222 | Best Acc: 99.761
-------------------------------------
Training, Epoch 9:
Accuracies by groups:
0, 0  acc:  9471 /  9767 =  96.969
0, 1  acc:  7517 /  7535 =  99.761
1, 0  acc:  2068 /  2480 =  83.387
1, 1  acc:    58 /   180 =  32.222
------------------------------------
Average acc: 19114 / 19962 =  95.752
Robust  acc:    58 /   180 =  32.222
------------------------------------
Accuracies by groups:
0, 0  acc:  9471 /  9767 =  96.969
0, 1  acc:  7517 /  7535 =  99.761
1, 0  acc:  2068 /  2480 =  83.387
1, 1  acc:    58 /   180 =  32.222
------------------------------------
Average acc: 19114 / 19962 =  95.752
Robust  acc:    58 /   180 =  32.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9471 /  9767 =  96.969
0, 1  acc:  7517 /  7535 =  99.761
1, 0  acc:  2068 /  2480 =  83.387
1, 1  acc:    58 /   180 =  32.222
------------------------------------
Average acc: 19114 / 19962 =  95.752
Robust  acc:    58 /   180 =  32.222
------------------------------------
Epoch:  11 | Train Loss: 0.000 | Train Acc: 95.527 | Val Loss: 0.001 | Val Acc: 95.430
Training:
Accuracies by groups:
0, 0  acc: 68463 / 71629 =  95.580
0, 1  acc: 66514 / 66874 =  99.462
1, 0  acc: 19959 / 22880 =  87.233
1, 1  acc:   553 /  1387 =  39.870
--------------------------------------
Average acc: 155489 / 162770 =  95.527
Robust  acc:   553 /  1387 =  39.870
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8075 /  8535 =  94.610
0, 1  acc:  8244 /  8276 =  99.613
1, 0  acc:  2578 /  2874 =  89.701
1, 1  acc:    62 /   182 =  34.066
------------------------------------
Average acc: 18959 / 19867 =  95.430
Robust  acc:    62 /   182 =  34.066
------------------------------------
New max robust acc: 34.065934065934066
- Saving best checkpoint at epoch 10
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=7-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=10-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.747
Robust Acc: 45.000 | Best Acc: 99.522
-------------------------------------
Training, Epoch 10:
Accuracies by groups:
0, 0  acc:  9332 /  9767 =  95.546
0, 1  acc:  7499 /  7535 =  99.522
1, 0  acc:  2201 /  2480 =  88.750
1, 1  acc:    81 /   180 =  45.000
------------------------------------
Average acc: 19113 / 19962 =  95.747
Robust  acc:    81 /   180 =  45.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9332 /  9767 =  95.546
0, 1  acc:  7499 /  7535 =  99.522
1, 0  acc:  2201 /  2480 =  88.750
1, 1  acc:    81 /   180 =  45.000
------------------------------------
Average acc: 19113 / 19962 =  95.747
Robust  acc:    81 /   180 =  45.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9332 /  9767 =  95.546
0, 1  acc:  7499 /  7535 =  99.522
1, 0  acc:  2201 /  2480 =  88.750
1, 1  acc:    81 /   180 =  45.000
------------------------------------
Average acc: 19113 / 19962 =  95.747
Robust  acc:    81 /   180 =  45.000
------------------------------------
Epoch:  12 | Train Loss: 0.000 | Train Acc: 95.520 | Val Loss: 0.001 | Val Acc: 95.430
Training:
Accuracies by groups:
0, 0  acc: 68263 / 71629 =  95.301
0, 1  acc: 66439 / 66874 =  99.350
1, 0  acc: 20160 / 22880 =  88.112
1, 1  acc:   616 /  1387 =  44.412
--------------------------------------
Average acc: 155478 / 162770 =  95.520
Robust  acc:   616 /  1387 =  44.412
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8108 /  8535 =  94.997
0, 1  acc:  8243 /  8276 =  99.601
1, 0  acc:  2548 /  2874 =  88.657
1, 1  acc:    60 /   182 =  32.967
------------------------------------
Average acc: 18959 / 19867 =  95.430
Robust  acc:    60 /   182 =  32.967
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.792
Robust Acc: 43.889 | Best Acc: 99.522
-------------------------------------
Training, Epoch 11:
Accuracies by groups:
0, 0  acc:  9371 /  9767 =  95.946
0, 1  acc:  7499 /  7535 =  99.522
1, 0  acc:  2173 /  2480 =  87.621
1, 1  acc:    79 /   180 =  43.889
------------------------------------
Average acc: 19122 / 19962 =  95.792
Robust  acc:    79 /   180 =  43.889
------------------------------------
Accuracies by groups:
0, 0  acc:  9371 /  9767 =  95.946
0, 1  acc:  7499 /  7535 =  99.522
1, 0  acc:  2173 /  2480 =  87.621
1, 1  acc:    79 /   180 =  43.889
------------------------------------
Average acc: 19122 / 19962 =  95.792
Robust  acc:    79 /   180 =  43.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9371 /  9767 =  95.946
0, 1  acc:  7499 /  7535 =  99.522
1, 0  acc:  2173 /  2480 =  87.621
1, 1  acc:    79 /   180 =  43.889
------------------------------------
Average acc: 19122 / 19962 =  95.792
Robust  acc:    79 /   180 =  43.889
------------------------------------
Epoch:  13 | Train Loss: 0.000 | Train Acc: 95.411 | Val Loss: 0.001 | Val Acc: 95.339
Training:
Accuracies by groups:
0, 0  acc: 68200 / 71629 =  95.213
0, 1  acc: 66401 / 66874 =  99.293
1, 0  acc: 20087 / 22880 =  87.793
1, 1  acc:   612 /  1387 =  44.124
--------------------------------------
Average acc: 155300 / 162770 =  95.411
Robust  acc:   612 /  1387 =  44.124
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8096 /  8535 =  94.856
0, 1  acc:  8238 /  8276 =  99.541
1, 0  acc:  2541 /  2874 =  88.413
1, 1  acc:    66 /   182 =  36.264
------------------------------------
Average acc: 18941 / 19867 =  95.339
Robust  acc:    66 /   182 =  36.264
------------------------------------
New max robust acc: 36.26373626373626
- Saving best checkpoint at epoch 12
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=10-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=12-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.772
Robust Acc: 42.778 | Best Acc: 99.482
-------------------------------------
Training, Epoch 12:
Accuracies by groups:
0, 0  acc:  9371 /  9767 =  95.946
0, 1  acc:  7496 /  7535 =  99.482
1, 0  acc:  2174 /  2480 =  87.661
1, 1  acc:    77 /   180 =  42.778
------------------------------------
Average acc: 19118 / 19962 =  95.772
Robust  acc:    77 /   180 =  42.778
------------------------------------
Accuracies by groups:
0, 0  acc:  9371 /  9767 =  95.946
0, 1  acc:  7496 /  7535 =  99.482
1, 0  acc:  2174 /  2480 =  87.661
1, 1  acc:    77 /   180 =  42.778
------------------------------------
Average acc: 19118 / 19962 =  95.772
Robust  acc:    77 /   180 =  42.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9371 /  9767 =  95.946
0, 1  acc:  7496 /  7535 =  99.482
1, 0  acc:  2174 /  2480 =  87.661
1, 1  acc:    77 /   180 =  42.778
------------------------------------
Average acc: 19118 / 19962 =  95.772
Robust  acc:    77 /   180 =  42.778
------------------------------------
Epoch:  14 | Train Loss: 0.000 | Train Acc: 95.284 | Val Loss: 0.001 | Val Acc: 95.087
Training:
Accuracies by groups:
0, 0  acc: 68307 / 71629 =  95.362
0, 1  acc: 66394 / 66874 =  99.282
1, 0  acc: 19778 / 22880 =  86.442
1, 1  acc:   614 /  1387 =  44.268
--------------------------------------
Average acc: 155093 / 162770 =  95.284
Robust  acc:   614 /  1387 =  44.268
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8211 /  8535 =  96.204
0, 1  acc:  8247 /  8276 =  99.650
1, 0  acc:  2378 /  2874 =  82.742
1, 1  acc:    55 /   182 =  30.220
------------------------------------
Average acc: 18891 / 19867 =  95.087
Robust  acc:    55 /   182 =  30.220
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.552
Robust Acc: 32.222 | Best Acc: 99.628
-------------------------------------
Training, Epoch 13:
Accuracies by groups:
0, 0  acc:  9490 /  9767 =  97.164
0, 1  acc:  7507 /  7535 =  99.628
1, 0  acc:  2019 /  2480 =  81.411
1, 1  acc:    58 /   180 =  32.222
------------------------------------
Average acc: 19074 / 19962 =  95.552
Robust  acc:    58 /   180 =  32.222
------------------------------------
Accuracies by groups:
0, 0  acc:  9490 /  9767 =  97.164
0, 1  acc:  7507 /  7535 =  99.628
1, 0  acc:  2019 /  2480 =  81.411
1, 1  acc:    58 /   180 =  32.222
------------------------------------
Average acc: 19074 / 19962 =  95.552
Robust  acc:    58 /   180 =  32.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9490 /  9767 =  97.164
0, 1  acc:  7507 /  7535 =  99.628
1, 0  acc:  2019 /  2480 =  81.411
1, 1  acc:    58 /   180 =  32.222
------------------------------------
Average acc: 19074 / 19962 =  95.552
Robust  acc:    58 /   180 =  32.222
------------------------------------
Epoch:  15 | Train Loss: 0.001 | Train Acc: 95.076 | Val Loss: 0.001 | Val Acc: 94.624
Training:
Accuracies by groups:
0, 0  acc: 68445 / 71629 =  95.555
0, 1  acc: 66343 / 66874 =  99.206
1, 0  acc: 19357 / 22880 =  84.602
1, 1  acc:   611 /  1387 =  44.052
--------------------------------------
Average acc: 154756 / 162770 =  95.076
Robust  acc:   611 /  1387 =  44.052
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8337 /  8535 =  97.680
0, 1  acc:  8254 /  8276 =  99.734
1, 0  acc:  2165 /  2874 =  75.331
1, 1  acc:    43 /   182 =  23.626
------------------------------------
Average acc: 18799 / 19867 =  94.624
Robust  acc:    43 /   182 =  23.626
------------------------------------
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.091
Robust Acc: 23.889 | Best Acc: 99.814
-------------------------------------
Training, Epoch 14:
Accuracies by groups:
0, 0  acc:  9595 /  9767 =  98.239
0, 1  acc:  7521 /  7535 =  99.814
1, 0  acc:  1823 /  2480 =  73.508
1, 1  acc:    43 /   180 =  23.889
------------------------------------
Average acc: 18982 / 19962 =  95.091
Robust  acc:    43 /   180 =  23.889
------------------------------------
Accuracies by groups:
0, 0  acc:  9595 /  9767 =  98.239
0, 1  acc:  7521 /  7535 =  99.814
1, 0  acc:  1823 /  2480 =  73.508
1, 1  acc:    43 /   180 =  23.889
------------------------------------
Average acc: 18982 / 19962 =  95.091
Robust  acc:    43 /   180 =  23.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9595 /  9767 =  98.239
0, 1  acc:  7521 /  7535 =  99.814
1, 0  acc:  1823 /  2480 =  73.508
1, 1  acc:    43 /   180 =  23.889
------------------------------------
Average acc: 18982 / 19962 =  95.091
Robust  acc:    43 /   180 =  23.889
------------------------------------
Epoch:  16 | Train Loss: 0.001 | Train Acc: 94.990 | Val Loss: 0.001 | Val Acc: 94.891
Training:
Accuracies by groups:
0, 0  acc: 68519 / 71629 =  95.658
0, 1  acc: 66330 / 66874 =  99.187
1, 0  acc: 19167 / 22880 =  83.772
1, 1  acc:   599 /  1387 =  43.187
--------------------------------------
Average acc: 154615 / 162770 =  94.990
Robust  acc:   599 /  1387 =  43.187
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8070 /  8535 =  94.552
0, 1  acc:  8188 /  8276 =  98.937
1, 0  acc:  2514 /  2874 =  87.474
1, 1  acc:    80 /   182 =  43.956
------------------------------------
Average acc: 18852 / 19867 =  94.891
Robust  acc:    80 /   182 =  43.956
------------------------------------
New max robust acc: 43.956043956043956
- Saving best checkpoint at epoch 15
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=12-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=15-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 95.376
Robust Acc: 46.667 | Best Acc: 99.084
-------------------------------------
Training, Epoch 15:
Accuracies by groups:
0, 0  acc:  9361 /  9767 =  95.843
0, 1  acc:  7466 /  7535 =  99.084
1, 0  acc:  2128 /  2480 =  85.806
1, 1  acc:    84 /   180 =  46.667
------------------------------------
Average acc: 19039 / 19962 =  95.376
Robust  acc:    84 /   180 =  46.667
------------------------------------
Accuracies by groups:
0, 0  acc:  9361 /  9767 =  95.843
0, 1  acc:  7466 /  7535 =  99.084
1, 0  acc:  2128 /  2480 =  85.806
1, 1  acc:    84 /   180 =  46.667
------------------------------------
Average acc: 19039 / 19962 =  95.376
Robust  acc:    84 /   180 =  46.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9361 /  9767 =  95.843
0, 1  acc:  7466 /  7535 =  99.084
1, 0  acc:  2128 /  2480 =  85.806
1, 1  acc:    84 /   180 =  46.667
------------------------------------
Average acc: 19039 / 19962 =  95.376
Robust  acc:    84 /   180 =  46.667
------------------------------------
Epoch:  17 | Train Loss: 0.001 | Train Acc: 94.817 | Val Loss: 0.001 | Val Acc: 94.030
Training:
Accuracies by groups:
0, 0  acc: 68690 / 71629 =  95.897
0, 1  acc: 66312 / 66874 =  99.160
1, 0  acc: 18768 / 22880 =  82.028
1, 1  acc:   564 /  1387 =  40.663
--------------------------------------
Average acc: 154334 / 162770 =  94.817
Robust  acc:   564 /  1387 =  40.663
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7884 /  8535 =  92.373
0, 1  acc:  8134 /  8276 =  98.284
1, 0  acc:  2577 /  2874 =  89.666
1, 1  acc:    86 /   182 =  47.253
------------------------------------
Average acc: 18681 / 19867 =  94.030
Robust  acc:    86 /   182 =  47.253
------------------------------------
New max robust acc: 47.25274725274725
- Saving best checkpoint at epoch 16
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=15-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=16-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.870
Robust Acc: 52.222 | Best Acc: 98.421
-------------------------------------
Training, Epoch 16:
Accuracies by groups:
0, 0  acc:  9239 /  9767 =  94.594
0, 1  acc:  7416 /  7535 =  98.421
1, 0  acc:  2189 /  2480 =  88.266
1, 1  acc:    94 /   180 =  52.222
------------------------------------
Average acc: 18938 / 19962 =  94.870
Robust  acc:    94 /   180 =  52.222
------------------------------------
Accuracies by groups:
0, 0  acc:  9239 /  9767 =  94.594
0, 1  acc:  7416 /  7535 =  98.421
1, 0  acc:  2189 /  2480 =  88.266
1, 1  acc:    94 /   180 =  52.222
------------------------------------
Average acc: 18938 / 19962 =  94.870
Robust  acc:    94 /   180 =  52.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9239 /  9767 =  94.594
0, 1  acc:  7416 /  7535 =  98.421
1, 0  acc:  2189 /  2480 =  88.266
1, 1  acc:    94 /   180 =  52.222
------------------------------------
Average acc: 18938 / 19962 =  94.870
Robust  acc:    94 /   180 =  52.222
------------------------------------
Epoch:  18 | Train Loss: 0.001 | Train Acc: 94.679 | Val Loss: 0.002 | Val Acc: 93.975
Training:
Accuracies by groups:
0, 0  acc: 68807 / 71629 =  96.060
0, 1  acc: 66337 / 66874 =  99.197
1, 0  acc: 18406 / 22880 =  80.446
1, 1  acc:   559 /  1387 =  40.303
--------------------------------------
Average acc: 154109 / 162770 =  94.679
Robust  acc:   559 /  1387 =  40.303
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7946 /  8535 =  93.099
0, 1  acc:  8080 /  8276 =  97.632
1, 0  acc:  2546 /  2874 =  88.587
1, 1  acc:    98 /   182 =  53.846
------------------------------------
Average acc: 18670 / 19867 =  93.975
Robust  acc:    98 /   182 =  53.846
------------------------------------
New max robust acc: 53.84615384615385
- Saving best checkpoint at epoch 17
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=16-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=17-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.001 | Avg Test Acc: 94.555
Robust Acc: 52.778 | Best Acc: 97.691
-------------------------------------
Training, Epoch 17:
Accuracies by groups:
0, 0  acc:  9263 /  9767 =  94.840
0, 1  acc:  7361 /  7535 =  97.691
1, 0  acc:  2156 /  2480 =  86.935
1, 1  acc:    95 /   180 =  52.778
------------------------------------
Average acc: 18875 / 19962 =  94.555
Robust  acc:    95 /   180 =  52.778
------------------------------------
Accuracies by groups:
0, 0  acc:  9263 /  9767 =  94.840
0, 1  acc:  7361 /  7535 =  97.691
1, 0  acc:  2156 /  2480 =  86.935
1, 1  acc:    95 /   180 =  52.778
------------------------------------
Average acc: 18875 / 19962 =  94.555
Robust  acc:    95 /   180 =  52.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9263 /  9767 =  94.840
0, 1  acc:  7361 /  7535 =  97.691
1, 0  acc:  2156 /  2480 =  86.935
1, 1  acc:    95 /   180 =  52.778
------------------------------------
Average acc: 18875 / 19962 =  94.555
Robust  acc:    95 /   180 =  52.778
------------------------------------
Epoch:  19 | Train Loss: 0.001 | Train Acc: 94.474 | Val Loss: 0.002 | Val Acc: 91.735
Training:
Accuracies by groups:
0, 0  acc: 68979 / 71629 =  96.300
0, 1  acc: 66379 / 66874 =  99.260
1, 0  acc: 17917 / 22880 =  78.309
1, 1  acc:   500 /  1387 =  36.049
--------------------------------------
Average acc: 153775 / 162770 =  94.474
Robust  acc:   500 /  1387 =  36.049
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8399 /  8535 =  98.407
0, 1  acc:  8265 /  8276 =  99.867
1, 0  acc:  1547 /  2874 =  53.827
1, 1  acc:    14 /   182 =   7.692
------------------------------------
Average acc: 18225 / 19867 =  91.735
Robust  acc:    14 /   182 =   7.692
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.556
Robust Acc: 11.111 | Best Acc: 99.934
-------------------------------------
Training, Epoch 18:
Accuracies by groups:
0, 0  acc:  9653 /  9767 =  98.833
0, 1  acc:  7530 /  7535 =  99.934
1, 0  acc:  1273 /  2480 =  51.331
1, 1  acc:    20 /   180 =  11.111
------------------------------------
Average acc: 18476 / 19962 =  92.556
Robust  acc:    20 /   180 =  11.111
------------------------------------
Accuracies by groups:
0, 0  acc:  9653 /  9767 =  98.833
0, 1  acc:  7530 /  7535 =  99.934
1, 0  acc:  1273 /  2480 =  51.331
1, 1  acc:    20 /   180 =  11.111
------------------------------------
Average acc: 18476 / 19962 =  92.556
Robust  acc:    20 /   180 =  11.111
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9653 /  9767 =  98.833
0, 1  acc:  7530 /  7535 =  99.934
1, 0  acc:  1273 /  2480 =  51.331
1, 1  acc:    20 /   180 =  11.111
------------------------------------
Average acc: 18476 / 19962 =  92.556
Robust  acc:    20 /   180 =  11.111
------------------------------------
Epoch:  20 | Train Loss: 0.001 | Train Acc: 94.206 | Val Loss: 0.002 | Val Acc: 94.126
Training:
Accuracies by groups:
0, 0  acc: 69183 / 71629 =  96.585
0, 1  acc: 66434 / 66874 =  99.342
1, 0  acc: 17267 / 22880 =  75.468
1, 1  acc:   455 /  1387 =  32.805
--------------------------------------
Average acc: 153339 / 162770 =  94.206
Robust  acc:   455 /  1387 =  32.805
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8179 /  8535 =  95.829
0, 1  acc:  8235 /  8276 =  99.505
1, 0  acc:  2242 /  2874 =  78.010
1, 1  acc:    44 /   182 =  24.176
------------------------------------
Average acc: 18700 / 19867 =  94.126
Robust  acc:    44 /   182 =  24.176
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 94.840
Robust Acc: 23.889 | Best Acc: 99.655
-------------------------------------
Training, Epoch 19:
Accuracies by groups:
0, 0  acc:  9480 /  9767 =  97.062
0, 1  acc:  7509 /  7535 =  99.655
1, 0  acc:  1900 /  2480 =  76.613
1, 1  acc:    43 /   180 =  23.889
------------------------------------
Average acc: 18932 / 19962 =  94.840
Robust  acc:    43 /   180 =  23.889
------------------------------------
Accuracies by groups:
0, 0  acc:  9480 /  9767 =  97.062
