Add files using upload-large-folder tool

data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/eval_log_last.json

@@ -0,0 +1 @@
+{"eval/last/epoch": 19, "eval/last/id_best": 16, "eval/last/lr_best": 8.1e-05, "eval/last/wd_best": 0.05, "eval/last/train/loss": 2.171297788619995, "eval/last/train/acc": 0.35084052982574754, "eval/last/train/acc_std": 0.0022872729527584545, "eval/last/train/f1": 0.28647551045480546, "eval/last/train/f1_std": 0.0023214594409422125, "eval/last/validation/loss": 2.4222354888916016, "eval/last/validation/acc": 0.27150239940937615, "eval/last/validation/acc_std": 0.00519698230893839, "eval/last/validation/f1": 0.20057228105758518, "eval/last/validation/f1_std": 0.004624768163389933, "eval/last/test/loss": 2.397343635559082, "eval/last/test/acc": 0.2792207792207792, "eval/last/test/acc_std": 0.00515904547244289, "eval/last/test/f1": 0.20079560877860803, "eval/last/test/f1_std": 0.0048771565982414795, "eval/last/testid/loss": 2.3290913105010986, "eval/last/testid/acc": 0.29535376903797955, "eval/last/testid/acc_std": 0.0056612386805859494, "eval/last/testid/f1": 0.22700868756848122, "eval/last/testid/f1_std": 0.005186164866650084}

data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/eval_table_best.csv

@@ -0,0 +1,5 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,attn,nsd_cococlip,best,6,0.000156,0.05,20,"[0.52, 1.0]",train,2.252119541168213,0.3276683364577891,0.002220140903461018,0.2622455758398266,0.0022267072748040464
+flat_mae,patch,attn,nsd_cococlip,best,6,0.000156,0.05,20,"[0.52, 1.0]",validation,2.415285348892212,0.2737172388335179,0.005144654041092983,0.20614164787225583,0.00463394310568278
+flat_mae,patch,attn,nsd_cococlip,best,6,0.000156,0.05,20,"[0.52, 1.0]",test,2.3897087574005127,0.2795918367346939,0.00530092004652869,0.20199755080867374,0.004819614958486068
+flat_mae,patch,attn,nsd_cococlip,best,6,0.000156,0.05,20,"[0.52, 1.0]",testid,2.375211715698242,0.2791594370541739,0.0055198147521058685,0.21536086551405484,0.004982723637039659

data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/eval_table_last.csv

@@ -0,0 +1,5 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,attn,nsd_cococlip,last,19,8.1e-05,0.05,16,"[0.27, 1.0]",train,2.171297788619995,0.35084052982574754,0.0022872729527584545,0.28647551045480546,0.0023214594409422125
+flat_mae,patch,attn,nsd_cococlip,last,19,8.1e-05,0.05,16,"[0.27, 1.0]",validation,2.4222354888916016,0.27150239940937615,0.00519698230893839,0.20057228105758518,0.004624768163389933
+flat_mae,patch,attn,nsd_cococlip,last,19,8.1e-05,0.05,16,"[0.27, 1.0]",test,2.397343635559082,0.2792207792207792,0.00515904547244289,0.20079560877860803,0.0048771565982414795
+flat_mae,patch,attn,nsd_cococlip,last,19,8.1e-05,0.05,16,"[0.27, 1.0]",testid,2.3290913105010986,0.29535376903797955,0.0056612386805859494,0.22700868756848122,0.005186164866650084

data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/log.txt

@@ -0,0 +1,962 @@
+fMRI foundation model probe eval
+version: 0.1.dev65+g4003a1397
+sha: 6c01b606db98add5848cecd23e5d599250c0bf86, status: clean, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-02-24 20:16:39
+config:
+output_root: experiments/data_scaling/output
+name_prefix: eval_probe
+remote_root: null