0, 1  acc:  7509 /  7535 =  99.655
1, 0  acc:  1900 /  2480 =  76.613
1, 1  acc:    43 /   180 =  23.889
------------------------------------
Average acc: 18932 / 19962 =  94.840
Robust  acc:    43 /   180 =  23.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9480 /  9767 =  97.062
0, 1  acc:  7509 /  7535 =  99.655
1, 0  acc:  1900 /  2480 =  76.613
1, 1  acc:    43 /   180 =  23.889
------------------------------------
Average acc: 18932 / 19962 =  94.840
Robust  acc:    43 /   180 =  23.889
------------------------------------
Epoch:  21 | Train Loss: 0.001 | Train Acc: 93.956 | Val Loss: 0.002 | Val Acc: 93.376
Training:
Accuracies by groups:
0, 0  acc: 69499 / 71629 =  97.026
0, 1  acc: 66477 / 66874 =  99.406
1, 0  acc: 16549 / 22880 =  72.330
1, 1  acc:   407 /  1387 =  29.344
--------------------------------------
Average acc: 152932 / 162770 =  93.956
Robust  acc:   407 /  1387 =  29.344
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7901 /  8535 =  92.572
0, 1  acc:  8100 /  8276 =  97.873
1, 0  acc:  2469 /  2874 =  85.908
1, 1  acc:    81 /   182 =  44.505
------------------------------------
Average acc: 18551 / 19867 =  93.376
Robust  acc:    81 /   182 =  44.505
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 94.394
Robust Acc: 40.556 | Best Acc: 98.354
-------------------------------------
Training, Epoch 20:
Accuracies by groups:
0, 0  acc:  9261 /  9767 =  94.819
0, 1  acc:  7411 /  7535 =  98.354
1, 0  acc:  2098 /  2480 =  84.597
1, 1  acc:    73 /   180 =  40.556
------------------------------------
Average acc: 18843 / 19962 =  94.394
Robust  acc:    73 /   180 =  40.556
------------------------------------
Accuracies by groups:
0, 0  acc:  9261 /  9767 =  94.819
0, 1  acc:  7411 /  7535 =  98.354
1, 0  acc:  2098 /  2480 =  84.597
1, 1  acc:    73 /   180 =  40.556
------------------------------------
Average acc: 18843 / 19962 =  94.394
Robust  acc:    73 /   180 =  40.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9261 /  9767 =  94.819
0, 1  acc:  7411 /  7535 =  98.354
1, 0  acc:  2098 /  2480 =  84.597
1, 1  acc:    73 /   180 =  40.556
------------------------------------
Average acc: 18843 / 19962 =  94.394
Robust  acc:    73 /   180 =  40.556
------------------------------------
Epoch:  22 | Train Loss: 0.001 | Train Acc: 93.613 | Val Loss: 0.003 | Val Acc: 85.821
Training:
Accuracies by groups:
0, 0  acc: 69758 / 71629 =  97.388
0, 1  acc: 66540 / 66874 =  99.501
1, 0  acc: 15713 / 22880 =  68.676
1, 1  acc:   363 /  1387 =  26.172
--------------------------------------
Average acc: 152374 / 162770 =  93.613
Robust  acc:   363 /  1387 =  26.172
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8532 /  8535 =  99.965
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:   240 /  2874 =   8.351
1, 1  acc:     2 /   182 =   1.099
------------------------------------
Average acc: 17050 / 19867 =  85.821
Robust  acc:     2 /   182 =   1.099
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 87.611
Robust Acc: 0.556 | Best Acc: 100.000
-------------------------------------
Training, Epoch 21:
Accuracies by groups:
0, 0  acc:  9762 /  9767 =  99.949
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:   191 /  2480 =   7.702
1, 1  acc:     1 /   180 =   0.556
------------------------------------
Average acc: 17489 / 19962 =  87.611
Robust  acc:     1 /   180 =   0.556
------------------------------------
Accuracies by groups:
0, 0  acc:  9762 /  9767 =  99.949
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:   191 /  2480 =   7.702
1, 1  acc:     1 /   180 =   0.556
------------------------------------
Average acc: 17489 / 19962 =  87.611
Robust  acc:     1 /   180 =   0.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9762 /  9767 =  99.949
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:   191 /  2480 =   7.702
1, 1  acc:     1 /   180 =   0.556
------------------------------------
Average acc: 17489 / 19962 =  87.611
Robust  acc:     1 /   180 =   0.556
------------------------------------
Epoch:  23 | Train Loss: 0.001 | Train Acc: 93.196 | Val Loss: 0.002 | Val Acc: 87.472
Training:
Accuracies by groups:
0, 0  acc: 69948 / 71629 =  97.653
0, 1  acc: 66554 / 66874 =  99.521
1, 0  acc: 14869 / 22880 =  64.987
1, 1  acc:   324 /  1387 =  23.360
--------------------------------------
Average acc: 151695 / 162770 =  93.196
Robust  acc:   324 /  1387 =  23.360
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8524 /  8535 =  99.871
0, 1  acc:  8274 /  8276 =  99.976
1, 0  acc:   572 /  2874 =  19.903
1, 1  acc:     8 /   182 =   4.396
------------------------------------
Average acc: 17378 / 19867 =  87.472
Robust  acc:     8 /   182 =   4.396
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 89.099
Robust Acc: 3.333 | Best Acc: 99.987
------------------------------------
Training, Epoch 22:
Accuracies by groups:
0, 0  acc:  9754 /  9767 =  99.867
0, 1  acc:  7534 /  7535 =  99.987
1, 0  acc:   492 /  2480 =  19.839
1, 1  acc:     6 /   180 =   3.333
------------------------------------
Average acc: 17786 / 19962 =  89.099
Robust  acc:     6 /   180 =   3.333
------------------------------------
Accuracies by groups:
0, 0  acc:  9754 /  9767 =  99.867
0, 1  acc:  7534 /  7535 =  99.987
1, 0  acc:   492 /  2480 =  19.839
1, 1  acc:     6 /   180 =   3.333
------------------------------------
Average acc: 17786 / 19962 =  89.099
Robust  acc:     6 /   180 =   3.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9754 /  9767 =  99.867
0, 1  acc:  7534 /  7535 =  99.987
1, 0  acc:   492 /  2480 =  19.839
1, 1  acc:     6 /   180 =   3.333
------------------------------------
Average acc: 17786 / 19962 =  89.099
Robust  acc:     6 /   180 =   3.333
------------------------------------
Epoch:  24 | Train Loss: 0.001 | Train Acc: 92.942 | Val Loss: 0.002 | Val Acc: 92.782
Training:
Accuracies by groups:
0, 0  acc: 70119 / 71629 =  97.892
0, 1  acc: 66595 / 66874 =  99.583
1, 0  acc: 14271 / 22880 =  62.373
1, 1  acc:   297 /  1387 =  21.413
--------------------------------------
Average acc: 151282 / 162770 =  92.942
Robust  acc:   297 /  1387 =  21.413
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8393 /  8535 =  98.336
0, 1  acc:  8255 /  8276 =  99.746
1, 0  acc:  1758 /  2874 =  61.169
1, 1  acc:    27 /   182 =  14.835
------------------------------------
Average acc: 18433 / 19867 =  92.782
Robust  acc:    27 /   182 =  14.835
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.152
Robust Acc: 13.889 | Best Acc: 99.814
-------------------------------------
Training, Epoch 23:
Accuracies by groups:
0, 0  acc:  9622 /  9767 =  98.515
0, 1  acc:  7521 /  7535 =  99.814
1, 0  acc:  1427 /  2480 =  57.540
1, 1  acc:    25 /   180 =  13.889
------------------------------------
Average acc: 18595 / 19962 =  93.152
Robust  acc:    25 /   180 =  13.889
------------------------------------
Accuracies by groups:
0, 0  acc:  9622 /  9767 =  98.515
0, 1  acc:  7521 /  7535 =  99.814
1, 0  acc:  1427 /  2480 =  57.540
1, 1  acc:    25 /   180 =  13.889
------------------------------------
Average acc: 18595 / 19962 =  93.152
Robust  acc:    25 /   180 =  13.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9622 /  9767 =  98.515
0, 1  acc:  7521 /  7535 =  99.814
1, 0  acc:  1427 /  2480 =  57.540
1, 1  acc:    25 /   180 =  13.889
------------------------------------
Average acc: 18595 / 19962 =  93.152
Robust  acc:    25 /   180 =  13.889
------------------------------------
Epoch:  25 | Train Loss: 0.001 | Train Acc: 92.543 | Val Loss: 0.003 | Val Acc: 90.804
Training:
Accuracies by groups:
0, 0  acc: 70392 / 71629 =  98.273
0, 1  acc: 66665 / 66874 =  99.687
1, 0  acc: 13304 / 22880 =  58.147
1, 1  acc:   272 /  1387 =  19.611
--------------------------------------
Average acc: 150633 / 162770 =  92.543
Robust  acc:   272 /  1387 =  19.611
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7512 /  8535 =  88.014
0, 1  acc:  7801 /  8276 =  94.261
1, 0  acc:  2611 /  2874 =  90.849
1, 1  acc:   116 /   182 =  63.736
------------------------------------
Average acc: 18040 / 19867 =  90.804
Robust  acc:   116 /   182 =  63.736
------------------------------------
New max robust acc: 63.73626373626373
- Saving best checkpoint at epoch 24
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=17-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=24-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.829
Robust Acc: 62.778 | Best Acc: 94.453
-------------------------------------
Training, Epoch 24:
Accuracies by groups:
0, 0  acc:  8878 /  9767 =  90.898
0, 1  acc:  7117 /  7535 =  94.453
1, 0  acc:  2223 /  2480 =  89.637
1, 1  acc:   113 /   180 =  62.778
------------------------------------
Average acc: 18331 / 19962 =  91.829
Robust  acc:   113 /   180 =  62.778
------------------------------------
Accuracies by groups:
0, 0  acc:  8878 /  9767 =  90.898
0, 1  acc:  7117 /  7535 =  94.453
1, 0  acc:  2223 /  2480 =  89.637
1, 1  acc:   113 /   180 =  62.778
------------------------------------
Average acc: 18331 / 19962 =  91.829
Robust  acc:   113 /   180 =  62.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8878 /  9767 =  90.898
0, 1  acc:  7117 /  7535 =  94.453
1, 0  acc:  2223 /  2480 =  89.637
1, 1  acc:   113 /   180 =  62.778
------------------------------------
Average acc: 18331 / 19962 =  91.829
Robust  acc:   113 /   180 =  62.778
------------------------------------
Epoch:  26 | Train Loss: 0.001 | Train Acc: 92.034 | Val Loss: 0.003 | Val Acc: 85.760
Training:
Accuracies by groups:
0, 0  acc: 70638 / 71629 =  98.616
0, 1  acc: 66704 / 66874 =  99.746
1, 0  acc: 12244 / 22880 =  53.514
1, 1  acc:   218 /  1387 =  15.717
--------------------------------------
Average acc: 149804 / 162770 =  92.034
Robust  acc:   218 /  1387 =  15.717
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6891 /  8535 =  80.738
0, 1  acc:  7261 /  8276 =  87.736
1, 0  acc:  2743 /  2874 =  95.442
1, 1  acc:   143 /   182 =  78.571
------------------------------------
Average acc: 17038 / 19867 =  85.760
Robust  acc:   143 /   182 =  78.571
------------------------------------
New max robust acc: 78.57142857142857
- Saving best checkpoint at epoch 25
replace: True
-> Updating checkpoint cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=24-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-bias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=25-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 86.830
Robust Acc: 74.444 | Best Acc: 94.153
-------------------------------------
Training, Epoch 25:
Accuracies by groups:
0, 0  acc:  8279 /  9767 =  84.765
0, 1  acc:  6585 /  7535 =  87.392
1, 0  acc:  2335 /  2480 =  94.153
1, 1  acc:   134 /   180 =  74.444
------------------------------------
Average acc: 17333 / 19962 =  86.830
Robust  acc:   134 /   180 =  74.444
------------------------------------
Accuracies by groups:
0, 0  acc:  8279 /  9767 =  84.765
0, 1  acc:  6585 /  7535 =  87.392
1, 0  acc:  2335 /  2480 =  94.153
1, 1  acc:   134 /   180 =  74.444
------------------------------------
Average acc: 17333 / 19962 =  86.830
Robust  acc:   134 /   180 =  74.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8279 /  9767 =  84.765
0, 1  acc:  6585 /  7535 =  87.392
1, 0  acc:  2335 /  2480 =  94.153
1, 1  acc:   134 /   180 =  74.444
------------------------------------
Average acc: 17333 / 19962 =  86.830
Robust  acc:   134 /   180 =  74.444
------------------------------------
Epoch:  27 | Train Loss: 0.001 | Train Acc: 91.388 | Val Loss: 0.002 | Val Acc: 91.121
Training:
Accuracies by groups:
0, 0  acc: 70912 / 71629 =  98.999
0, 1  acc: 66770 / 66874 =  99.844
1, 0  acc: 10903 / 22880 =  47.653
1, 1  acc:   168 /  1387 =  12.112
--------------------------------------
Average acc: 148753 / 162770 =  91.388
Robust  acc:   168 /  1387 =  12.112
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8468 /  8535 =  99.215
0, 1  acc:  8268 /  8276 =  99.903
1, 0  acc:  1352 /  2874 =  47.042
1, 1  acc:    15 /   182 =   8.242
------------------------------------
Average acc: 18103 / 19867 =  91.121
Robust  acc:    15 /   182 =   8.242
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.654
Robust Acc: 10.556 | Best Acc: 99.934
-------------------------------------
Training, Epoch 26:
Accuracies by groups:
0, 0  acc:  9701 /  9767 =  99.324
0, 1  acc:  7530 /  7535 =  99.934
1, 0  acc:  1046 /  2480 =  42.177
1, 1  acc:    19 /   180 =  10.556
------------------------------------
Average acc: 18296 / 19962 =  91.654
Robust  acc:    19 /   180 =  10.556
------------------------------------
Accuracies by groups:
0, 0  acc:  9701 /  9767 =  99.324
0, 1  acc:  7530 /  7535 =  99.934
1, 0  acc:  1046 /  2480 =  42.177
1, 1  acc:    19 /   180 =  10.556
------------------------------------
Average acc: 18296 / 19962 =  91.654
Robust  acc:    19 /   180 =  10.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9701 /  9767 =  99.324
0, 1  acc:  7530 /  7535 =  99.934
1, 0  acc:  1046 /  2480 =  42.177
1, 1  acc:    19 /   180 =  10.556
------------------------------------
Average acc: 18296 / 19962 =  91.654
Robust  acc:    19 /   180 =  10.556
------------------------------------
Epoch:  28 | Train Loss: 0.001 | Train Acc: 90.566 | Val Loss: 0.002 | Val Acc: 92.359
Training:
Accuracies by groups:
0, 0  acc: 71130 / 71629 =  99.303
0, 1  acc: 66791 / 66874 =  99.876
1, 0  acc:  9372 / 22880 =  40.962
1, 1  acc:   121 /  1387 =   8.724
--------------------------------------
Average acc: 147414 / 162770 =  90.566
Robust  acc:   121 /  1387 =   8.724
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8416 /  8535 =  98.606
0, 1  acc:  8264 /  8276 =  99.855
1, 0  acc:  1648 /  2874 =  57.342
1, 1  acc:    21 /   182 =  11.538
------------------------------------
Average acc: 18349 / 19867 =  92.359
Robust  acc:    21 /   182 =  11.538
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.756
Robust Acc: 12.778 | Best Acc: 99.841
-------------------------------------
Training, Epoch 27:
Accuracies by groups:
0, 0  acc:  9658 /  9767 =  98.884
0, 1  acc:  7523 /  7535 =  99.841
1, 0  acc:  1312 /  2480 =  52.903
1, 1  acc:    23 /   180 =  12.778
------------------------------------
Average acc: 18516 / 19962 =  92.756
Robust  acc:    23 /   180 =  12.778
------------------------------------
Accuracies by groups:
0, 0  acc:  9658 /  9767 =  98.884
0, 1  acc:  7523 /  7535 =  99.841
1, 0  acc:  1312 /  2480 =  52.903
1, 1  acc:    23 /   180 =  12.778
------------------------------------
Average acc: 18516 / 19962 =  92.756
Robust  acc:    23 /   180 =  12.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9658 /  9767 =  98.884
0, 1  acc:  7523 /  7535 =  99.841
1, 0  acc:  1312 /  2480 =  52.903
1, 1  acc:    23 /   180 =  12.778
------------------------------------
Average acc: 18516 / 19962 =  92.756
Robust  acc:    23 /   180 =  12.778
------------------------------------
Epoch:  29 | Train Loss: 0.001 | Train Acc: 89.628 | Val Loss: 0.002 | Val Acc: 88.046
Training:
Accuracies by groups:
0, 0  acc: 71286 / 71629 =  99.521
0, 1  acc: 66830 / 66874 =  99.934
1, 0  acc:  7691 / 22880 =  33.615
1, 1  acc:    80 /  1387 =   5.768
--------------------------------------
Average acc: 145887 / 162770 =  89.628
Robust  acc:    80 /  1387 =   5.768
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8520 /  8535 =  99.824
0, 1  acc:  8275 /  8276 =  99.988
1, 0  acc:   693 /  2874 =  24.113
1, 1  acc:     4 /   182 =   2.198
------------------------------------
Average acc: 17492 / 19867 =  88.046
Robust  acc:     4 /   182 =   2.198
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 89.455
Robust Acc: 1.667 | Best Acc: 100.000
-------------------------------------
Training, Epoch 28:
Accuracies by groups:
0, 0  acc:  9753 /  9767 =  99.857
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:   566 /  2480 =  22.823
1, 1  acc:     3 /   180 =   1.667
------------------------------------
Average acc: 17857 / 19962 =  89.455
Robust  acc:     3 /   180 =   1.667
------------------------------------
Accuracies by groups:
0, 0  acc:  9753 /  9767 =  99.857
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:   566 /  2480 =  22.823
1, 1  acc:     3 /   180 =   1.667
------------------------------------
Average acc: 17857 / 19962 =  89.455
Robust  acc:     3 /   180 =   1.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9753 /  9767 =  99.857
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:   566 /  2480 =  22.823
1, 1  acc:     3 /   180 =   1.667
------------------------------------
Average acc: 17857 / 19962 =  89.455
Robust  acc:     3 /   180 =   1.667
------------------------------------
Epoch:  30 | Train Loss: 0.001 | Train Acc: 88.478 | Val Loss: 0.002 | Val Acc: 85.302
Training:
Accuracies by groups:
0, 0  acc: 71450 / 71629 =  99.750
0, 1  acc: 66849 / 66874 =  99.963
1, 0  acc:  5668 / 22880 =  24.773
1, 1  acc:    49 /  1387 =   3.533
--------------------------------------
Average acc: 144016 / 162770 =  88.478
Robust  acc:    49 /  1387 =   3.533
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8532 /  8535 =  99.965
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:   139 /  2874 =   4.836
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16947 / 19867 =  85.302
Robust  acc:     0 /   182 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 87.226
Robust Acc: 0.556 | Best Acc: 100.000
-------------------------------------
Training, Epoch 29:
Accuracies by groups:
0, 0  acc:  9766 /  9767 =  99.990
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:   110 /  2480 =   4.435
1, 1  acc:     1 /   180 =   0.556
------------------------------------
Average acc: 17412 / 19962 =  87.226
Robust  acc:     1 /   180 =   0.556
------------------------------------
Accuracies by groups:
0, 0  acc:  9766 /  9767 =  99.990
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:   110 /  2480 =   4.435
1, 1  acc:     1 /   180 =   0.556
------------------------------------
Average acc: 17412 / 19962 =  87.226
Robust  acc:     1 /   180 =   0.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9766 /  9767 =  99.990
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:   110 /  2480 =   4.435
1, 1  acc:     1 /   180 =   0.556
------------------------------------
Average acc: 17412 / 19962 =  87.226
Robust  acc:     1 /   180 =   0.556
------------------------------------
Epoch:  31 | Train Loss: 0.001 | Train Acc: 87.093 | Val Loss: 0.002 | Val Acc: 87.376
Training:
Accuracies by groups:
0, 0  acc: 71562 / 71629 =  99.906
0, 1  acc: 66867 / 66874 =  99.990
1, 0  acc:  3307 / 22880 =  14.454
1, 1  acc:    25 /  1387 =   1.802
--------------------------------------
Average acc: 141761 / 162770 =  87.093
Robust  acc:    25 /  1387 =   1.802
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8516 /  8535 =  99.777