+notes: data scaling experiment n200_2; eval v2 (nsd_cococlip patch attn)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n200_2/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: true
+  norm: true
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn
+model: flat_mae
+representation: patch
+classifier: attn
+dataset: nsd_cococlip
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=True, reg_tokens=0, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: nsd_cococlip (flat)
+train (n=32539):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'run', 'trial_id', 'nsd_id', 'category_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 32539
+}),
+    labels=[ 3  4  5  6 10 11 17 18 19 20 22 23 25 30 31 33 36 37 38 53 55 59 61 74],
+    counts=[1286 1180 1639 1868  834  824 1026 1042  913 1853 1503 2092 1001 1410
+  794 1241 1904 1872 2267 1428  889  904 1447 1322]
+)
+
+validation (n=5418):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'run', 'trial_id', 'nsd_id', 'category_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 5418
+}),
+    labels=[ 3  4  5  6 10 11 17 18 19 20 22 23 25 30 31 33 36 37 38 53 55 59 61 74],
+    counts=[197 161 276 345 126 142 143 185 112 295 285 387 169 250 159 193 316 334
+ 343 215 172 141 226 246]
+)
+
+test (n=5390):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'run', 'trial_id', 'nsd_id', 'category_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 5390
+}),
+    labels=[ 3  4  5  6 10 11 17 18 19 20 22 23 25 30 31 33 36 37 38 53 55 59 61 74],
+    counts=[202 172 274 298 144 180 134 182 186 293 218 343 165 185 140 177 346 333
+ 345 271 165 140 251 246]
+)
+
+testid (n=5187):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'run', 'trial_id', 'nsd_id', 'category_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 5187
+}),
+    labels=[ 3  4  5  6 10 11 17 18 19 20 22 23 25 30 31 33 36 37 38 53 55 59 61 74],
+    counts=[197 159 267 273 123 153 175 184 139 310 215 386 153 230 118 192 330 306
+ 349 223 143 127 249 186]
+)
+
+running backbone on example batch to get embedding dim
+embedding feature dim (patch): 768
+initializing sweep of classifier heads
+classifiers:
+ModuleList(
+  (0-48): 49 x AttnPoolClassifier(
+    (kv): Linear(in_features=768, out_features=1536, bias=True)
+    (linear): Linear(in_features=768, out_features=24, bias=True)
+  )
+)
+classifier params (train): 58.8M (58.8M)
+setting up optimizer
+total batch size: 128 = 64 bs per gpu x 2 accum
+lr: 3.00e-04
+full schedule: epochs = 20 (steps = 4000) (decay = True)
+warmup: epochs = 5 (steps = 1000)
+start training for 20 epochs
+train: [0]  [  0/400]  eta: 0:24:24  lr: nan  time: 3.6625  data: 3.1096  max mem: 21740
+train: [0]  [ 20/400]  eta: 0:03:53  lr: 0.000003  loss: 3.1849 (3.1796)  grad: 0.1680 (0.1715)  time: 0.4609  data: 0.0029  max mem: 22448
+train: [0]  [ 40/400]  eta: 0:03:12  lr: 0.000006  loss: 3.1686 (3.1709)  grad: 0.1680 (0.1697)  time: 0.4541  data: 0.0049  max mem: 22448
+train: [0]  [ 60/400]  eta: 0:02:52  lr: 0.000009  loss: 3.1621 (3.1690)  grad: 0.1653 (0.1687)  time: 0.4520  data: 0.0049  max mem: 22448
+train: [0]  [ 80/400]  eta: 0:02:37  lr: 0.000012  loss: 3.1665 (3.1672)  grad: 0.1639 (0.1671)  time: 0.4397  data: 0.0049  max mem: 22448