0, 1  acc:  8275 /  8276 =  99.988
1, 0  acc:   562 /  2874 =  19.555
1, 1  acc:     6 /   182 =   3.297
------------------------------------
Average acc: 17359 / 19867 =  87.376
Robust  acc:     6 /   182 =   3.297
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 88.909
Robust Acc: 0.000 | Best Acc: 99.973
------------------------------------
Training, Epoch 30:
Accuracies by groups:
0, 0  acc:  9749 /  9767 =  99.816
0, 1  acc:  7533 /  7535 =  99.973
1, 0  acc:   466 /  2480 =  18.790
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17748 / 19962 =  88.909
Robust  acc:     0 /   180 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9749 /  9767 =  99.816
0, 1  acc:  7533 /  7535 =  99.973
1, 0  acc:   466 /  2480 =  18.790
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17748 / 19962 =  88.909
Robust  acc:     0 /   180 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9749 /  9767 =  99.816
0, 1  acc:  7533 /  7535 =  99.973
1, 0  acc:   466 /  2480 =  18.790
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17748 / 19962 =  88.909
Robust  acc:     0 /   180 =   0.000
------------------------------------
Epoch:  32 | Train Loss: 0.001 | Train Acc: 85.698 | Val Loss: 0.003 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71616 / 71629 =  99.982
0, 1  acc: 66873 / 66874 =  99.999
1, 0  acc:   994 / 22880 =   4.344
1, 1  acc:     7 /  1387 =   0.505
--------------------------------------
Average acc: 139490 / 162770 =  85.698
Robust  acc:     7 /  1387 =   0.505
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 31:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  33 | Train Loss: 0.001 | Train Acc: 85.142 | Val Loss: 0.002 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:    83 / 22880 =   0.363
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138586 / 162770 =  85.142
Robust  acc:     0 /  1387 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 32:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  34 | Train Loss: 0.001 | Train Acc: 85.094 | Val Loss: 0.002 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     4 / 22880 =   0.017
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138507 / 162770 =  85.094
Robust  acc:     0 /  1387 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 33:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  35 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.002 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 34:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  36 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.002 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 35:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  37 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.002 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 36:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  38 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.002 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 37:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  39 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.003 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 38:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  40 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.002 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 39:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  41 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.003 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 40:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  42 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.003 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 41:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  43 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.003 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 42:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  44 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.003 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 43:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  45 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.003 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 44:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  46 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.003 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 45:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  47 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.003 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 46:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  48 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.003 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 47:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  49 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.002 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 48:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:  50 | Train Loss: 0.001 | Train Acc: 85.091 | Val Loss: 0.002 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 71629 / 71629 = 100.000
0, 1  acc: 66874 / 66874 = 100.000
1, 0  acc:     0 / 22880 =   0.000
1, 1  acc:     0 /  1387 =   0.000
--------------------------------------
Average acc: 138503 / 162770 =  85.091
Robust  acc:     0 / 22880 =   0.000
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 49:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
replace: True
Checkpoint saved at ./model/celebA/config/cp-bias-end-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=49-cpre=-1-cpb=-1.pt
======
# Calculate probability ...
======
======
# Load biased model ...
======
Epoch:   1 | Train Loss: 0.005 | Train Acc: 60.529 | Val Loss: 0.003 | Val Acc: 84.618
Training:
Accuracies by groups:
0, 0  acc: 51310 / 51310 = 100.000
0, 1  acc: 47213 / 47213 = 100.000
1, 0  acc:     0 / 58265 =   0.000
1, 1  acc:     0 /  5982 =   0.000
-------------------------------------
Average acc: 98523 / 162770 =  60.529
Robust  acc:     0 / 58265 =   0.000
-------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8535 /  8535 = 100.000
0, 1  acc:  8276 /  8276 = 100.000
1, 0  acc:     0 /  2874 =   0.000
1, 1  acc:     0 /   182 =   0.000
------------------------------------
Average acc: 16811 / 19867 =  84.618
Robust  acc:     0 /  2874 =   0.000
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 86.675
Robust Acc: 0.000 | Best Acc: 100.000
-------------------------------------
Training, Epoch 0:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9767 /  9767 = 100.000
0, 1  acc:  7535 /  7535 = 100.000
1, 0  acc:     0 /  2480 =   0.000
1, 1  acc:     0 /   180 =   0.000
------------------------------------
Average acc: 17302 / 19962 =  86.675
Robust  acc:     0 /  2480 =   0.000
------------------------------------
Epoch:   2 | Train Loss: 0.004 | Train Acc: 75.881 | Val Loss: 0.003 | Val Acc: 89.279
Training:
Accuracies by groups:
0, 0  acc: 48549 / 50935 =  95.316
0, 1  acc: 46602 / 47453 =  98.207
1, 0  acc: 27065 / 58401 =  46.343
1, 1  acc:  1295 /  5981 =  21.652
--------------------------------------
Average acc: 123511 / 162770 =  75.881
Robust  acc:  1295 /  5981 =  21.652
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7246 /  8535 =  84.897
0, 1  acc:  7778 /  8276 =  93.983
1, 0  acc:  2586 /  2874 =  89.979
1, 1  acc:   127 /   182 =  69.780
------------------------------------
Average acc: 17737 / 19867 =  89.279
Robust  acc:   127 /   182 =  69.780
------------------------------------
New max robust acc: 69.78021978021978
- Saving best checkpoint at epoch 1
replace: True
Checkpoint saved at ./model/celebA/config/cp-debias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=1-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 90.687
Robust Acc: 60.556 | Best Acc: 94.532
-------------------------------------
Training, Epoch 1:
Accuracies by groups:
0, 0  acc:  8648 /  9767 =  88.543
0, 1  acc:  7123 /  7535 =  94.532
1, 0  acc:  2223 /  2480 =  89.637
1, 1  acc:   109 /   180 =  60.556
------------------------------------
Average acc: 18103 / 19962 =  90.687
Robust  acc:   109 /   180 =  60.556
------------------------------------
Accuracies by groups:
0, 0  acc:  8648 /  9767 =  88.543
0, 1  acc:  7123 /  7535 =  94.532
1, 0  acc:  2223 /  2480 =  89.637
1, 1  acc:   109 /   180 =  60.556
------------------------------------
Average acc: 18103 / 19962 =  90.687
Robust  acc:   109 /   180 =  60.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8648 /  9767 =  88.543
0, 1  acc:  7123 /  7535 =  94.532
1, 0  acc:  2223 /  2480 =  89.637
1, 1  acc:   109 /   180 =  60.556
------------------------------------
Average acc: 18103 / 19962 =  90.687
Robust  acc:   109 /   180 =  60.556
------------------------------------
Epoch:   3 | Train Loss: 0.004 | Train Acc: 83.957 | Val Loss: 0.002 | Val Acc: 91.498
Training:
Accuracies by groups:
0, 0  acc: 45958 / 51314 =  89.562
0, 1  acc: 44982 / 47313 =  95.073
1, 0  acc: 42738 / 58130 =  73.521
1, 1  acc:  2978 /  6013 =  49.526
--------------------------------------
Average acc: 136656 / 162770 =  83.957
Robust  acc:  2978 /  6013 =  49.526
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8214 /  8535 =  96.239
0, 1  acc:  8219 /  8276 =  99.311
1, 0  acc:  1706 /  2874 =  59.360
1, 1  acc:    39 /   182 =  21.429
------------------------------------
Average acc: 18178 / 19867 =  91.498
Robust  acc:    39 /   182 =  21.429
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.851
Robust Acc: 22.222 | Best Acc: 99.496
-------------------------------------
Training, Epoch 2:
Accuracies by groups:
0, 0  acc:  9496 /  9767 =  97.225
0, 1  acc:  7497 /  7535 =  99.496
1, 0  acc:  1502 /  2480 =  60.565
1, 1  acc:    40 /   180 =  22.222
------------------------------------
Average acc: 18535 / 19962 =  92.851
Robust  acc:    40 /   180 =  22.222
------------------------------------
Accuracies by groups:
0, 0  acc:  9496 /  9767 =  97.225
0, 1  acc:  7497 /  7535 =  99.496
1, 0  acc:  1502 /  2480 =  60.565
1, 1  acc:    40 /   180 =  22.222
------------------------------------
Average acc: 18535 / 19962 =  92.851
Robust  acc:    40 /   180 =  22.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9496 /  9767 =  97.225
0, 1  acc:  7497 /  7535 =  99.496
1, 0  acc:  1502 /  2480 =  60.565
1, 1  acc:    40 /   180 =  22.222
------------------------------------
Average acc: 18535 / 19962 =  92.851
Robust  acc:    40 /   180 =  22.222
------------------------------------
Epoch:   4 | Train Loss: 0.003 | Train Acc: 85.547 | Val Loss: 0.003 | Val Acc: 82.549
Training:
Accuracies by groups:
0, 0  acc: 44571 / 51052 =  87.305
0, 1  acc: 44516 / 47416 =  93.884
1, 0  acc: 46721 / 58355 =  80.063
1, 1  acc:  3437 /  5947 =  57.794
--------------------------------------
Average acc: 139245 / 162770 =  85.547
Robust  acc:  3437 /  5947 =  57.794
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6425 /  8535 =  75.278
0, 1  acc:  7050 /  8276 =  85.186
1, 0  acc:  2762 /  2874 =  96.103
1, 1  acc:   163 /   182 =  89.560
------------------------------------
Average acc: 16400 / 19867 =  82.549
Robust  acc:  6425 /  8535 =  75.278
------------------------------------
New max robust acc: 75.27826596367898
- Saving best checkpoint at epoch 3
replace: True
-> Updating checkpoint cp-debias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=1-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-debias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=3-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 84.526
Robust Acc: 80.792 | Best Acc: 96.452
-------------------------------------
Training, Epoch 3:
Accuracies by groups:
0, 0  acc:  7891 /  9767 =  80.792
0, 1  acc:  6443 /  7535 =  85.508
1, 0  acc:  2392 /  2480 =  96.452
1, 1  acc:   147 /   180 =  81.667
------------------------------------
Average acc: 16873 / 19962 =  84.526
Robust  acc:  7891 /  9767 =  80.792
------------------------------------
Accuracies by groups:
0, 0  acc:  7891 /  9767 =  80.792
0, 1  acc:  6443 /  7535 =  85.508
1, 0  acc:  2392 /  2480 =  96.452
1, 1  acc:   147 /   180 =  81.667
------------------------------------
Average acc: 16873 / 19962 =  84.526
Robust  acc:  7891 /  9767 =  80.792
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7891 /  9767 =  80.792
0, 1  acc:  6443 /  7535 =  85.508
1, 0  acc:  2392 /  2480 =  96.452
1, 1  acc:   147 /   180 =  81.667
------------------------------------
Average acc: 16873 / 19962 =  84.526
Robust  acc:  7891 /  9767 =  80.792
------------------------------------
Epoch:   5 | Train Loss: 0.003 | Train Acc: 86.262 | Val Loss: 0.004 | Val Acc: 71.702
Training:
Accuracies by groups:
0, 0  acc: 43890 / 50997 =  86.064
0, 1  acc: 44123 / 47280 =  93.323
1, 0  acc: 48707 / 58564 =  83.169
1, 1  acc:  3688 /  5929 =  62.203
--------------------------------------
Average acc: 140408 / 162770 =  86.262
Robust  acc:  3688 /  5929 =  62.203
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5050 /  8535 =  59.168
0, 1  acc:  6171 /  8276 =  74.565
1, 0  acc:  2846 /  2874 =  99.026
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 14245 / 19867 =  71.702
Robust  acc:  5050 /  8535 =  59.168
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 73.610
Robust Acc: 65.598 | Best Acc: 98.992
-------------------------------------
Training, Epoch 4:
Accuracies by groups:
0, 0  acc:  6407 /  9767 =  65.598
0, 1  acc:  5664 /  7535 =  75.169
1, 0  acc:  2455 /  2480 =  98.992
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 14694 / 19962 =  73.610
Robust  acc:  6407 /  9767 =  65.598
------------------------------------
Accuracies by groups:
0, 0  acc:  6407 /  9767 =  65.598
0, 1  acc:  5664 /  7535 =  75.169
1, 0  acc:  2455 /  2480 =  98.992
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 14694 / 19962 =  73.610
Robust  acc:  6407 /  9767 =  65.598
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6407 /  9767 =  65.598
0, 1  acc:  5664 /  7535 =  75.169
1, 0  acc:  2455 /  2480 =  98.992
1, 1  acc:   168 /   180 =  93.333
------------------------------------
Average acc: 14694 / 19962 =  73.610
Robust  acc:  6407 /  9767 =  65.598
------------------------------------
Epoch:   6 | Train Loss: 0.003 | Train Acc: 86.837 | Val Loss: 0.002 | Val Acc: 89.686
Training:
Accuracies by groups:
0, 0  acc: 43938 / 51154 =  85.894
0, 1  acc: 44576 / 47828 =  93.201
1, 0  acc: 48971 / 57843 =  84.662
1, 1  acc:  3859 /  5945 =  64.912
--------------------------------------
Average acc: 141344 / 162770 =  86.837
Robust  acc:  3859 /  5945 =  64.912
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7218 /  8535 =  84.569
0, 1  acc:  7850 /  8276 =  94.853
1, 0  acc:  2621 /  2874 =  91.197
1, 1  acc:   129 /   182 =  70.879
------------------------------------
Average acc: 17818 / 19867 =  89.686
Robust  acc:   129 /   182 =  70.879
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.013
Robust Acc: 66.111 | Best Acc: 95.129
-------------------------------------
Training, Epoch 5:
Accuracies by groups:
0, 0  acc:  8592 /  9767 =  87.970
0, 1  acc:  7168 /  7535 =  95.129
1, 0  acc:  2289 /  2480 =  92.298
1, 1  acc:   119 /   180 =  66.111
------------------------------------
Average acc: 18168 / 19962 =  91.013
Robust  acc:   119 /   180 =  66.111
------------------------------------
Accuracies by groups:
0, 0  acc:  8592 /  9767 =  87.970
0, 1  acc:  7168 /  7535 =  95.129
1, 0  acc:  2289 /  2480 =  92.298
1, 1  acc:   119 /   180 =  66.111
------------------------------------
Average acc: 18168 / 19962 =  91.013
Robust  acc:   119 /   180 =  66.111
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8592 /  9767 =  87.970
0, 1  acc:  7168 /  7535 =  95.129
1, 0  acc:  2289 /  2480 =  92.298
1, 1  acc:   119 /   180 =  66.111
------------------------------------
Average acc: 18168 / 19962 =  91.013
Robust  acc:   119 /   180 =  66.111
------------------------------------
Epoch:   7 | Train Loss: 0.003 | Train Acc: 87.016 | Val Loss: 0.003 | Val Acc: 85.811
Training:
Accuracies by groups:
0, 0  acc: 43740 / 51029 =  85.716
0, 1  acc: 44294 / 47520 =  93.211
1, 0  acc: 49795 / 58308 =  85.400
1, 1  acc:  3807 /  5913 =  64.384
--------------------------------------
Average acc: 141636 / 162770 =  87.016
Robust  acc:  3807 /  5913 =  64.384
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6674 /  8535 =  78.196
0, 1  acc:  7442 /  8276 =  89.923
1, 0  acc:  2775 /  2874 =  96.555
1, 1  acc:   157 /   182 =  86.264
------------------------------------
Average acc: 17048 / 19867 =  85.811
Robust  acc:  6674 /  8535 =  78.196
------------------------------------
New max robust acc: 78.19566490919743
- Saving best checkpoint at epoch 6
replace: True
-> Updating checkpoint cp-debias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=3-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-debias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=6-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 87.401
Robust Acc: 80.556 | Best Acc: 96.613
-------------------------------------
Training, Epoch 6:
Accuracies by groups:
0, 0  acc:  8100 /  9767 =  82.932
0, 1  acc:  6806 /  7535 =  90.325
1, 0  acc:  2396 /  2480 =  96.613
1, 1  acc:   145 /   180 =  80.556
------------------------------------
Average acc: 17447 / 19962 =  87.401
Robust  acc:   145 /   180 =  80.556
------------------------------------
Accuracies by groups:
0, 0  acc:  8100 /  9767 =  82.932
0, 1  acc:  6806 /  7535 =  90.325
1, 0  acc:  2396 /  2480 =  96.613
1, 1  acc:   145 /   180 =  80.556
------------------------------------
Average acc: 17447 / 19962 =  87.401
Robust  acc:   145 /   180 =  80.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8100 /  9767 =  82.932
0, 1  acc:  6806 /  7535 =  90.325
1, 0  acc:  2396 /  2480 =  96.613
1, 1  acc:   145 /   180 =  80.556
------------------------------------
Average acc: 17447 / 19962 =  87.401
Robust  acc:   145 /   180 =  80.556
------------------------------------
Epoch:   8 | Train Loss: 0.003 | Train Acc: 87.237 | Val Loss: 0.002 | Val Acc: 88.096
Training:
Accuracies by groups:
0, 0  acc: 43975 / 51295 =  85.730
0, 1  acc: 44107 / 47229 =  93.390
1, 0  acc: 49973 / 58239 =  85.807
1, 1  acc:  3940 /  6007 =  65.590
--------------------------------------
Average acc: 141995 / 162770 =  87.237
Robust  acc:  3940 /  6007 =  65.590
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7003 /  8535 =  82.050
0, 1  acc:  7601 /  8276 =  91.844
1, 0  acc:  2744 /  2874 =  95.477
1, 1  acc:   154 /   182 =  84.615
------------------------------------
Average acc: 17502 / 19867 =  88.096
Robust  acc:  7003 /  8535 =  82.050
------------------------------------
New max robust acc: 82.05038078500293
- Saving best checkpoint at epoch 7
replace: True
-> Updating checkpoint cp-debias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=6-cpre=-1-cpb=-1.pt...