+train: [0]  [100/400]  eta: 0:02:25  lr: 0.000015  loss: 3.1567 (3.1656)  grad: 0.1564 (0.1655)  time: 0.4584  data: 0.0053  max mem: 22448
+train: [0]  [120/400]  eta: 0:02:14  lr: 0.000018  loss: 3.1508 (3.1632)  grad: 0.1498 (0.1631)  time: 0.4545  data: 0.0052  max mem: 22448
+train: [0]  [140/400]  eta: 0:02:03  lr: 0.000021  loss: 3.1542 (3.1621)  grad: 0.1522 (0.1629)  time: 0.4537  data: 0.0049  max mem: 22448
+train: [0]  [160/400]  eta: 0:01:54  lr: 0.000024  loss: 3.1515 (3.1587)  grad: 0.1684 (0.1639)  time: 0.4674  data: 0.0051  max mem: 22448
+train: [0]  [180/400]  eta: 0:01:43  lr: 0.000027  loss: 3.1258 (3.1560)  grad: 0.1655 (0.1635)  time: 0.4487  data: 0.0049  max mem: 22448
+train: [0]  [200/400]  eta: 0:01:34  lr: 0.000030  loss: 3.1364 (3.1549)  grad: 0.1492 (0.1621)  time: 0.4591  data: 0.0049  max mem: 22448
+train: [0]  [220/400]  eta: 0:01:24  lr: 0.000033  loss: 3.1409 (3.1537)  grad: 0.1540 (0.1616)  time: 0.4377  data: 0.0047  max mem: 22448
+train: [0]  [240/400]  eta: 0:01:14  lr: 0.000036  loss: 3.1343 (3.1521)  grad: 0.1564 (0.1613)  time: 0.4421  data: 0.0048  max mem: 22448
+train: [0]  [260/400]  eta: 0:01:04  lr: 0.000039  loss: 3.1244 (3.1500)  grad: 0.1529 (0.1608)  time: 0.4431  data: 0.0048  max mem: 22448
+train: [0]  [280/400]  eta: 0:00:55  lr: 0.000042  loss: 3.1098 (3.1469)  grad: 0.1528 (0.1605)  time: 0.4401  data: 0.0048  max mem: 22448
+train: [0]  [300/400]  eta: 0:00:46  lr: 0.000045  loss: 3.0939 (3.1424)  grad: 0.1534 (0.1606)  time: 0.4399  data: 0.0048  max mem: 22448
+train: [0]  [320/400]  eta: 0:00:36  lr: 0.000048  loss: 3.0794 (3.1391)  grad: 0.1653 (0.1613)  time: 0.4412  data: 0.0048  max mem: 22448
+train: [0]  [340/400]  eta: 0:00:27  lr: 0.000051  loss: 3.0884 (3.1363)  grad: 0.1655 (0.1615)  time: 0.4365  data: 0.0049  max mem: 22448
+train: [0]  [360/400]  eta: 0:00:18  lr: 0.000054  loss: 3.0753 (3.1326)  grad: 0.1691 (0.1624)  time: 0.4376  data: 0.0045  max mem: 22448
+train: [0]  [380/400]  eta: 0:00:09  lr: 0.000057  loss: 3.0690 (3.1290)  grad: 0.1794 (0.1632)  time: 0.4454  data: 0.0049  max mem: 22448
+train: [0]  [399/400]  eta: 0:00:00  lr: 0.000060  loss: 3.0676 (3.1266)  grad: 0.1825 (0.1640)  time: 0.4384  data: 0.0047  max mem: 22448
+train: [0] Total time: 0:03:02 (0.4562 s / it)
+train: [0] Summary: lr: 0.000060  loss: 3.0676 (3.1266)  grad: 0.1825 (0.1640)
+eval (validation): [0]  [ 0/85]  eta: 0:04:12    time: 2.9750  data: 2.7438  max mem: 22448
+eval (validation): [0]  [20/85]  eta: 0:00:30    time: 0.3380  data: 0.0040  max mem: 22448
+eval (validation): [0]  [40/85]  eta: 0:00:17    time: 0.3274  data: 0.0040  max mem: 22448
+eval (validation): [0]  [60/85]  eta: 0:00:09    time: 0.3242  data: 0.0042  max mem: 22448
+eval (validation): [0]  [80/85]  eta: 0:00:01    time: 0.3205  data: 0.0040  max mem: 22448
+eval (validation): [0]  [84/85]  eta: 0:00:00    time: 0.3099  data: 0.0039  max mem: 22448
+eval (validation): [0] Total time: 0:00:30 (0.3609 s / it)
+cv: [0]  best hparam: (31, 1.0) (045) ('045_lr3.1e+01_wd1.0e+00')  loss: 2.607  acc: 0.218  f1: 0.153
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-best.pth
+train: [1]  [  0/400]  eta: 0:21:01  lr: nan  time: 3.1546  data: 2.8181  max mem: 22448
+train: [1]  [ 20/400]  eta: 0:03:43  lr: 0.000063  loss: 3.0186 (3.0243)  grad: 0.1771 (0.1753)  time: 0.4606  data: 0.0035  max mem: 22448
+train: [1]  [ 40/400]  eta: 0:03:06  lr: 0.000066  loss: 3.0209 (3.0263)  grad: 0.1674 (0.1713)  time: 0.4441  data: 0.0050  max mem: 22448
+train: [1]  [ 60/400]  eta: 0:02:49  lr: 0.000069  loss: 3.0071 (3.0125)  grad: 0.1673 (0.1727)  time: 0.4538  data: 0.0051  max mem: 22448
+train: [1]  [ 80/400]  eta: 0:02:33  lr: 0.000072  loss: 2.9928 (3.0120)  grad: 0.1753 (0.1751)  time: 0.4278  data: 0.0047  max mem: 22448
+train: [1]  [100/400]  eta: 0:02:21  lr: 0.000075  loss: 3.0088 (3.0071)  grad: 0.1762 (0.1756)  time: 0.4310  data: 0.0047  max mem: 22448
+train: [1]  [120/400]  eta: 0:02:10  lr: 0.000078  loss: 2.9930 (3.0035)  grad: 0.1793 (0.1765)  time: 0.4467  data: 0.0050  max mem: 22448
+train: [1]  [140/400]  eta: 0:02:00  lr: 0.000081  loss: 2.9868 (3.0009)  grad: 0.1814 (0.1779)  time: 0.4377  data: 0.0048  max mem: 22448
+train: [1]  [160/400]  eta: 0:01:50  lr: 0.000084  loss: 2.9930 (3.0015)  grad: 0.1832 (0.1785)  time: 0.4380  data: 0.0048  max mem: 22448
+train: [1]  [180/400]  eta: 0:01:40  lr: 0.000087  loss: 2.9971 (3.0008)  grad: 0.1826 (0.1792)  time: 0.4488  data: 0.0048  max mem: 22448
+train: [1]  [200/400]  eta: 0:01:31  lr: 0.000090  loss: 2.9666 (2.9989)  grad: 0.1799 (0.1798)  time: 0.4372  data: 0.0050  max mem: 22448
+train: [1]  [220/400]  eta: 0:01:21  lr: 0.000093  loss: 2.9374 (2.9913)  grad: 0.1944 (0.1819)  time: 0.4330  data: 0.0049  max mem: 22448
+train: [1]  [240/400]  eta: 0:01:12  lr: 0.000096  loss: 2.9201 (2.9877)  grad: 0.1944 (0.1821)  time: 0.4394  data: 0.0049  max mem: 22448
+train: [1]  [260/400]  eta: 0:01:03  lr: 0.000099  loss: 2.9355 (2.9864)  grad: 0.1827 (0.1826)  time: 0.4422  data: 0.0049  max mem: 22448
+train: [1]  [280/400]  eta: 0:00:54  lr: 0.000102  loss: 2.9345 (2.9820)  grad: 0.1832 (0.1827)  time: 0.4350  data: 0.0048  max mem: 22448
+train: [1]  [300/400]  eta: 0:00:44  lr: 0.000105  loss: 2.9301 (2.9802)  grad: 0.1865 (0.1833)  time: 0.4325  data: 0.0048  max mem: 22448
+train: [1]  [320/400]  eta: 0:00:35  lr: 0.000108  loss: 2.9353 (2.9762)  grad: 0.1920 (0.1842)  time: 0.4404  data: 0.0050  max mem: 22448
+train: [1]  [340/400]  eta: 0:00:26  lr: 0.000111  loss: 2.8974 (2.9721)  grad: 0.1943 (0.1850)  time: 0.4509  data: 0.0049  max mem: 22448
+train: [1]  [360/400]  eta: 0:00:17  lr: 0.000114  loss: 2.9191 (2.9702)  grad: 0.1934 (0.1855)  time: 0.4364  data: 0.0048  max mem: 22448
+train: [1]  [380/400]  eta: 0:00:08  lr: 0.000117  loss: 2.9191 (2.9663)  grad: 0.1941 (0.1864)  time: 0.4307  data: 0.0051  max mem: 22448
+train: [1]  [399/400]  eta: 0:00:00  lr: 0.000120  loss: 2.9132 (2.9648)  grad: 0.2020 (0.1875)  time: 0.4409  data: 0.0050  max mem: 22448
+train: [1] Total time: 0:02:59 (0.4477 s / it)
+train: [1] Summary: lr: 0.000120  loss: 2.9132 (2.9648)  grad: 0.2020 (0.1875)
+eval (validation): [1]  [ 0/85]  eta: 0:04:37    time: 3.2594  data: 2.9670  max mem: 22448
+eval (validation): [1]  [20/85]  eta: 0:00:30    time: 0.3366  data: 0.0038  max mem: 22448
+eval (validation): [1]  [40/85]  eta: 0:00:18    time: 0.3353  data: 0.0038  max mem: 22448
+eval (validation): [1]  [60/85]  eta: 0:00:09    time: 0.3143  data: 0.0042  max mem: 22448
+eval (validation): [1]  [80/85]  eta: 0:00:01    time: 0.3323  data: 0.0044  max mem: 22448