Checkpoint saved at ./model/celebA/config/cp-debias-wga-best-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=7-cpre=-1-cpb=-1.pt
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 89.315
Robust Acc: 78.333 | Best Acc: 95.161
-------------------------------------
Training, Epoch 7:
Accuracies by groups:
0, 0  acc:  8365 /  9767 =  85.646
0, 1  acc:  6963 /  7535 =  92.409
1, 0  acc:  2360 /  2480 =  95.161
1, 1  acc:   141 /   180 =  78.333
------------------------------------
Average acc: 17829 / 19962 =  89.315
Robust  acc:   141 /   180 =  78.333
------------------------------------
Accuracies by groups:
0, 0  acc:  8365 /  9767 =  85.646
0, 1  acc:  6963 /  7535 =  92.409
1, 0  acc:  2360 /  2480 =  95.161
1, 1  acc:   141 /   180 =  78.333
------------------------------------
Average acc: 17829 / 19962 =  89.315
Robust  acc:   141 /   180 =  78.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8365 /  9767 =  85.646
0, 1  acc:  6963 /  7535 =  92.409
1, 0  acc:  2360 /  2480 =  95.161
1, 1  acc:   141 /   180 =  78.333
------------------------------------
Average acc: 17829 / 19962 =  89.315
Robust  acc:   141 /   180 =  78.333
------------------------------------
Epoch:   9 | Train Loss: 0.003 | Train Acc: 87.534 | Val Loss: 0.003 | Val Acc: 84.834
Training:
Accuracies by groups:
0, 0  acc: 43817 / 51013 =  85.894
0, 1  acc: 44055 / 47180 =  93.376
1, 0  acc: 50679 / 58679 =  86.367
1, 1  acc:  3928 /  5898 =  66.599
--------------------------------------
Average acc: 142479 / 162770 =  87.534
Robust  acc:  3928 /  5898 =  66.599
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6590 /  8535 =  77.211
0, 1  acc:  7370 /  8276 =  89.053
1, 0  acc:  2741 /  2874 =  95.372
1, 1  acc:   153 /   182 =  84.066
------------------------------------
Average acc: 16854 / 19867 =  84.834
Robust  acc:  6590 /  8535 =  77.211
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 86.244
Robust Acc: 80.000 | Best Acc: 96.169
-------------------------------------
Training, Epoch 8:
Accuracies by groups:
0, 0  acc:  7967 /  9767 =  81.571
0, 1  acc:  6720 /  7535 =  89.184
1, 0  acc:  2385 /  2480 =  96.169
1, 1  acc:   144 /   180 =  80.000
------------------------------------
Average acc: 17216 / 19962 =  86.244
Robust  acc:   144 /   180 =  80.000
------------------------------------
Accuracies by groups:
0, 0  acc:  7967 /  9767 =  81.571
0, 1  acc:  6720 /  7535 =  89.184
1, 0  acc:  2385 /  2480 =  96.169
1, 1  acc:   144 /   180 =  80.000
------------------------------------
Average acc: 17216 / 19962 =  86.244
Robust  acc:   144 /   180 =  80.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7967 /  9767 =  81.571
0, 1  acc:  6720 /  7535 =  89.184
1, 0  acc:  2385 /  2480 =  96.169
1, 1  acc:   144 /   180 =  80.000
------------------------------------
Average acc: 17216 / 19962 =  86.244
Robust  acc:   144 /   180 =  80.000
------------------------------------
Epoch:  10 | Train Loss: 0.003 | Train Acc: 87.466 | Val Loss: 0.002 | Val Acc: 92.394
Training:
Accuracies by groups:
0, 0  acc: 44225 / 51673 =  85.586
0, 1  acc: 44235 / 47357 =  93.408
1, 0  acc: 49855 / 57738 =  86.347
1, 1  acc:  4053 /  6002 =  67.527
--------------------------------------
Average acc: 142368 / 162770 =  87.466
Robust  acc:  4053 /  6002 =  67.527
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7701 /  8535 =  90.228
0, 1  acc:  8010 /  8276 =  96.786
1, 0  acc:  2531 /  2874 =  88.065
1, 1  acc:   114 /   182 =  62.637
------------------------------------
Average acc: 18356 / 19867 =  92.394
Robust  acc:   114 /   182 =  62.637
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.543
Robust Acc: 60.000 | Best Acc: 96.881
-------------------------------------
Training, Epoch 9:
Accuracies by groups:
0, 0  acc:  9057 /  9767 =  92.731
0, 1  acc:  7300 /  7535 =  96.881
1, 0  acc:  2208 /  2480 =  89.032
1, 1  acc:   108 /   180 =  60.000
------------------------------------
Average acc: 18673 / 19962 =  93.543
Robust  acc:   108 /   180 =  60.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9057 /  9767 =  92.731
0, 1  acc:  7300 /  7535 =  96.881
1, 0  acc:  2208 /  2480 =  89.032
1, 1  acc:   108 /   180 =  60.000
------------------------------------
Average acc: 18673 / 19962 =  93.543
Robust  acc:   108 /   180 =  60.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9057 /  9767 =  92.731
0, 1  acc:  7300 /  7535 =  96.881
1, 0  acc:  2208 /  2480 =  89.032
1, 1  acc:   108 /   180 =  60.000
------------------------------------
Average acc: 18673 / 19962 =  93.543
Robust  acc:   108 /   180 =  60.000
------------------------------------
Epoch:  11 | Train Loss: 0.003 | Train Acc: 87.516 | Val Loss: 0.004 | Val Acc: 75.286
Training:
Accuracies by groups:
0, 0  acc: 43467 / 50784 =  85.592
0, 1  acc: 44658 / 47667 =  93.687
1, 0  acc: 50393 / 58362 =  86.346
1, 1  acc:  3932 /  5957 =  66.006
--------------------------------------
Average acc: 142450 / 162770 =  87.516
Robust  acc:  3932 /  5957 =  66.006
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5433 /  8535 =  63.656
0, 1  acc:  6498 /  8276 =  78.516
1, 0  acc:  2849 /  2874 =  99.130
1, 1  acc:   177 /   182 =  97.253
------------------------------------
Average acc: 14957 / 19867 =  75.286
Robust  acc:  5433 /  8535 =  63.656
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 76.846
Robust Acc: 69.069 | Best Acc: 98.952
-------------------------------------
Training, Epoch 10:
Accuracies by groups:
0, 0  acc:  6746 /  9767 =  69.069
0, 1  acc:  5973 /  7535 =  79.270
1, 0  acc:  2454 /  2480 =  98.952
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 15340 / 19962 =  76.846
Robust  acc:  6746 /  9767 =  69.069
------------------------------------
Accuracies by groups:
0, 0  acc:  6746 /  9767 =  69.069
0, 1  acc:  5973 /  7535 =  79.270
1, 0  acc:  2454 /  2480 =  98.952
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 15340 / 19962 =  76.846
Robust  acc:  6746 /  9767 =  69.069
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6746 /  9767 =  69.069
0, 1  acc:  5973 /  7535 =  79.270
1, 0  acc:  2454 /  2480 =  98.952
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 15340 / 19962 =  76.846
Robust  acc:  6746 /  9767 =  69.069
------------------------------------
Epoch:  12 | Train Loss: 0.003 | Train Acc: 87.611 | Val Loss: 0.002 | Val Acc: 91.519
Training:
Accuracies by groups:
0, 0  acc: 43582 / 50837 =  85.729
0, 1  acc: 44510 / 47544 =  93.619
1, 0  acc: 50522 / 58295 =  86.666
1, 1  acc:  3990 /  6094 =  65.474
--------------------------------------
Average acc: 142604 / 162770 =  87.611
Robust  acc:  3990 /  6094 =  65.474
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7422 /  8535 =  86.960
0, 1  acc:  7978 /  8276 =  96.399
1, 0  acc:  2657 /  2874 =  92.450
1, 1  acc:   125 /   182 =  68.681
------------------------------------
Average acc: 18182 / 19867 =  91.519
Robust  acc:   125 /   182 =  68.681
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.606
Robust Acc: 67.778 | Best Acc: 96.403
-------------------------------------
Training, Epoch 11:
Accuracies by groups:
0, 0  acc:  8801 /  9767 =  90.110
0, 1  acc:  7264 /  7535 =  96.403
1, 0  acc:  2299 /  2480 =  92.702
1, 1  acc:   122 /   180 =  67.778
------------------------------------
Average acc: 18486 / 19962 =  92.606
Robust  acc:   122 /   180 =  67.778
------------------------------------
Accuracies by groups:
0, 0  acc:  8801 /  9767 =  90.110
0, 1  acc:  7264 /  7535 =  96.403
1, 0  acc:  2299 /  2480 =  92.702
1, 1  acc:   122 /   180 =  67.778
------------------------------------
Average acc: 18486 / 19962 =  92.606
Robust  acc:   122 /   180 =  67.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8801 /  9767 =  90.110
0, 1  acc:  7264 /  7535 =  96.403
1, 0  acc:  2299 /  2480 =  92.702
1, 1  acc:   122 /   180 =  67.778
------------------------------------
Average acc: 18486 / 19962 =  92.606
Robust  acc:   122 /   180 =  67.778
------------------------------------
Epoch:  13 | Train Loss: 0.003 | Train Acc: 87.762 | Val Loss: 0.004 | Val Acc: 73.056
Training:
Accuracies by groups:
0, 0  acc: 43805 / 51106 =  85.714
0, 1  acc: 44485 / 47457 =  93.737
1, 0  acc: 50587 / 58182 =  86.946
1, 1  acc:  3973 /  6025 =  65.942
--------------------------------------
Average acc: 142850 / 162770 =  87.762
Robust  acc:  3973 /  6025 =  65.942
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5216 /  8535 =  61.113
0, 1  acc:  6268 /  8276 =  75.737
1, 0  acc:  2852 /  2874 =  99.235
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 14514 / 19867 =  73.056
Robust  acc:  5216 /  8535 =  61.113
------------------------------------
-------------------------------------------
Avg Test Loss: 0.004 | Avg Test Acc: 74.917
Robust Acc: 67.134 | Best Acc: 99.234
-------------------------------------
Training, Epoch 12:
Accuracies by groups:
0, 0  acc:  6557 /  9767 =  67.134
0, 1  acc:  5770 /  7535 =  76.576
1, 0  acc:  2461 /  2480 =  99.234
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 14955 / 19962 =  74.917
Robust  acc:  6557 /  9767 =  67.134
------------------------------------
Accuracies by groups:
0, 0  acc:  6557 /  9767 =  67.134
0, 1  acc:  5770 /  7535 =  76.576
1, 0  acc:  2461 /  2480 =  99.234
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 14955 / 19962 =  74.917
Robust  acc:  6557 /  9767 =  67.134
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  6557 /  9767 =  67.134
0, 1  acc:  5770 /  7535 =  76.576
1, 0  acc:  2461 /  2480 =  99.234
1, 1  acc:   167 /   180 =  92.778
------------------------------------
Average acc: 14955 / 19962 =  74.917
Robust  acc:  6557 /  9767 =  67.134
------------------------------------
Epoch:  14 | Train Loss: 0.003 | Train Acc: 87.605 | Val Loss: 0.003 | Val Acc: 78.567
Training:
Accuracies by groups:
0, 0  acc: 43817 / 51267 =  85.468
0, 1  acc: 44236 / 47229 =  93.663
1, 0  acc: 50497 / 58153 =  86.835
1, 1  acc:  4044 /  6121 =  66.068
--------------------------------------
Average acc: 142594 / 162770 =  87.605
Robust  acc:  4044 /  6121 =  66.068
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5813 /  8535 =  68.108
0, 1  acc:  6792 /  8276 =  82.069
1, 0  acc:  2836 /  2874 =  98.678
1, 1  acc:   168 /   182 =  92.308
------------------------------------
Average acc: 15609 / 19867 =  78.567
Robust  acc:  5813 /  8535 =  68.108
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 80.593
Robust Acc: 74.250 | Best Acc: 98.306
-------------------------------------
Training, Epoch 13:
Accuracies by groups:
0, 0  acc:  7252 /  9767 =  74.250
0, 1  acc:  6239 /  7535 =  82.800
1, 0  acc:  2438 /  2480 =  98.306
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 16088 / 19962 =  80.593
Robust  acc:  7252 /  9767 =  74.250
------------------------------------
Accuracies by groups:
0, 0  acc:  7252 /  9767 =  74.250
0, 1  acc:  6239 /  7535 =  82.800
1, 0  acc:  2438 /  2480 =  98.306
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 16088 / 19962 =  80.593
Robust  acc:  7252 /  9767 =  74.250
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7252 /  9767 =  74.250
0, 1  acc:  6239 /  7535 =  82.800
1, 0  acc:  2438 /  2480 =  98.306
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 16088 / 19962 =  80.593
Robust  acc:  7252 /  9767 =  74.250
------------------------------------
Epoch:  15 | Train Loss: 0.003 | Train Acc: 87.845 | Val Loss: 0.002 | Val Acc: 88.896
Training:
Accuracies by groups:
0, 0  acc: 43978 / 51176 =  85.935
0, 1  acc: 44236 / 47301 =  93.520
1, 0  acc: 50787 / 58299 =  87.115
1, 1  acc:  3984 /  5994 =  66.466
--------------------------------------
Average acc: 142985 / 162770 =  87.845
Robust  acc:  3984 /  5994 =  66.466
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7076 /  8535 =  82.906
0, 1  acc:  7720 /  8276 =  93.282
1, 0  acc:  2727 /  2874 =  94.885
1, 1  acc:   138 /   182 =  75.824
------------------------------------
Average acc: 17661 / 19867 =  88.896
Robust  acc:   138 /   182 =  75.824
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 90.181
Robust Acc: 70.000 | Best Acc: 94.677
-------------------------------------
Training, Epoch 14:
Accuracies by groups:
0, 0  acc:  8465 /  9767 =  86.669
0, 1  acc:  7063 /  7535 =  93.736
1, 0  acc:  2348 /  2480 =  94.677
1, 1  acc:   126 /   180 =  70.000
------------------------------------
Average acc: 18002 / 19962 =  90.181
Robust  acc:   126 /   180 =  70.000
------------------------------------
Accuracies by groups:
0, 0  acc:  8465 /  9767 =  86.669
0, 1  acc:  7063 /  7535 =  93.736
1, 0  acc:  2348 /  2480 =  94.677
1, 1  acc:   126 /   180 =  70.000
------------------------------------
Average acc: 18002 / 19962 =  90.181
Robust  acc:   126 /   180 =  70.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8465 /  9767 =  86.669
0, 1  acc:  7063 /  7535 =  93.736
1, 0  acc:  2348 /  2480 =  94.677
1, 1  acc:   126 /   180 =  70.000
------------------------------------
Average acc: 18002 / 19962 =  90.181
Robust  acc:   126 /   180 =  70.000
------------------------------------
Epoch:  16 | Train Loss: 0.003 | Train Acc: 87.691 | Val Loss: 0.002 | Val Acc: 93.124
Training:
Accuracies by groups:
0, 0  acc: 43666 / 51183 =  85.313
0, 1  acc: 44179 / 47139 =  93.721
1, 0  acc: 50899 / 58455 =  87.074
1, 1  acc:  3991 /  5993 =  66.594
--------------------------------------
Average acc: 142735 / 162770 =  87.691
Robust  acc:  3991 /  5993 =  66.594
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8266 /  8535 =  96.848
0, 1  acc:  8239 /  8276 =  99.553
1, 0  acc:  1947 /  2874 =  67.745
1, 1  acc:    49 /   182 =  26.923
------------------------------------
Average acc: 18501 / 19867 =  93.124
Robust  acc:    49 /   182 =  26.923
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.974
Robust Acc: 25.556 | Best Acc: 99.681
-------------------------------------
Training, Epoch 15:
Accuracies by groups:
0, 0  acc:  9561 /  9767 =  97.891
0, 1  acc:  7511 /  7535 =  99.681
1, 0  acc:  1641 /  2480 =  66.169
1, 1  acc:    46 /   180 =  25.556
------------------------------------
Average acc: 18759 / 19962 =  93.974
Robust  acc:    46 /   180 =  25.556
------------------------------------
Accuracies by groups:
0, 0  acc:  9561 /  9767 =  97.891
0, 1  acc:  7511 /  7535 =  99.681
1, 0  acc:  1641 /  2480 =  66.169
1, 1  acc:    46 /   180 =  25.556
------------------------------------
Average acc: 18759 / 19962 =  93.974
Robust  acc:    46 /   180 =  25.556
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9561 /  9767 =  97.891
0, 1  acc:  7511 /  7535 =  99.681
1, 0  acc:  1641 /  2480 =  66.169
1, 1  acc:    46 /   180 =  25.556
------------------------------------
Average acc: 18759 / 19962 =  93.974