+eval (validation): [1]  [84/85]  eta: 0:00:00    time: 0.3309  data: 0.0044  max mem: 22448
+eval (validation): [1] Total time: 0:00:31 (0.3670 s / it)
+cv: [1]  best hparam: (14, 1.0) (040) ('040_lr1.4e+01_wd1.0e+00')  loss: 2.516  acc: 0.243  f1: 0.166
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-best.pth
+train: [2]  [  0/400]  eta: 0:26:25  lr: nan  time: 3.9629  data: 3.5690  max mem: 22448
+train: [2]  [ 20/400]  eta: 0:03:52  lr: 0.000123  loss: 2.9222 (2.8929)  grad: 0.2230 (0.2215)  time: 0.4449  data: 0.0025  max mem: 22448
+train: [2]  [ 40/400]  eta: 0:03:11  lr: 0.000126  loss: 2.9127 (2.8953)  grad: 0.2169 (0.2176)  time: 0.4479  data: 0.0049  max mem: 22448
+train: [2]  [ 60/400]  eta: 0:02:50  lr: 0.000129  loss: 2.9118 (2.8954)  grad: 0.2075 (0.2136)  time: 0.4408  data: 0.0048  max mem: 22448
+train: [2]  [ 80/400]  eta: 0:02:36  lr: 0.000132  loss: 2.8790 (2.8911)  grad: 0.2049 (0.2132)  time: 0.4432  data: 0.0050  max mem: 22448
+train: [2]  [100/400]  eta: 0:02:24  lr: 0.000135  loss: 2.8554 (2.8837)  grad: 0.2121 (0.2135)  time: 0.4512  data: 0.0048  max mem: 22448
+train: [2]  [120/400]  eta: 0:02:12  lr: 0.000138  loss: 2.8554 (2.8774)  grad: 0.2181 (0.2156)  time: 0.4315  data: 0.0046  max mem: 22448
+train: [2]  [140/400]  eta: 0:02:01  lr: 0.000141  loss: 2.8687 (2.8807)  grad: 0.2269 (0.2183)  time: 0.4422  data: 0.0047  max mem: 22448
+train: [2]  [160/400]  eta: 0:01:51  lr: 0.000144  loss: 2.8665 (2.8769)  grad: 0.2431 (0.2224)  time: 0.4503  data: 0.0049  max mem: 22448
+train: [2]  [180/400]  eta: 0:01:41  lr: 0.000147  loss: 2.8530 (2.8728)  grad: 0.2437 (0.2246)  time: 0.4383  data: 0.0047  max mem: 22448
+train: [2]  [200/400]  eta: 0:01:32  lr: 0.000150  loss: 2.8682 (2.8714)  grad: 0.2474 (0.2311)  time: 0.4397  data: 0.0047  max mem: 22448
+train: [2]  [220/400]  eta: 0:01:22  lr: 0.000153  loss: 2.9370 (2.8890)  grad: 0.3308 (0.2646)  time: 0.4286  data: 0.0049  max mem: 22448
+WARNING: classifier 48 (50, 1.0) diverged (loss=66.15 > 63.56) at step 520. Freezing.
+train: [2]  [240/400]  eta: 0:01:12  lr: 0.000156  loss: 3.1591 (2.9481)  grad: 0.7622 (0.3475)  time: 0.4342  data: 0.0051  max mem: 22448
+train: [2]  [260/400]  eta: 0:01:03  lr: 0.000159  loss: 2.9171 (2.9392)  grad: 0.2338 (0.3372)  time: 0.4540  data: 0.0052  max mem: 22448
+train: [2]  [280/400]  eta: 0:00:54  lr: 0.000162  loss: 2.8333 (2.9320)  grad: 0.2300 (0.3306)  time: 0.4470  data: 0.0050  max mem: 22448
+train: [2]  [300/400]  eta: 0:00:45  lr: 0.000165  loss: 2.8333 (2.9247)  grad: 0.2352 (0.3241)  time: 0.4375  data: 0.0049  max mem: 22448
+train: [2]  [320/400]  eta: 0:00:36  lr: 0.000168  loss: 2.8334 (2.9184)  grad: 0.2275 (0.3178)  time: 0.4582  data: 0.0051  max mem: 22448
+train: [2]  [340/400]  eta: 0:00:27  lr: 0.000171  loss: 2.8334 (2.9147)  grad: 0.2332 (0.3132)  time: 0.4412  data: 0.0049  max mem: 22448
+train: [2]  [360/400]  eta: 0:00:18  lr: 0.000174  loss: 2.9109 (2.9175)  grad: 0.2451 (0.3204)  time: 0.4304  data: 0.0048  max mem: 22448
+train: [2]  [380/400]  eta: 0:00:09  lr: 0.000177  loss: 3.1165 (2.9485)  grad: 0.6340 (0.3645)  time: 0.4426  data: 0.0048  max mem: 22448
+WARNING: classifier 47 (43, 1.0) diverged (loss=68.26 > 63.56) at step 591. Freezing.
+train: [2]  [399/400]  eta: 0:00:00  lr: 0.000180  loss: 3.1165 (2.9489)  grad: 0.5983 (0.3654)  time: 0.4340  data: 0.0048  max mem: 22448
+train: [2] Total time: 0:03:00 (0.4513 s / it)
+train: [2] Summary: lr: 0.000180  loss: 3.1165 (2.9489)  grad: 0.5983 (0.3654)
+eval (validation): [2]  [ 0/85]  eta: 0:04:32    time: 3.2011  data: 2.9676  max mem: 22448
+eval (validation): [2]  [20/85]  eta: 0:00:32    time: 0.3598  data: 0.0040  max mem: 22448
+eval (validation): [2]  [40/85]  eta: 0:00:18    time: 0.3120  data: 0.0039  max mem: 22448
+eval (validation): [2]  [60/85]  eta: 0:00:09    time: 0.3436  data: 0.0045  max mem: 22448
+eval (validation): [2]  [80/85]  eta: 0:00:01    time: 0.3201  data: 0.0039  max mem: 22448
+eval (validation): [2]  [84/85]  eta: 0:00:00    time: 0.3108  data: 0.0039  max mem: 22448
+eval (validation): [2] Total time: 0:00:31 (0.3695 s / it)
+cv: [2]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 2.454  acc: 0.261  f1: 0.189
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-best.pth
+train: [3]  [  0/400]  eta: 0:22:09  lr: nan  time: 3.3226  data: 2.9903  max mem: 22448
+train: [3]  [ 20/400]  eta: 0:03:33  lr: 0.000183  loss: 2.7712 (2.7956)  grad: 0.2212 (0.2170)  time: 0.4249  data: 0.0039  max mem: 22448
+train: [3]  [ 40/400]  eta: 0:02:59  lr: 0.000186  loss: 2.8045 (2.8164)  grad: 0.2222 (0.2229)  time: 0.4298  data: 0.0036  max mem: 22448
+train: [3]  [ 60/400]  eta: 0:02:41  lr: 0.000189  loss: 2.7729 (2.8028)  grad: 0.2232 (0.2227)  time: 0.4281  data: 0.0049  max mem: 22448
+train: [3]  [ 80/400]  eta: 0:02:30  lr: 0.000192  loss: 2.7623 (2.8020)  grad: 0.2189 (0.2227)  time: 0.4510  data: 0.0049  max mem: 22448
+train: [3]  [100/400]  eta: 0:02:19  lr: 0.000195  loss: 2.7823 (2.7979)  grad: 0.2203 (0.2236)  time: 0.4405  data: 0.0046  max mem: 22448
+train: [3]  [120/400]  eta: 0:02:08  lr: 0.000198  loss: 2.7787 (2.7951)  grad: 0.2265 (0.2249)  time: 0.4322  data: 0.0049  max mem: 22448
+train: [3]  [140/400]  eta: 0:01:58  lr: 0.000201  loss: 2.7756 (2.7950)  grad: 0.2432 (0.2282)  time: 0.4515  data: 0.0049  max mem: 22448
+train: [3]  [160/400]  eta: 0:01:49  lr: 0.000204  loss: 2.7880 (2.7961)  grad: 0.2444 (0.2295)  time: 0.4354  data: 0.0049  max mem: 22448
+train: [3]  [180/400]  eta: 0:01:39  lr: 0.000207  loss: 2.7748 (2.7917)  grad: 0.2345 (0.2304)  time: 0.4315  data: 0.0050  max mem: 22448
+train: [3]  [200/400]  eta: 0:01:29  lr: 0.000210  loss: 2.7748 (2.7947)  grad: 0.2371 (0.2322)  time: 0.4295  data: 0.0049  max mem: 22448
+train: [3]  [220/400]  eta: 0:01:20  lr: 0.000213  loss: 2.7794 (2.7952)  grad: 0.2441 (0.2332)  time: 0.4336  data: 0.0049  max mem: 22448
+train: [3]  [240/400]  eta: 0:01:11  lr: 0.000216  loss: 2.7741 (2.7943)  grad: 0.2441 (0.2347)  time: 0.4429  data: 0.0052  max mem: 22448