Robust  acc:    46 /   180 =  25.556
------------------------------------
Epoch:  17 | Train Loss: 0.003 | Train Acc: 87.811 | Val Loss: 0.002 | Val Acc: 93.220
Training:
Accuracies by groups:
0, 0  acc: 44000 / 51403 =  85.598
0, 1  acc: 43986 / 47001 =  93.585
1, 0  acc: 50891 / 58361 =  87.200
1, 1  acc:  4053 /  6005 =  67.494
--------------------------------------
Average acc: 142930 / 162770 =  87.811
Robust  acc:  4053 /  6005 =  67.494
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7837 /  8535 =  91.822
0, 1  acc:  8089 /  8276 =  97.740
1, 0  acc:  2495 /  2874 =  86.813
1, 1  acc:    99 /   182 =  54.396
------------------------------------
Average acc: 18520 / 19867 =  93.220
Robust  acc:    99 /   182 =  54.396
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.954
Robust Acc: 54.444 | Best Acc: 97.532
-------------------------------------
Training, Epoch 16:
Accuracies by groups:
0, 0  acc:  9177 /  9767 =  93.959
0, 1  acc:  7349 /  7535 =  97.532
1, 0  acc:  2131 /  2480 =  85.927
1, 1  acc:    98 /   180 =  54.444
------------------------------------
Average acc: 18755 / 19962 =  93.954
Robust  acc:    98 /   180 =  54.444
------------------------------------
Accuracies by groups:
0, 0  acc:  9177 /  9767 =  93.959
0, 1  acc:  7349 /  7535 =  97.532
1, 0  acc:  2131 /  2480 =  85.927
1, 1  acc:    98 /   180 =  54.444
------------------------------------
Average acc: 18755 / 19962 =  93.954
Robust  acc:    98 /   180 =  54.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9177 /  9767 =  93.959
0, 1  acc:  7349 /  7535 =  97.532
1, 0  acc:  2131 /  2480 =  85.927
1, 1  acc:    98 /   180 =  54.444
------------------------------------
Average acc: 18755 / 19962 =  93.954
Robust  acc:    98 /   180 =  54.444
------------------------------------
Epoch:  18 | Train Loss: 0.003 | Train Acc: 87.874 | Val Loss: 0.002 | Val Acc: 90.371
Training:
Accuracies by groups:
0, 0  acc: 43851 / 51333 =  85.425
0, 1  acc: 44310 / 47251 =  93.776
1, 0  acc: 50877 / 58197 =  87.422
1, 1  acc:  3995 /  5989 =  66.706
--------------------------------------
Average acc: 143033 / 162770 =  87.874
Robust  acc:  3995 /  5989 =  66.706
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7269 /  8535 =  85.167
0, 1  acc:  7873 /  8276 =  95.130
1, 0  acc:  2685 /  2874 =  93.424
1, 1  acc:   127 /   182 =  69.780
------------------------------------
Average acc: 17954 / 19867 =  90.371
Robust  acc:   127 /   182 =  69.780
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.479
Robust Acc: 67.778 | Best Acc: 94.758
-------------------------------------
Training, Epoch 17:
Accuracies by groups:
0, 0  acc:  8683 /  9767 =  88.901
0, 1  acc:  7140 /  7535 =  94.758
1, 0  acc:  2316 /  2480 =  93.387
1, 1  acc:   122 /   180 =  67.778
------------------------------------
Average acc: 18261 / 19962 =  91.479
Robust  acc:   122 /   180 =  67.778
------------------------------------
Accuracies by groups:
0, 0  acc:  8683 /  9767 =  88.901
0, 1  acc:  7140 /  7535 =  94.758
1, 0  acc:  2316 /  2480 =  93.387
1, 1  acc:   122 /   180 =  67.778
------------------------------------
Average acc: 18261 / 19962 =  91.479
Robust  acc:   122 /   180 =  67.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8683 /  9767 =  88.901
0, 1  acc:  7140 /  7535 =  94.758
1, 0  acc:  2316 /  2480 =  93.387
1, 1  acc:   122 /   180 =  67.778
------------------------------------
Average acc: 18261 / 19962 =  91.479
Robust  acc:   122 /   180 =  67.778
------------------------------------
Epoch:  19 | Train Loss: 0.003 | Train Acc: 88.040 | Val Loss: 0.002 | Val Acc: 87.280
Training:
Accuracies by groups:
0, 0  acc: 43887 / 51202 =  85.713
0, 1  acc: 44506 / 47382 =  93.930
1, 0  acc: 51028 / 58286 =  87.548
1, 1  acc:  3881 /  5900 =  65.780
--------------------------------------
Average acc: 143302 / 162770 =  88.040
Robust  acc:  3881 /  5900 =  65.780
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6833 /  8535 =  80.059
0, 1  acc:  7600 /  8276 =  91.832
1, 0  acc:  2754 /  2874 =  95.825
1, 1  acc:   153 /   182 =  84.066
------------------------------------
Average acc: 17340 / 19867 =  87.280
Robust  acc:  6833 /  8535 =  80.059
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 88.643
Robust Acc: 75.000 | Best Acc: 96.169
-------------------------------------
Training, Epoch 18:
Accuracies by groups:
0, 0  acc:  8224 /  9767 =  84.202
0, 1  acc:  6951 /  7535 =  92.250
1, 0  acc:  2385 /  2480 =  96.169
1, 1  acc:   135 /   180 =  75.000
------------------------------------
Average acc: 17695 / 19962 =  88.643
Robust  acc:   135 /   180 =  75.000
------------------------------------
Accuracies by groups:
0, 0  acc:  8224 /  9767 =  84.202
0, 1  acc:  6951 /  7535 =  92.250
1, 0  acc:  2385 /  2480 =  96.169
1, 1  acc:   135 /   180 =  75.000
------------------------------------
Average acc: 17695 / 19962 =  88.643
Robust  acc:   135 /   180 =  75.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8224 /  9767 =  84.202
0, 1  acc:  6951 /  7535 =  92.250
1, 0  acc:  2385 /  2480 =  96.169
1, 1  acc:   135 /   180 =  75.000
------------------------------------
Average acc: 17695 / 19962 =  88.643
Robust  acc:   135 /   180 =  75.000
------------------------------------
Epoch:  20 | Train Loss: 0.003 | Train Acc: 87.955 | Val Loss: 0.002 | Val Acc: 92.062
Training:
Accuracies by groups:
0, 0  acc: 44037 / 51313 =  85.820
0, 1  acc: 44650 / 47564 =  93.874
1, 0  acc: 50665 / 58042 =  87.290
1, 1  acc:  3812 /  5851 =  65.151
--------------------------------------
Average acc: 143164 / 162770 =  87.955
Robust  acc:  3812 /  5851 =  65.151
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7549 /  8535 =  88.448
0, 1  acc:  8064 /  8276 =  97.438
1, 0  acc:  2564 /  2874 =  89.214
1, 1  acc:   113 /   182 =  62.088
------------------------------------
Average acc: 18290 / 19867 =  92.062
Robust  acc:   113 /   182 =  62.088
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.232
Robust Acc: 52.778 | Best Acc: 97.651
-------------------------------------
Training, Epoch 19:
Accuracies by groups:
0, 0  acc:  8921 /  9767 =  91.338
0, 1  acc:  7358 /  7535 =  97.651
1, 0  acc:  2237 /  2480 =  90.202
1, 1  acc:    95 /   180 =  52.778
------------------------------------
Average acc: 18611 / 19962 =  93.232
Robust  acc:    95 /   180 =  52.778
------------------------------------
Accuracies by groups:
0, 0  acc:  8921 /  9767 =  91.338
0, 1  acc:  7358 /  7535 =  97.651
1, 0  acc:  2237 /  2480 =  90.202
1, 1  acc:    95 /   180 =  52.778
------------------------------------
Average acc: 18611 / 19962 =  93.232
Robust  acc:    95 /   180 =  52.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8921 /  9767 =  91.338
0, 1  acc:  7358 /  7535 =  97.651
1, 0  acc:  2237 /  2480 =  90.202
1, 1  acc:    95 /   180 =  52.778
------------------------------------
Average acc: 18611 / 19962 =  93.232
Robust  acc:    95 /   180 =  52.778
------------------------------------
Epoch:  21 | Train Loss: 0.003 | Train Acc: 87.861 | Val Loss: 0.002 | Val Acc: 90.915
Training:
Accuracies by groups:
0, 0  acc: 43745 / 51186 =  85.463
0, 1  acc: 44254 / 47294 =  93.572
1, 0  acc: 51134 / 58378 =  87.591
1, 1  acc:  3878 /  5912 =  65.595
--------------------------------------
Average acc: 143011 / 162770 =  87.861
Robust  acc:  3878 /  5912 =  65.595
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7391 /  8535 =  86.596
0, 1  acc:  7872 /  8276 =  95.118
1, 0  acc:  2667 /  2874 =  92.797
1, 1  acc:   132 /   182 =  72.527
------------------------------------
Average acc: 18062 / 19867 =  90.915
Robust  acc:   132 /   182 =  72.527
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.250
Robust Acc: 66.111 | Best Acc: 95.700
-------------------------------------
Training, Epoch 20:
Accuracies by groups:
0, 0  acc:  8762 /  9767 =  89.710
0, 1  acc:  7211 /  7535 =  95.700
1, 0  acc:  2323 /  2480 =  93.669
1, 1  acc:   119 /   180 =  66.111
------------------------------------
Average acc: 18415 / 19962 =  92.250
Robust  acc:   119 /   180 =  66.111
------------------------------------
Accuracies by groups:
0, 0  acc:  8762 /  9767 =  89.710
0, 1  acc:  7211 /  7535 =  95.700
1, 0  acc:  2323 /  2480 =  93.669
1, 1  acc:   119 /   180 =  66.111
------------------------------------
Average acc: 18415 / 19962 =  92.250
Robust  acc:   119 /   180 =  66.111
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8762 /  9767 =  89.710
0, 1  acc:  7211 /  7535 =  95.700
1, 0  acc:  2323 /  2480 =  93.669
1, 1  acc:   119 /   180 =  66.111
------------------------------------
Average acc: 18415 / 19962 =  92.250
Robust  acc:   119 /   180 =  66.111
------------------------------------
Epoch:  22 | Train Loss: 0.003 | Train Acc: 88.033 | Val Loss: 0.002 | Val Acc: 88.926
Training:
Accuracies by groups:
0, 0  acc: 43790 / 51204 =  85.521
0, 1  acc: 44439 / 47381 =  93.791
1, 0  acc: 51141 / 58319 =  87.692
1, 1  acc:  3922 /  5866 =  66.860
--------------------------------------
Average acc: 143292 / 162770 =  88.033
Robust  acc:  3922 /  5866 =  66.860
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7164 /  8535 =  83.937
0, 1  acc:  7650 /  8276 =  92.436
1, 0  acc:  2704 /  2874 =  94.085
1, 1  acc:   149 /   182 =  81.868
------------------------------------
Average acc: 17667 / 19867 =  88.926
Robust  acc:   149 /   182 =  81.868
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 89.961
Robust Acc: 76.111 | Best Acc: 94.194
-------------------------------------
Training, Epoch 21:
Accuracies by groups:
0, 0  acc:  8572 /  9767 =  87.765
0, 1  acc:  6913 /  7535 =  91.745
1, 0  acc:  2336 /  2480 =  94.194
1, 1  acc:   137 /   180 =  76.111
------------------------------------
Average acc: 17958 / 19962 =  89.961
Robust  acc:   137 /   180 =  76.111
------------------------------------
Accuracies by groups:
0, 0  acc:  8572 /  9767 =  87.765
0, 1  acc:  6913 /  7535 =  91.745
1, 0  acc:  2336 /  2480 =  94.194
1, 1  acc:   137 /   180 =  76.111
------------------------------------
Average acc: 17958 / 19962 =  89.961
Robust  acc:   137 /   180 =  76.111
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8572 /  9767 =  87.765
0, 1  acc:  6913 /  7535 =  91.745
1, 0  acc:  2336 /  2480 =  94.194
1, 1  acc:   137 /   180 =  76.111
------------------------------------
Average acc: 17958 / 19962 =  89.961
Robust  acc:   137 /   180 =  76.111
------------------------------------
Epoch:  23 | Train Loss: 0.003 | Train Acc: 88.159 | Val Loss: 0.002 | Val Acc: 88.509
Training:
Accuracies by groups:
0, 0  acc: 43915 / 51342 =  85.534
0, 1  acc: 44622 / 47472 =  93.996
1, 0  acc: 50947 / 58006 =  87.831
1, 1  acc:  4013 /  5950 =  67.445
--------------------------------------
Average acc: 143497 / 162770 =  88.159
Robust  acc:  4013 /  5950 =  67.445
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6962 /  8535 =  81.570
0, 1  acc:  7720 /  8276 =  93.282
1, 0  acc:  2743 /  2874 =  95.442
1, 1  acc:   159 /   182 =  87.363
------------------------------------
Average acc: 17584 / 19867 =  88.509
Robust  acc:  6962 /  8535 =  81.570
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 89.791
Robust Acc: 75.000 | Best Acc: 95.565
-------------------------------------
Training, Epoch 22:
Accuracies by groups:
0, 0  acc:  8370 /  9767 =  85.697
0, 1  acc:  7049 /  7535 =  93.550
1, 0  acc:  2370 /  2480 =  95.565
1, 1  acc:   135 /   180 =  75.000
------------------------------------
Average acc: 17924 / 19962 =  89.791
Robust  acc:   135 /   180 =  75.000
------------------------------------
Accuracies by groups:
0, 0  acc:  8370 /  9767 =  85.697
0, 1  acc:  7049 /  7535 =  93.550
1, 0  acc:  2370 /  2480 =  95.565
1, 1  acc:   135 /   180 =  75.000
------------------------------------
Average acc: 17924 / 19962 =  89.791
Robust  acc:   135 /   180 =  75.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8370 /  9767 =  85.697
0, 1  acc:  7049 /  7535 =  93.550
1, 0  acc:  2370 /  2480 =  95.565
1, 1  acc:   135 /   180 =  75.000
------------------------------------
Average acc: 17924 / 19962 =  89.791
Robust  acc:   135 /   180 =  75.000
------------------------------------
Epoch:  24 | Train Loss: 0.003 | Train Acc: 88.062 | Val Loss: 0.002 | Val Acc: 91.861
Training:
Accuracies by groups:
0, 0  acc: 43643 / 51276 =  85.114
0, 1  acc: 44460 / 47269 =  94.057
1, 0  acc: 51187 / 58198 =  87.953
1, 1  acc:  4049 /  6027 =  67.181
--------------------------------------
Average acc: 143339 / 162770 =  88.062
Robust  acc:  4049 /  6027 =  67.181
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7549 /  8535 =  88.448
0, 1  acc:  8007 /  8276 =  96.750
1, 0  acc:  2581 /  2874 =  89.805
1, 1  acc:   113 /   182 =  62.088
------------------------------------
Average acc: 18250 / 19867 =  91.861
Robust  acc:   113 /   182 =  62.088
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.037
Robust Acc: 52.222 | Best Acc: 97.133
-------------------------------------
Training, Epoch 23:
Accuracies by groups:
0, 0  acc:  8914 /  9767 =  91.267
0, 1  acc:  7319 /  7535 =  97.133
1, 0  acc:  2245 /  2480 =  90.524
1, 1  acc:    94 /   180 =  52.222
------------------------------------
Average acc: 18572 / 19962 =  93.037
Robust  acc:    94 /   180 =  52.222
------------------------------------
Accuracies by groups:
0, 0  acc:  8914 /  9767 =  91.267
0, 1  acc:  7319 /  7535 =  97.133
1, 0  acc:  2245 /  2480 =  90.524
1, 1  acc:    94 /   180 =  52.222
------------------------------------
Average acc: 18572 / 19962 =  93.037
Robust  acc:    94 /   180 =  52.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8914 /  9767 =  91.267
0, 1  acc:  7319 /  7535 =  97.133
1, 0  acc:  2245 /  2480 =  90.524
1, 1  acc:    94 /   180 =  52.222
------------------------------------
Average acc: 18572 / 19962 =  93.037
Robust  acc:    94 /   180 =  52.222
------------------------------------
Epoch:  25 | Train Loss: 0.003 | Train Acc: 88.037 | Val Loss: 0.002 | Val Acc: 93.134
Training:
Accuracies by groups:
0, 0  acc: 43652 / 51229 =  85.210
0, 1  acc: 44309 / 47253 =  93.770
1, 0  acc: 51204 / 58167 =  88.029
1, 1  acc:  4133 /  6121 =  67.522
--------------------------------------
Average acc: 143298 / 162770 =  88.037
Robust  acc:  4133 /  6121 =  67.522
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7761 /  8535 =  90.931
0, 1  acc:  8090 /  8276 =  97.753
1, 0  acc:  2546 /  2874 =  88.587
1, 1  acc:   106 /   182 =  58.242
------------------------------------
Average acc: 18503 / 19867 =  93.134
Robust  acc:   106 /   182 =  58.242
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 94.069
Robust Acc: 55.000 | Best Acc: 97.784
-------------------------------------
Training, Epoch 24:
Accuracies by groups:
0, 0  acc:  9119 /  9767 =  93.365
0, 1  acc:  7368 /  7535 =  97.784
1, 0  acc:  2192 /  2480 =  88.387
1, 1  acc:    99 /   180 =  55.000
------------------------------------
Average acc: 18778 / 19962 =  94.069
Robust  acc:    99 /   180 =  55.000
------------------------------------
Accuracies by groups:
0, 0  acc:  9119 /  9767 =  93.365
0, 1  acc:  7368 /  7535 =  97.784
1, 0  acc:  2192 /  2480 =  88.387
1, 1  acc:    99 /   180 =  55.000
------------------------------------
Average acc: 18778 / 19962 =  94.069
Robust  acc:    99 /   180 =  55.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9119 /  9767 =  93.365
0, 1  acc:  7368 /  7535 =  97.784
1, 0  acc:  2192 /  2480 =  88.387
1, 1  acc:    99 /   180 =  55.000
------------------------------------
Average acc: 18778 / 19962 =  94.069
Robust  acc:    99 /   180 =  55.000
------------------------------------
Epoch:  26 | Train Loss: 0.003 | Train Acc: 88.144 | Val Loss: 0.003 | Val Acc: 86.359
Training:
Accuracies by groups:
0, 0  acc: 43700 / 51212 =  85.332
0, 1  acc: 44001 / 46877 =  93.865
1, 0  acc: 51723 / 58655 =  88.182
1, 1  acc:  4048 /  6026 =  67.176
--------------------------------------
Average acc: 143472 / 162770 =  88.144
Robust  acc:  4048 /  6026 =  67.176
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6643 /  8535 =  77.832
0, 1  acc:  7552 /  8276 =  91.252
1, 0  acc:  2799 /  2874 =  97.390
1, 1  acc:   163 /   182 =  89.560
------------------------------------
Average acc: 17157 / 19867 =  86.359
Robust  acc:  6643 /  8535 =  77.832
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 87.762
Robust Acc: 82.502 | Best Acc: 97.177
-------------------------------------
Training, Epoch 25:
Accuracies by groups:
0, 0  acc:  8058 /  9767 =  82.502
0, 1  acc:  6900 /  7535 =  91.573
1, 0  acc:  2410 /  2480 =  97.177
1, 1  acc:   151 /   180 =  83.889
------------------------------------
Average acc: 17519 / 19962 =  87.762
Robust  acc:  8058 /  9767 =  82.502
------------------------------------
Accuracies by groups:
0, 0  acc:  8058 /  9767 =  82.502
0, 1  acc:  6900 /  7535 =  91.573
1, 0  acc:  2410 /  2480 =  97.177
1, 1  acc:   151 /   180 =  83.889
------------------------------------
Average acc: 17519 / 19962 =  87.762
Robust  acc:  8058 /  9767 =  82.502
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8058 /  9767 =  82.502
0, 1  acc:  6900 /  7535 =  91.573
1, 0  acc:  2410 /  2480 =  97.177
1, 1  acc:   151 /   180 =  83.889
------------------------------------
Average acc: 17519 / 19962 =  87.762
Robust  acc:  8058 /  9767 =  82.502
------------------------------------
Epoch:  27 | Train Loss: 0.002 | Train Acc: 88.243 | Val Loss: 0.003 | Val Acc: 78.930
Training:
Accuracies by groups:
0, 0  acc: 43658 / 51090 =  85.453
0, 1  acc: 44642 / 47578 =  93.829
1, 0  acc: 51147 / 58045 =  88.116
1, 1  acc:  4186 /  6057 =  69.110
--------------------------------------
Average acc: 143633 / 162770 =  88.243
Robust  acc:  4186 /  6057 =  69.110
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5886 /  8535 =  68.963
0, 1  acc:  6767 /  8276 =  81.767
1, 0  acc:  2851 /  2874 =  99.200
1, 1  acc:   177 /   182 =  97.253
------------------------------------
Average acc: 15681 / 19867 =  78.930
Robust  acc:  5886 /  8535 =  68.963
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 80.628
Robust Acc: 74.598 | Best Acc: 98.831
-------------------------------------
Training, Epoch 26:
Accuracies by groups:
0, 0  acc:  7286 /  9767 =  74.598
0, 1  acc:  6197 /  7535 =  82.243
1, 0  acc:  2451 /  2480 =  98.831
1, 1  acc:   161 /   180 =  89.444
------------------------------------
Average acc: 16095 / 19962 =  80.628
Robust  acc:  7286 /  9767 =  74.598
------------------------------------
Accuracies by groups:
0, 0  acc:  7286 /  9767 =  74.598
0, 1  acc:  6197 /  7535 =  82.243
1, 0  acc:  2451 /  2480 =  98.831
1, 1  acc:   161 /   180 =  89.444
------------------------------------
Average acc: 16095 / 19962 =  80.628
Robust  acc:  7286 /  9767 =  74.598
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7286 /  9767 =  74.598
0, 1  acc:  6197 /  7535 =  82.243
1, 0  acc:  2451 /  2480 =  98.831
1, 1  acc:   161 /   180 =  89.444
------------------------------------
Average acc: 16095 / 19962 =  80.628
Robust  acc:  7286 /  9767 =  74.598
------------------------------------
Epoch:  28 | Train Loss: 0.002 | Train Acc: 88.414 | Val Loss: 0.003 | Val Acc: 83.505
Training:
Accuracies by groups:
0, 0  acc: 43772 / 51016 =  85.801
0, 1  acc: 44452 / 47293 =  93.993
1, 0  acc: 51672 / 58466 =  88.380
1, 1  acc:  4016 /  5995 =  66.989
--------------------------------------
Average acc: 143912 / 162770 =  88.414
Robust  acc:  4016 /  5995 =  66.989
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6343 /  8535 =  74.318
0, 1  acc:  7269 /  8276 =  87.832
1, 0  acc:  2810 /  2874 =  97.773
1, 1  acc:   168 /   182 =  92.308
------------------------------------
Average acc: 16590 / 19867 =  83.505
Robust  acc:  6343 /  8535 =  74.318
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 85.432
Robust Acc: 79.922 | Best Acc: 97.581
-------------------------------------
Training, Epoch 27:
Accuracies by groups:
0, 0  acc:  7806 /  9767 =  79.922
0, 1  acc:  6684 /  7535 =  88.706
1, 0  acc:  2420 /  2480 =  97.581
1, 1  acc:   144 /   180 =  80.000
------------------------------------
Average acc: 17054 / 19962 =  85.432
Robust  acc:  7806 /  9767 =  79.922
------------------------------------
Accuracies by groups:
0, 0  acc:  7806 /  9767 =  79.922
0, 1  acc:  6684 /  7535 =  88.706
1, 0  acc:  2420 /  2480 =  97.581
1, 1  acc:   144 /   180 =  80.000
------------------------------------
Average acc: 17054 / 19962 =  85.432
Robust  acc:  7806 /  9767 =  79.922
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7806 /  9767 =  79.922
0, 1  acc:  6684 /  7535 =  88.706
1, 0  acc:  2420 /  2480 =  97.581
1, 1  acc:   144 /   180 =  80.000
------------------------------------
Average acc: 17054 / 19962 =  85.432
Robust  acc:  7806 /  9767 =  79.922
------------------------------------
Epoch:  29 | Train Loss: 0.002 | Train Acc: 88.343 | Val Loss: 0.002 | Val Acc: 92.576
Training:
Accuracies by groups:
0, 0  acc: 43627 / 51115 =  85.351
0, 1  acc: 44618 / 47413 =  94.105
1, 0  acc: 51549 / 58277 =  88.455
1, 1  acc:  4002 /  5965 =  67.091
--------------------------------------
Average acc: 143796 / 162770 =  88.343
Robust  acc:  4002 /  5965 =  67.091
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8401 /  8535 =  98.430
0, 1  acc:  8266 /  8276 =  99.879
1, 0  acc:  1700 /  2874 =  59.151
1, 1  acc:    25 /   182 =  13.736
------------------------------------
Average acc: 18392 / 19867 =  92.576
Robust  acc:    25 /   182 =  13.736
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.982
Robust Acc: 14.444 | Best Acc: 99.881
-------------------------------------
Training, Epoch 28:
Accuracies by groups:
0, 0  acc:  9655 /  9767 =  98.853
0, 1  acc:  7526 /  7535 =  99.881
1, 0  acc:  1354 /  2480 =  54.597
1, 1  acc:    26 /   180 =  14.444
------------------------------------
Average acc: 18561 / 19962 =  92.982
Robust  acc:    26 /   180 =  14.444
------------------------------------
Accuracies by groups:
0, 0  acc:  9655 /  9767 =  98.853
0, 1  acc:  7526 /  7535 =  99.881
1, 0  acc:  1354 /  2480 =  54.597
1, 1  acc:    26 /   180 =  14.444
------------------------------------
Average acc: 18561 / 19962 =  92.982
Robust  acc:    26 /   180 =  14.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9655 /  9767 =  98.853
0, 1  acc:  7526 /  7535 =  99.881
1, 0  acc:  1354 /  2480 =  54.597
1, 1  acc:    26 /   180 =  14.444
------------------------------------
Average acc: 18561 / 19962 =  92.982
Robust  acc:    26 /   180 =  14.444
------------------------------------
Epoch:  30 | Train Loss: 0.002 | Train Acc: 88.193 | Val Loss: 0.002 | Val Acc: 93.300
Training:
Accuracies by groups:
0, 0  acc: 43806 / 51230 =  85.508
0, 1  acc: 44505 / 47302 =  94.087
1, 0  acc: 51181 / 58192 =  87.952
1, 1  acc:  4060 /  6046 =  67.152
--------------------------------------
Average acc: 143552 / 162770 =  88.193
Robust  acc:  4060 /  6046 =  67.152
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8084 /  8535 =  94.716
0, 1  acc:  8191 /  8276 =  98.973
1, 0  acc:  2192 /  2874 =  76.270
1, 1  acc:    69 /   182 =  37.912
------------------------------------
Average acc: 18536 / 19867 =  93.300
Robust  acc:    69 /   182 =  37.912
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.803
Robust Acc: 32.222 | Best Acc: 99.071
-------------------------------------
Training, Epoch 29:
Accuracies by groups:
0, 0  acc:  9392 /  9767 =  96.161
0, 1  acc:  7465 /  7535 =  99.071
1, 0  acc:  1810 /  2480 =  72.984
1, 1  acc:    58 /   180 =  32.222
------------------------------------
Average acc: 18725 / 19962 =  93.803
Robust  acc:    58 /   180 =  32.222
------------------------------------
Accuracies by groups:
0, 0  acc:  9392 /  9767 =  96.161
0, 1  acc:  7465 /  7535 =  99.071
1, 0  acc:  1810 /  2480 =  72.984
1, 1  acc:    58 /   180 =  32.222
------------------------------------
Average acc: 18725 / 19962 =  93.803
Robust  acc:    58 /   180 =  32.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9392 /  9767 =  96.161
0, 1  acc:  7465 /  7535 =  99.071
1, 0  acc:  1810 /  2480 =  72.984
1, 1  acc:    58 /   180 =  32.222
------------------------------------
Average acc: 18725 / 19962 =  93.803
Robust  acc:    58 /   180 =  32.222
------------------------------------
Epoch:  31 | Train Loss: 0.002 | Train Acc: 88.471 | Val Loss: 0.003 | Val Acc: 79.499
Training:
Accuracies by groups:
0, 0  acc: 43586 / 50928 =  85.584
0, 1  acc: 44596 / 47366 =  94.152
1, 0  acc: 51723 / 58438 =  88.509
1, 1  acc:  4100 /  6038 =  67.903
--------------------------------------
Average acc: 144005 / 162770 =  88.471
Robust  acc:  4100 /  6038 =  67.903
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5758 /  8535 =  67.463
0, 1  acc:  7059 /  8276 =  85.295
1, 0  acc:  2808 /  2874 =  97.704
1, 1  acc:   169 /   182 =  92.857
------------------------------------
Average acc: 15794 / 19867 =  79.499
Robust  acc:  5758 /  8535 =  67.463
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 81.695
Robust Acc: 74.035 | Best Acc: 97.984
-------------------------------------
Training, Epoch 30:
Accuracies by groups:
0, 0  acc:  7231 /  9767 =  74.035
0, 1  acc:  6499 /  7535 =  86.251
1, 0  acc:  2430 /  2480 =  97.984
1, 1  acc:   148 /   180 =  82.222
------------------------------------
Average acc: 16308 / 19962 =  81.695
Robust  acc:  7231 /  9767 =  74.035
------------------------------------
Accuracies by groups:
0, 0  acc:  7231 /  9767 =  74.035
0, 1  acc:  6499 /  7535 =  86.251
1, 0  acc:  2430 /  2480 =  97.984
1, 1  acc:   148 /   180 =  82.222
------------------------------------
Average acc: 16308 / 19962 =  81.695
Robust  acc:  7231 /  9767 =  74.035
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7231 /  9767 =  74.035
0, 1  acc:  6499 /  7535 =  86.251
1, 0  acc:  2430 /  2480 =  97.984
1, 1  acc:   148 /   180 =  82.222
------------------------------------
Average acc: 16308 / 19962 =  81.695
Robust  acc:  7231 /  9767 =  74.035
------------------------------------
Epoch:  32 | Train Loss: 0.002 | Train Acc: 88.500 | Val Loss: 0.002 | Val Acc: 91.609
Training:
Accuracies by groups:
0, 0  acc: 43966 / 51210 =  85.854
0, 1  acc: 44708 / 47559 =  94.005
1, 0  acc: 51418 / 58068 =  88.548
1, 1  acc:  3959 /  5933 =  66.728
--------------------------------------
Average acc: 144051 / 162770 =  88.500
Robust  acc:  3959 /  5933 =  66.728
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7401 /  8535 =  86.714
0, 1  acc:  7979 /  8276 =  96.411
1, 0  acc:  2696 /  2874 =  93.807
1, 1  acc:   124 /   182 =  68.132
------------------------------------
Average acc: 18200 / 19867 =  91.609
Robust  acc:   124 /   182 =  68.132
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 92.841
Robust Acc: 57.222 | Best Acc: 96.629
-------------------------------------
Training, Epoch 31:
Accuracies by groups:
0, 0  acc:  8830 /  9767 =  90.406
0, 1  acc:  7281 /  7535 =  96.629
1, 0  acc:  2319 /  2480 =  93.508
1, 1  acc:   103 /   180 =  57.222
------------------------------------
Average acc: 18533 / 19962 =  92.841
Robust  acc:   103 /   180 =  57.222
------------------------------------
Accuracies by groups:
0, 0  acc:  8830 /  9767 =  90.406
0, 1  acc:  7281 /  7535 =  96.629
1, 0  acc:  2319 /  2480 =  93.508
1, 1  acc:   103 /   180 =  57.222
------------------------------------
Average acc: 18533 / 19962 =  92.841
Robust  acc:   103 /   180 =  57.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8830 /  9767 =  90.406
0, 1  acc:  7281 /  7535 =  96.629
1, 0  acc:  2319 /  2480 =  93.508
1, 1  acc:   103 /   180 =  57.222
------------------------------------
Average acc: 18533 / 19962 =  92.841
Robust  acc:   103 /   180 =  57.222
------------------------------------
Epoch:  33 | Train Loss: 0.002 | Train Acc: 88.301 | Val Loss: 0.003 | Val Acc: 86.329
Training:
Accuracies by groups:
0, 0  acc: 43800 / 51284 =  85.407
0, 1  acc: 44535 / 47336 =  94.083
1, 0  acc: 51460 / 58197 =  88.424
1, 1  acc:  3932 /  5953 =  66.051
--------------------------------------
Average acc: 143727 / 162770 =  88.301
Robust  acc:  3932 /  5953 =  66.051
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6680 /  8535 =  78.266
0, 1  acc:  7542 /  8276 =  91.131
1, 0  acc:  2776 /  2874 =  96.590
1, 1  acc:   153 /   182 =  84.066
------------------------------------
Average acc: 17151 / 19867 =  86.329
Robust  acc:  6680 /  8535 =  78.266
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 87.992
Robust Acc: 73.889 | Best Acc: 96.694
-------------------------------------
Training, Epoch 32:
Accuracies by groups:
0, 0  acc:  8103 /  9767 =  82.963
0, 1  acc:  6931 /  7535 =  91.984
1, 0  acc:  2398 /  2480 =  96.694
1, 1  acc:   133 /   180 =  73.889
------------------------------------
Average acc: 17565 / 19962 =  87.992
Robust  acc:   133 /   180 =  73.889
------------------------------------
Accuracies by groups:
0, 0  acc:  8103 /  9767 =  82.963
0, 1  acc:  6931 /  7535 =  91.984
1, 0  acc:  2398 /  2480 =  96.694
1, 1  acc:   133 /   180 =  73.889
------------------------------------
Average acc: 17565 / 19962 =  87.992
Robust  acc:   133 /   180 =  73.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8103 /  9767 =  82.963
0, 1  acc:  6931 /  7535 =  91.984
1, 0  acc:  2398 /  2480 =  96.694
1, 1  acc:   133 /   180 =  73.889
------------------------------------
Average acc: 17565 / 19962 =  87.992