+train: [3]  [260/400]  eta: 0:01:02  lr: 0.000219  loss: 2.7850 (2.7929)  grad: 0.2421 (0.2359)  time: 0.4566  data: 0.0050  max mem: 22448
+train: [3]  [280/400]  eta: 0:00:53  lr: 0.000222  loss: 2.7787 (2.7907)  grad: 0.2434 (0.2373)  time: 0.4283  data: 0.0046  max mem: 22448
+train: [3]  [300/400]  eta: 0:00:44  lr: 0.000225  loss: 2.8008 (2.7943)  grad: 0.2628 (0.2431)  time: 0.4371  data: 0.0047  max mem: 22448
+train: [3]  [320/400]  eta: 0:00:35  lr: 0.000228  loss: 2.9331 (2.8212)  grad: 0.4418 (0.2884)  time: 0.4439  data: 0.0047  max mem: 22448
+WARNING: classifier 46 (36, 1.0) diverged (loss=80.56 > 63.56) at step 763. Freezing.
+train: [3]  [340/400]  eta: 0:00:26  lr: 0.000231  loss: 2.9592 (2.8431)  grad: 0.6130 (0.3189)  time: 0.4349  data: 0.0048  max mem: 22448
+train: [3]  [360/400]  eta: 0:00:17  lr: 0.000234  loss: 2.8824 (2.8458)  grad: 0.2953 (0.3205)  time: 0.4251  data: 0.0050  max mem: 22448
+train: [3]  [380/400]  eta: 0:00:08  lr: 0.000237  loss: 2.9278 (2.8612)  grad: 0.4424 (0.3481)  time: 0.4279  data: 0.0048  max mem: 22448
+WARNING: classifier 45 (31, 1.0) diverged (loss=73.52 > 63.56) at step 797. Freezing.
+train: [3]  [399/400]  eta: 0:00:00  lr: 0.000240  loss: 3.2573 (2.8950)  grad: 0.9422 (0.3878)  time: 0.4258  data: 0.0049  max mem: 22448
+train: [3] Total time: 0:02:57 (0.4433 s / it)
+train: [3] Summary: lr: 0.000240  loss: 3.2573 (2.8950)  grad: 0.9422 (0.3878)
+eval (validation): [3]  [ 0/85]  eta: 0:04:28    time: 3.1614  data: 2.9331  max mem: 22448
+eval (validation): [3]  [20/85]  eta: 0:00:29    time: 0.3220  data: 0.0036  max mem: 22448
+eval (validation): [3]  [40/85]  eta: 0:00:18    time: 0.3617  data: 0.0041  max mem: 22448
+eval (validation): [3]  [60/85]  eta: 0:00:09    time: 0.3455  data: 0.0043  max mem: 22448
+eval (validation): [3]  [80/85]  eta: 0:00:01    time: 0.3174  data: 0.0041  max mem: 22448
+eval (validation): [3]  [84/85]  eta: 0:00:00    time: 0.3113  data: 0.0038  max mem: 22448
+eval (validation): [3] Total time: 0:00:31 (0.3724 s / it)
+cv: [3]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 2.480  acc: 0.263  f1: 0.186
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-best.pth
+train: [4]  [  0/400]  eta: 0:21:48  lr: nan  time: 3.2708  data: 2.8880  max mem: 22448
+train: [4]  [ 20/400]  eta: 0:03:36  lr: 0.000243  loss: 2.6738 (2.7161)  grad: 0.2289 (0.2290)  time: 0.4357  data: 0.0038  max mem: 22448
+train: [4]  [ 40/400]  eta: 0:03:00  lr: 0.000246  loss: 2.6966 (2.7220)  grad: 0.2299 (0.2317)  time: 0.4288  data: 0.0050  max mem: 22448
+train: [4]  [ 60/400]  eta: 0:02:45  lr: 0.000249  loss: 2.7327 (2.7258)  grad: 0.2299 (0.2326)  time: 0.4531  data: 0.0051  max mem: 22448
+train: [4]  [ 80/400]  eta: 0:02:32  lr: 0.000252  loss: 2.7226 (2.7277)  grad: 0.2279 (0.2312)  time: 0.4503  data: 0.0051  max mem: 22448