Robust  acc:   133 /   180 =  73.889
------------------------------------
Epoch:  34 | Train Loss: 0.002 | Train Acc: 88.410 | Val Loss: 0.002 | Val Acc: 90.960
Training:
Accuracies by groups:
0, 0  acc: 43626 / 51035 =  85.483
0, 1  acc: 44189 / 46981 =  94.057
1, 0  acc: 52012 / 58713 =  88.587
1, 1  acc:  4078 /  6041 =  67.505
--------------------------------------
Average acc: 143905 / 162770 =  88.410
Robust  acc:  4078 /  6041 =  67.505
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7321 /  8535 =  85.776
0, 1  acc:  7937 /  8276 =  95.904
1, 0  acc:  2681 /  2874 =  93.285
1, 1  acc:   132 /   182 =  72.527
------------------------------------
Average acc: 18071 / 19867 =  90.960
Robust  acc:   132 /   182 =  72.527
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.955
Robust Acc: 60.000 | Best Acc: 96.324
-------------------------------------
Training, Epoch 33:
Accuracies by groups:
0, 0  acc:  8697 /  9767 =  89.045
0, 1  acc:  7258 /  7535 =  96.324
1, 0  acc:  2293 /  2480 =  92.460
1, 1  acc:   108 /   180 =  60.000
------------------------------------
Average acc: 18356 / 19962 =  91.955
Robust  acc:   108 /   180 =  60.000
------------------------------------
Accuracies by groups:
0, 0  acc:  8697 /  9767 =  89.045
0, 1  acc:  7258 /  7535 =  96.324
1, 0  acc:  2293 /  2480 =  92.460
1, 1  acc:   108 /   180 =  60.000
------------------------------------
Average acc: 18356 / 19962 =  91.955
Robust  acc:   108 /   180 =  60.000
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8697 /  9767 =  89.045
0, 1  acc:  7258 /  7535 =  96.324
1, 0  acc:  2293 /  2480 =  92.460
1, 1  acc:   108 /   180 =  60.000
------------------------------------
Average acc: 18356 / 19962 =  91.955
Robust  acc:   108 /   180 =  60.000
------------------------------------
Epoch:  35 | Train Loss: 0.002 | Train Acc: 88.554 | Val Loss: 0.003 | Val Acc: 80.515
Training:
Accuracies by groups:
0, 0  acc: 43905 / 51123 =  85.881
0, 1  acc: 44784 / 47612 =  94.060
1, 0  acc: 51594 / 58146 =  88.732
1, 1  acc:  3856 /  5889 =  65.478
--------------------------------------
Average acc: 144139 / 162770 =  88.554
Robust  acc:  3856 /  5889 =  65.478
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6023 /  8535 =  70.568
0, 1  acc:  6955 /  8276 =  84.038
1, 0  acc:  2840 /  2874 =  98.817
1, 1  acc:   178 /   182 =  97.802
------------------------------------
Average acc: 15996 / 19867 =  80.515
Robust  acc:  6023 /  8535 =  70.568
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 82.271
Robust Acc: 76.369 | Best Acc: 98.508
-------------------------------------
Training, Epoch 34:
Accuracies by groups:
0, 0  acc:  7459 /  9767 =  76.369
0, 1  acc:  6367 /  7535 =  84.499
1, 0  acc:  2443 /  2480 =  98.508
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 16423 / 19962 =  82.271
Robust  acc:  7459 /  9767 =  76.369
------------------------------------
Accuracies by groups:
0, 0  acc:  7459 /  9767 =  76.369
0, 1  acc:  6367 /  7535 =  84.499
1, 0  acc:  2443 /  2480 =  98.508
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 16423 / 19962 =  82.271
Robust  acc:  7459 /  9767 =  76.369
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7459 /  9767 =  76.369
0, 1  acc:  6367 /  7535 =  84.499
1, 0  acc:  2443 /  2480 =  98.508
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 16423 / 19962 =  82.271
Robust  acc:  7459 /  9767 =  76.369
------------------------------------
Epoch:  36 | Train Loss: 0.002 | Train Acc: 88.479 | Val Loss: 0.002 | Val Acc: 92.379
Training:
Accuracies by groups:
0, 0  acc: 43730 / 51124 =  85.537
0, 1  acc: 44719 / 47514 =  94.118
1, 0  acc: 51611 / 58198 =  88.682
1, 1  acc:  3958 /  5934 =  66.700
--------------------------------------
Average acc: 144018 / 162770 =  88.479
Robust  acc:  3958 /  5934 =  66.700
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7629 /  8535 =  89.385
0, 1  acc:  7977 /  8276 =  96.387
1, 0  acc:  2628 /  2874 =  91.441
1, 1  acc:   119 /   182 =  65.385
------------------------------------
Average acc: 18353 / 19867 =  92.379
Robust  acc:   119 /   182 =  65.385
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.267
Robust Acc: 59.444 | Best Acc: 96.350
-------------------------------------
Training, Epoch 35:
Accuracies by groups:
0, 0  acc:  8983 /  9767 =  91.973
0, 1  acc:  7260 /  7535 =  96.350
1, 0  acc:  2268 /  2480 =  91.452
1, 1  acc:   107 /   180 =  59.444
------------------------------------
Average acc: 18618 / 19962 =  93.267
Robust  acc:   107 /   180 =  59.444
------------------------------------
Accuracies by groups:
0, 0  acc:  8983 /  9767 =  91.973
0, 1  acc:  7260 /  7535 =  96.350
1, 0  acc:  2268 /  2480 =  91.452
1, 1  acc:   107 /   180 =  59.444
------------------------------------
Average acc: 18618 / 19962 =  93.267
Robust  acc:   107 /   180 =  59.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8983 /  9767 =  91.973
0, 1  acc:  7260 /  7535 =  96.350
1, 0  acc:  2268 /  2480 =  91.452
1, 1  acc:   107 /   180 =  59.444
------------------------------------
Average acc: 18618 / 19962 =  93.267
Robust  acc:   107 /   180 =  59.444
------------------------------------
Epoch:  37 | Train Loss: 0.002 | Train Acc: 88.402 | Val Loss: 0.003 | Val Acc: 81.457
Training:
Accuracies by groups:
0, 0  acc: 43956 / 51418 =  85.488
0, 1  acc: 44207 / 47012 =  94.033
1, 0  acc: 51697 / 58376 =  88.559
1, 1  acc:  4032 /  5964 =  67.606
--------------------------------------
Average acc: 143892 / 162770 =  88.402
Robust  acc:  4032 /  5964 =  67.606
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6155 /  8535 =  72.115
0, 1  acc:  7029 /  8276 =  84.932
1, 0  acc:  2830 /  2874 =  98.469
1, 1  acc:   169 /   182 =  92.857
------------------------------------
Average acc: 16183 / 19867 =  81.457
Robust  acc:  6155 /  8535 =  72.115
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 82.817
Robust Acc: 77.148 | Best Acc: 98.185
-------------------------------------
Training, Epoch 36:
Accuracies by groups:
0, 0  acc:  7535 /  9767 =  77.148
0, 1  acc:  6408 /  7535 =  85.043
1, 0  acc:  2435 /  2480 =  98.185
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 16532 / 19962 =  82.817
Robust  acc:  7535 /  9767 =  77.148
------------------------------------
Accuracies by groups:
0, 0  acc:  7535 /  9767 =  77.148
0, 1  acc:  6408 /  7535 =  85.043
1, 0  acc:  2435 /  2480 =  98.185
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 16532 / 19962 =  82.817
Robust  acc:  7535 /  9767 =  77.148
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7535 /  9767 =  77.148
0, 1  acc:  6408 /  7535 =  85.043
1, 0  acc:  2435 /  2480 =  98.185
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 16532 / 19962 =  82.817
Robust  acc:  7535 /  9767 =  77.148
------------------------------------
Epoch:  38 | Train Loss: 0.002 | Train Acc: 88.433 | Val Loss: 0.002 | Val Acc: 90.522
Training:
Accuracies by groups:
0, 0  acc: 43462 / 50901 =  85.385
0, 1  acc: 44427 / 47218 =  94.089
1, 0  acc: 51966 / 58569 =  88.726
1, 1  acc:  4088 /  6082 =  67.215
--------------------------------------
Average acc: 143943 / 162770 =  88.433
Robust  acc:  4088 /  6082 =  67.215
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7332 /  8535 =  85.905
0, 1  acc:  7794 /  8276 =  94.176
1, 0  acc:  2713 /  2874 =  94.398
1, 1  acc:   145 /   182 =  79.670
------------------------------------
Average acc: 17984 / 19867 =  90.522
Robust  acc:   145 /   182 =  79.670
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.524
Robust Acc: 68.889 | Best Acc: 94.360
-------------------------------------
Training, Epoch 37:
Accuracies by groups:
0, 0  acc:  8708 /  9767 =  89.157
0, 1  acc:  7110 /  7535 =  94.360
1, 0  acc:  2328 /  2480 =  93.871
1, 1  acc:   124 /   180 =  68.889
------------------------------------
Average acc: 18270 / 19962 =  91.524
Robust  acc:   124 /   180 =  68.889
------------------------------------
Accuracies by groups:
0, 0  acc:  8708 /  9767 =  89.157
0, 1  acc:  7110 /  7535 =  94.360
1, 0  acc:  2328 /  2480 =  93.871
1, 1  acc:   124 /   180 =  68.889
------------------------------------
Average acc: 18270 / 19962 =  91.524
Robust  acc:   124 /   180 =  68.889
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8708 /  9767 =  89.157
0, 1  acc:  7110 /  7535 =  94.360
1, 0  acc:  2328 /  2480 =  93.871
1, 1  acc:   124 /   180 =  68.889
------------------------------------
Average acc: 18270 / 19962 =  91.524
Robust  acc:   124 /   180 =  68.889
------------------------------------
Epoch:  39 | Train Loss: 0.002 | Train Acc: 88.581 | Val Loss: 0.003 | Val Acc: 80.057
Training:
Accuracies by groups:
0, 0  acc: 43657 / 51071 =  85.483
0, 1  acc: 44620 / 47311 =  94.312
1, 0  acc: 51814 / 58411 =  88.706
1, 1  acc:  4092 /  5977 =  68.462
--------------------------------------
Average acc: 144183 / 162770 =  88.581
Robust  acc:  4092 /  5977 =  68.462
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  5953 /  8535 =  69.748
0, 1  acc:  6941 /  8276 =  83.869
1, 0  acc:  2834 /  2874 =  98.608
1, 1  acc:   177 /   182 =  97.253
------------------------------------
Average acc: 15905 / 19867 =  80.057
Robust  acc:  5953 /  8535 =  69.748
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 81.279
Robust Acc: 75.008 | Best Acc: 98.548
-------------------------------------
Training, Epoch 38:
Accuracies by groups:
0, 0  acc:  7326 /  9767 =  75.008
0, 1  acc:  6298 /  7535 =  83.583
1, 0  acc:  2444 /  2480 =  98.548
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 16225 / 19962 =  81.279
Robust  acc:  7326 /  9767 =  75.008
------------------------------------
Accuracies by groups:
0, 0  acc:  7326 /  9767 =  75.008
0, 1  acc:  6298 /  7535 =  83.583
1, 0  acc:  2444 /  2480 =  98.548
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 16225 / 19962 =  81.279
Robust  acc:  7326 /  9767 =  75.008
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7326 /  9767 =  75.008
0, 1  acc:  6298 /  7535 =  83.583
1, 0  acc:  2444 /  2480 =  98.548
1, 1  acc:   157 /   180 =  87.222
------------------------------------
Average acc: 16225 / 19962 =  81.279
Robust  acc:  7326 /  9767 =  75.008
------------------------------------
Epoch:  40 | Train Loss: 0.002 | Train Acc: 88.547 | Val Loss: 0.002 | Val Acc: 93.230
Training:
Accuracies by groups:
0, 0  acc: 44047 / 51477 =  85.566
0, 1  acc: 44352 / 47142 =  94.082
1, 0  acc: 51825 / 58394 =  88.751
1, 1  acc:  3904 /  5757 =  67.813
--------------------------------------
Average acc: 144128 / 162770 =  88.547
Robust  acc:  3904 /  5757 =  67.813
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7721 /  8535 =  90.463
0, 1  acc:  8118 /  8276 =  98.091
1, 0  acc:  2585 /  2874 =  89.944
1, 1  acc:    98 /   182 =  53.846
------------------------------------
Average acc: 18522 / 19867 =  93.230
Robust  acc:    98 /   182 =  53.846
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.959
Robust Acc: 49.444 | Best Acc: 98.129
-------------------------------------
Training, Epoch 39:
Accuracies by groups:
0, 0  acc:  9065 /  9767 =  92.813
0, 1  acc:  7394 /  7535 =  98.129
1, 0  acc:  2208 /  2480 =  89.032
1, 1  acc:    89 /   180 =  49.444
------------------------------------
Average acc: 18756 / 19962 =  93.959
Robust  acc:    89 /   180 =  49.444
------------------------------------
Accuracies by groups:
0, 0  acc:  9065 /  9767 =  92.813
0, 1  acc:  7394 /  7535 =  98.129
1, 0  acc:  2208 /  2480 =  89.032
1, 1  acc:    89 /   180 =  49.444
------------------------------------
Average acc: 18756 / 19962 =  93.959
Robust  acc:    89 /   180 =  49.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9065 /  9767 =  92.813
0, 1  acc:  7394 /  7535 =  98.129
1, 0  acc:  2208 /  2480 =  89.032
1, 1  acc:    89 /   180 =  49.444
------------------------------------
Average acc: 18756 / 19962 =  93.959
Robust  acc:    89 /   180 =  49.444
------------------------------------
Epoch:  41 | Train Loss: 0.002 | Train Acc: 88.657 | Val Loss: 0.002 | Val Acc: 93.577
Training:
Accuracies by groups:
0, 0  acc: 43968 / 51280 =  85.741
0, 1  acc: 44264 / 46987 =  94.205
1, 0  acc: 52099 / 58627 =  88.865
1, 1  acc:  3976 /  5876 =  67.665
--------------------------------------
Average acc: 144307 / 162770 =  88.657
Robust  acc:  3976 /  5876 =  67.665
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8117 /  8535 =  95.103
0, 1  acc:  8176 /  8276 =  98.792
1, 0  acc:  2237 /  2874 =  77.836
1, 1  acc:    61 /   182 =  33.516
------------------------------------
Average acc: 18591 / 19867 =  93.577
Robust  acc:    61 /   182 =  33.516
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 94.229
Robust Acc: 34.444 | Best Acc: 98.699
-------------------------------------
Training, Epoch 40:
Accuracies by groups:
0, 0  acc:  9424 /  9767 =  96.488
0, 1  acc:  7437 /  7535 =  98.699
1, 0  acc:  1887 /  2480 =  76.089
1, 1  acc:    62 /   180 =  34.444
------------------------------------
Average acc: 18810 / 19962 =  94.229
Robust  acc:    62 /   180 =  34.444
------------------------------------
Accuracies by groups:
0, 0  acc:  9424 /  9767 =  96.488
0, 1  acc:  7437 /  7535 =  98.699
1, 0  acc:  1887 /  2480 =  76.089
1, 1  acc:    62 /   180 =  34.444
------------------------------------
Average acc: 18810 / 19962 =  94.229
Robust  acc:    62 /   180 =  34.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9424 /  9767 =  96.488
0, 1  acc:  7437 /  7535 =  98.699
1, 0  acc:  1887 /  2480 =  76.089
1, 1  acc:    62 /   180 =  34.444
------------------------------------
Average acc: 18810 / 19962 =  94.229
Robust  acc:    62 /   180 =  34.444
------------------------------------
Epoch:  42 | Train Loss: 0.002 | Train Acc: 88.726 | Val Loss: 0.002 | Val Acc: 91.076
Training:
Accuracies by groups:
0, 0  acc: 43834 / 51240 =  85.546
0, 1  acc: 44714 / 47439 =  94.256
1, 0  acc: 51776 / 58073 =  89.157
1, 1  acc:  4096 /  6018 =  68.062
--------------------------------------
Average acc: 144420 / 162770 =  88.726
Robust  acc:  4096 /  6018 =  68.062
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7318 /  8535 =  85.741
0, 1  acc:  7961 /  8276 =  96.194
1, 0  acc:  2688 /  2874 =  93.528
1, 1  acc:   127 /   182 =  69.780
------------------------------------
Average acc: 18094 / 19867 =  91.076
Robust  acc:   127 /   182 =  69.780
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 91.930
Robust Acc: 58.333 | Best Acc: 96.563
-------------------------------------
Training, Epoch 41:
Accuracies by groups:
0, 0  acc:  8659 /  9767 =  88.656
0, 1  acc:  7276 /  7535 =  96.563
1, 0  acc:  2311 /  2480 =  93.185
1, 1  acc:   105 /   180 =  58.333
------------------------------------
Average acc: 18351 / 19962 =  91.930
Robust  acc:   105 /   180 =  58.333
------------------------------------
Accuracies by groups:
0, 0  acc:  8659 /  9767 =  88.656
0, 1  acc:  7276 /  7535 =  96.563
1, 0  acc:  2311 /  2480 =  93.185
1, 1  acc:   105 /   180 =  58.333
------------------------------------
Average acc: 18351 / 19962 =  91.930
Robust  acc:   105 /   180 =  58.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8659 /  9767 =  88.656
0, 1  acc:  7276 /  7535 =  96.563
1, 0  acc:  2311 /  2480 =  93.185
1, 1  acc:   105 /   180 =  58.333
------------------------------------
Average acc: 18351 / 19962 =  91.930
Robust  acc:   105 /   180 =  58.333
------------------------------------
Epoch:  43 | Train Loss: 0.002 | Train Acc: 88.629 | Val Loss: 0.002 | Val Acc: 87.653
Training:
Accuracies by groups:
0, 0  acc: 43784 / 51270 =  85.399
0, 1  acc: 44454 / 47115 =  94.352
1, 0  acc: 51905 / 58447 =  88.807
1, 1  acc:  4118 /  5938 =  69.350
--------------------------------------
Average acc: 144261 / 162770 =  88.629
Robust  acc:  4118 /  5938 =  69.350
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6865 /  8535 =  80.434
0, 1  acc:  7617 /  8276 =  92.037
1, 0  acc:  2772 /  2874 =  96.451
1, 1  acc:   160 /   182 =  87.912
------------------------------------
Average acc: 17414 / 19867 =  87.653
Robust  acc:  6865 /  8535 =  80.434
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 89.104
Robust Acc: 77.222 | Best Acc: 96.250
-------------------------------------
Training, Epoch 42:
Accuracies by groups:
0, 0  acc:  8291 /  9767 =  84.888
0, 1  acc:  6970 /  7535 =  92.502
1, 0  acc:  2387 /  2480 =  96.250
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 17787 / 19962 =  89.104
Robust  acc:   139 /   180 =  77.222
------------------------------------
Accuracies by groups:
0, 0  acc:  8291 /  9767 =  84.888
0, 1  acc:  6970 /  7535 =  92.502
1, 0  acc:  2387 /  2480 =  96.250
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 17787 / 19962 =  89.104
Robust  acc:   139 /   180 =  77.222
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8291 /  9767 =  84.888
0, 1  acc:  6970 /  7535 =  92.502
1, 0  acc:  2387 /  2480 =  96.250
1, 1  acc:   139 /   180 =  77.222
------------------------------------
Average acc: 17787 / 19962 =  89.104
Robust  acc:   139 /   180 =  77.222
------------------------------------
Epoch:  44 | Train Loss: 0.002 | Train Acc: 88.758 | Val Loss: 0.002 | Val Acc: 88.931
Training:
Accuracies by groups:
0, 0  acc: 43946 / 51263 =  85.727
0, 1  acc: 44567 / 47283 =  94.256
1, 0  acc: 51894 / 58246 =  89.095
1, 1  acc:  4065 /  5978 =  67.999
--------------------------------------
Average acc: 144472 / 162770 =  88.758
Robust  acc:  4065 /  5978 =  67.999
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6969 /  8535 =  81.652
0, 1  acc:  7815 /  8276 =  94.430
1, 0  acc:  2738 /  2874 =  95.268
1, 1  acc:   146 /   182 =  80.220
------------------------------------
Average acc: 17668 / 19867 =  88.931
Robust  acc:   146 /   182 =  80.220
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 90.196
Robust Acc: 71.667 | Best Acc: 95.121
-------------------------------------
Training, Epoch 43:
Accuracies by groups:
0, 0  acc:  8372 /  9767 =  85.717
0, 1  acc:  7145 /  7535 =  94.824
1, 0  acc:  2359 /  2480 =  95.121
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18005 / 19962 =  90.196
Robust  acc:   129 /   180 =  71.667
------------------------------------
Accuracies by groups:
0, 0  acc:  8372 /  9767 =  85.717
0, 1  acc:  7145 /  7535 =  94.824
1, 0  acc:  2359 /  2480 =  95.121
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18005 / 19962 =  90.196
Robust  acc:   129 /   180 =  71.667
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8372 /  9767 =  85.717
0, 1  acc:  7145 /  7535 =  94.824
1, 0  acc:  2359 /  2480 =  95.121
1, 1  acc:   129 /   180 =  71.667
------------------------------------
Average acc: 18005 / 19962 =  90.196
Robust  acc:   129 /   180 =  71.667
------------------------------------
Epoch:  45 | Train Loss: 0.002 | Train Acc: 88.709 | Val Loss: 0.002 | Val Acc: 92.571
Training:
Accuracies by groups:
0, 0  acc: 44055 / 51450 =  85.627
0, 1  acc: 44822 / 47470 =  94.422
1, 0  acc: 51576 / 58024 =  88.887
1, 1  acc:  3939 /  5826 =  67.611
--------------------------------------
Average acc: 144392 / 162770 =  88.709
Robust  acc:  3939 /  5826 =  67.611
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7634 /  8535 =  89.443
0, 1  acc:  8033 /  8276 =  97.064
1, 0  acc:  2614 /  2874 =  90.953
1, 1  acc:   110 /   182 =  60.440
------------------------------------
Average acc: 18391 / 19867 =  92.571
Robust  acc:   110 /   182 =  60.440
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.272
Robust Acc: 57.778 | Best Acc: 97.200
-------------------------------------
Training, Epoch 44:
Accuracies by groups:
0, 0  acc:  8971 /  9767 =  91.850
0, 1  acc:  7324 /  7535 =  97.200
1, 0  acc:  2220 /  2480 =  89.516
1, 1  acc:   104 /   180 =  57.778
------------------------------------
Average acc: 18619 / 19962 =  93.272
Robust  acc:   104 /   180 =  57.778
------------------------------------
Accuracies by groups:
0, 0  acc:  8971 /  9767 =  91.850
0, 1  acc:  7324 /  7535 =  97.200
1, 0  acc:  2220 /  2480 =  89.516
1, 1  acc:   104 /   180 =  57.778
------------------------------------
Average acc: 18619 / 19962 =  93.272
Robust  acc:   104 /   180 =  57.778
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  8971 /  9767 =  91.850
0, 1  acc:  7324 /  7535 =  97.200
1, 0  acc:  2220 /  2480 =  89.516
1, 1  acc:   104 /   180 =  57.778
------------------------------------
Average acc: 18619 / 19962 =  93.272
Robust  acc:   104 /   180 =  57.778
------------------------------------
Epoch:  46 | Train Loss: 0.002 | Train Acc: 88.560 | Val Loss: 0.003 | Val Acc: 81.638
Training:
Accuracies by groups:
0, 0  acc: 43614 / 51100 =  85.350
0, 1  acc: 44686 / 47429 =  94.217
1, 0  acc: 51862 / 58321 =  88.925
1, 1  acc:  3987 /  5920 =  67.348
--------------------------------------
Average acc: 144149 / 162770 =  88.560
Robust  acc:  3987 /  5920 =  67.348
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6100 /  8535 =  71.470
0, 1  acc:  7123 /  8276 =  86.068
1, 0  acc:  2824 /  2874 =  98.260
1, 1  acc:   172 /   182 =  94.505
------------------------------------
Average acc: 16219 / 19867 =  81.638
Robust  acc:  6100 /  8535 =  71.470
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 83.649
Robust Acc: 77.670 | Best Acc: 98.065
-------------------------------------
Training, Epoch 45:
Accuracies by groups:
0, 0  acc:  7586 /  9767 =  77.670
0, 1  acc:  6526 /  7535 =  86.609
1, 0  acc:  2432 /  2480 =  98.065
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 16698 / 19962 =  83.649
Robust  acc:  7586 /  9767 =  77.670
------------------------------------
Accuracies by groups:
0, 0  acc:  7586 /  9767 =  77.670
0, 1  acc:  6526 /  7535 =  86.609
1, 0  acc:  2432 /  2480 =  98.065
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 16698 / 19962 =  83.649
Robust  acc:  7586 /  9767 =  77.670
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7586 /  9767 =  77.670
0, 1  acc:  6526 /  7535 =  86.609
1, 0  acc:  2432 /  2480 =  98.065
1, 1  acc:   154 /   180 =  85.556
------------------------------------
Average acc: 16698 / 19962 =  83.649
Robust  acc:  7586 /  9767 =  77.670
------------------------------------
Epoch:  47 | Train Loss: 0.002 | Train Acc: 88.701 | Val Loss: 0.003 | Val Acc: 84.723
Training:
Accuracies by groups:
0, 0  acc: 43959 / 51345 =  85.615
0, 1  acc: 44691 / 47366 =  94.352
1, 0  acc: 51666 / 58109 =  88.912
1, 1  acc:  4062 /  5950 =  68.269
--------------------------------------
Average acc: 144378 / 162770 =  88.701
Robust  acc:  4062 /  5950 =  68.269
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6531 /  8535 =  76.520
0, 1  acc:  7315 /  8276 =  88.388
1, 0  acc:  2813 /  2874 =  97.878
1, 1  acc:   173 /   182 =  95.055
------------------------------------
Average acc: 16832 / 19867 =  84.723
Robust  acc:  6531 /  8535 =  76.520
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 85.983
Robust Acc: 80.854 | Best Acc: 97.863
-------------------------------------
Training, Epoch 46:
Accuracies by groups:
0, 0  acc:  7897 /  9767 =  80.854
0, 1  acc:  6694 /  7535 =  88.839
1, 0  acc:  2427 /  2480 =  97.863
1, 1  acc:   146 /   180 =  81.111
------------------------------------
Average acc: 17164 / 19962 =  85.983
Robust  acc:  7897 /  9767 =  80.854
------------------------------------
Accuracies by groups:
0, 0  acc:  7897 /  9767 =  80.854
0, 1  acc:  6694 /  7535 =  88.839
1, 0  acc:  2427 /  2480 =  97.863
1, 1  acc:   146 /   180 =  81.111
------------------------------------
Average acc: 17164 / 19962 =  85.983
Robust  acc:  7897 /  9767 =  80.854
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7897 /  9767 =  80.854
0, 1  acc:  6694 /  7535 =  88.839
1, 0  acc:  2427 /  2480 =  97.863
1, 1  acc:   146 /   180 =  81.111
------------------------------------
Average acc: 17164 / 19962 =  85.983
Robust  acc:  7897 /  9767 =  80.854
------------------------------------
Epoch:  48 | Train Loss: 0.002 | Train Acc: 88.626 | Val Loss: 0.002 | Val Acc: 93.975
Training:
Accuracies by groups:
0, 0  acc: 43865 / 51247 =  85.595
0, 1  acc: 44886 / 47528 =  94.441
1, 0  acc: 51627 / 58137 =  88.802
1, 1  acc:  3878 /  5858 =  66.200
--------------------------------------
Average acc: 144256 / 162770 =  88.626
Robust  acc:  3878 /  5858 =  66.200
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  8166 /  8535 =  95.677
0, 1  acc:  8217 /  8276 =  99.287
1, 0  acc:  2224 /  2874 =  77.383
1, 1  acc:    63 /   182 =  34.615
------------------------------------
Average acc: 18670 / 19867 =  93.975
Robust  acc:    63 /   182 =  34.615
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 94.600
Robust Acc: 38.333 | Best Acc: 99.244
-------------------------------------
Training, Epoch 47:
Accuracies by groups:
0, 0  acc:  9452 /  9767 =  96.775
0, 1  acc:  7478 /  7535 =  99.244
1, 0  acc:  1885 /  2480 =  76.008
1, 1  acc:    69 /   180 =  38.333
------------------------------------
Average acc: 18884 / 19962 =  94.600
Robust  acc:    69 /   180 =  38.333
------------------------------------
Accuracies by groups:
0, 0  acc:  9452 /  9767 =  96.775
0, 1  acc:  7478 /  7535 =  99.244
1, 0  acc:  1885 /  2480 =  76.008
1, 1  acc:    69 /   180 =  38.333
------------------------------------
Average acc: 18884 / 19962 =  94.600
Robust  acc:    69 /   180 =  38.333
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9452 /  9767 =  96.775
0, 1  acc:  7478 /  7535 =  99.244
1, 0  acc:  1885 /  2480 =  76.008
1, 1  acc:    69 /   180 =  38.333
------------------------------------
Average acc: 18884 / 19962 =  94.600
Robust  acc:    69 /   180 =  38.333
------------------------------------
Epoch:  49 | Train Loss: 0.002 | Train Acc: 88.627 | Val Loss: 0.002 | Val Acc: 92.933
Training:
Accuracies by groups:
0, 0  acc: 43368 / 50803 =  85.365
0, 1  acc: 44982 / 47650 =  94.401
1, 0  acc: 51782 / 58233 =  88.922
1, 1  acc:  4126 /  6084 =  67.817
--------------------------------------
Average acc: 144258 / 162770 =  88.627
Robust  acc:  4126 /  6084 =  67.817
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  7665 /  8535 =  89.807
0, 1  acc:  8099 /  8276 =  97.861
1, 0  acc:  2598 /  2874 =  90.397
1, 1  acc:   101 /   182 =  55.495
------------------------------------
Average acc: 18463 / 19867 =  92.933
Robust  acc:   101 /   182 =  55.495
------------------------------------
-------------------------------------------
Avg Test Loss: 0.002 | Avg Test Acc: 93.638
Robust Acc: 49.444 | Best Acc: 97.850
-------------------------------------
Training, Epoch 48:
Accuracies by groups:
0, 0  acc:  9011 /  9767 =  92.260
0, 1  acc:  7373 /  7535 =  97.850
1, 0  acc:  2219 /  2480 =  89.476
1, 1  acc:    89 /   180 =  49.444
------------------------------------
Average acc: 18692 / 19962 =  93.638
Robust  acc:    89 /   180 =  49.444
------------------------------------
Accuracies by groups:
0, 0  acc:  9011 /  9767 =  92.260
0, 1  acc:  7373 /  7535 =  97.850
1, 0  acc:  2219 /  2480 =  89.476
1, 1  acc:    89 /   180 =  49.444
------------------------------------
Average acc: 18692 / 19962 =  93.638
Robust  acc:    89 /   180 =  49.444
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  9011 /  9767 =  92.260
0, 1  acc:  7373 /  7535 =  97.850
1, 0  acc:  2219 /  2480 =  89.476
1, 1  acc:    89 /   180 =  49.444
------------------------------------
Average acc: 18692 / 19962 =  93.638
Robust  acc:    89 /   180 =  49.444
------------------------------------
Epoch:  50 | Train Loss: 0.002 | Train Acc: 88.720 | Val Loss: 0.003 | Val Acc: 82.217
Training:
Accuracies by groups:
0, 0  acc: 43759 / 51104 =  85.627
0, 1  acc: 44266 / 46926 =  94.332
1, 0  acc: 52399 / 58880 =  88.993
1, 1  acc:  3985 /  5860 =  68.003
--------------------------------------
Average acc: 144409 / 162770 =  88.720
Robust  acc:  3985 /  5860 =  68.003
--------------------------------------
Validating:
Accuracies by groups:
0, 0  acc:  6065 /  8535 =  71.060
0, 1  acc:  7257 /  8276 =  87.687
1, 0  acc:  2839 /  2874 =  98.782
1, 1  acc:   173 /   182 =  95.055
------------------------------------
Average acc: 16334 / 19867 =  82.217
Robust  acc:  6065 /  8535 =  71.060
------------------------------------
-------------------------------------------
Avg Test Loss: 0.003 | Avg Test Acc: 83.408
Robust Acc: 76.287 | Best Acc: 98.508
-------------------------------------
Training, Epoch 49:
Accuracies by groups:
0, 0  acc:  7451 /  9767 =  76.287
0, 1  acc:  6597 /  7535 =  87.551
1, 0  acc:  2443 /  2480 =  98.508
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 16650 / 19962 =  83.408
Robust  acc:  7451 /  9767 =  76.287
------------------------------------
Accuracies by groups:
0, 0  acc:  7451 /  9767 =  76.287
0, 1  acc:  6597 /  7535 =  87.551
1, 0  acc:  2443 /  2480 =  98.508
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 16650 / 19962 =  83.408
Robust  acc:  7451 /  9767 =  76.287
------------------------------------
Testing:
Accuracies by groups:
0, 0  acc:  7451 /  9767 =  76.287
0, 1  acc:  6597 /  7535 =  87.551
1, 0  acc:  2443 /  2480 =  98.508
1, 1  acc:   159 /   180 =  88.333
------------------------------------
Average acc: 16650 / 19962 =  83.408
Robust  acc:  7451 /  9767 =  76.287
------------------------------------
replace: True
Checkpoint saved at ./model/celebA/config/cp-debias-end-cnc-celebA-sw=re-na=32-np=32-nn=32-nne=0-tsr=1-t=0.05-bf=32-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=0-r=0-cpe=49-cpre=-1-cpb=-1.pt