+train: [4]  [100/400]  eta: 0:02:20  lr: 0.000255  loss: 2.7565 (2.7389)  grad: 0.2281 (0.2324)  time: 0.4290  data: 0.0046  max mem: 22448
+train: [4]  [120/400]  eta: 0:02:10  lr: 0.000258  loss: 2.7519 (2.7342)  grad: 0.2352 (0.2338)  time: 0.4564  data: 0.0049  max mem: 22448
+train: [4]  [140/400]  eta: 0:01:59  lr: 0.000261  loss: 2.6905 (2.7338)  grad: 0.2375 (0.2355)  time: 0.4327  data: 0.0049  max mem: 22448
+train: [4]  [160/400]  eta: 0:01:49  lr: 0.000264  loss: 2.7460 (2.7349)  grad: 0.2439 (0.2377)  time: 0.4295  data: 0.0051  max mem: 22448
+train: [4]  [180/400]  eta: 0:01:40  lr: 0.000267  loss: 2.7399 (2.7393)  grad: 0.2487 (0.2395)  time: 0.4354  data: 0.0049  max mem: 22448
+train: [4]  [200/400]  eta: 0:01:30  lr: 0.000270  loss: 2.7265 (2.7363)  grad: 0.2487 (0.2397)  time: 0.4301  data: 0.0051  max mem: 22448
+train: [4]  [220/400]  eta: 0:01:21  lr: 0.000273  loss: 2.7417 (2.7399)  grad: 0.2491 (0.2405)  time: 0.4306  data: 0.0048  max mem: 22448
+train: [4]  [240/400]  eta: 0:01:12  lr: 0.000276  loss: 2.7470 (2.7389)  grad: 0.2415 (0.2412)  time: 0.4511  data: 0.0052  max mem: 22448
+train: [4]  [260/400]  eta: 0:01:02  lr: 0.000279  loss: 2.7411 (2.7409)  grad: 0.2508 (0.2426)  time: 0.4452  data: 0.0051  max mem: 22448
+train: [4]  [280/400]  eta: 0:00:53  lr: 0.000282  loss: 2.7455 (2.7415)  grad: 0.2522 (0.2442)  time: 0.4323  data: 0.0050  max mem: 22448
+train: [4]  [300/400]  eta: 0:00:44  lr: 0.000285  loss: 2.7471 (2.7428)  grad: 0.2522 (0.2446)  time: 0.4386  data: 0.0050  max mem: 22448
+train: [4]  [320/400]  eta: 0:00:35  lr: 0.000288  loss: 2.7464 (2.7427)  grad: 0.2370 (0.2436)  time: 0.4419  data: 0.0052  max mem: 22448
+train: [4]  [340/400]  eta: 0:00:26  lr: 0.000291  loss: 2.7306 (2.7421)  grad: 0.2317 (0.2431)  time: 0.4466  data: 0.0049  max mem: 22448
+train: [4]  [360/400]  eta: 0:00:17  lr: 0.000294  loss: 2.7106 (2.7424)  grad: 0.2348 (0.2429)  time: 0.4276  data: 0.0049  max mem: 22448
+train: [4]  [380/400]  eta: 0:00:08  lr: 0.000297  loss: 2.7051 (2.7412)  grad: 0.2434 (0.2430)  time: 0.4361  data: 0.0050  max mem: 22448
+train: [4]  [399/400]  eta: 0:00:00  lr: 0.000300  loss: 2.6937 (2.7385)  grad: 0.2438 (0.2429)  time: 0.4303  data: 0.0050  max mem: 22448
+train: [4] Total time: 0:02:58 (0.4457 s / it)
+train: [4] Summary: lr: 0.000300  loss: 2.6937 (2.7385)  grad: 0.2438 (0.2429)
+eval (validation): [4]  [ 0/85]  eta: 0:04:21    time: 3.0709  data: 2.8388  max mem: 22448
+eval (validation): [4]  [20/85]  eta: 0:00:30    time: 0.3448  data: 0.0049  max mem: 22448
+eval (validation): [4]  [40/85]  eta: 0:00:18    time: 0.3370  data: 0.0034  max mem: 22448
+eval (validation): [4]  [60/85]  eta: 0:00:09    time: 0.3193  data: 0.0042  max mem: 22448
+eval (validation): [4]  [80/85]  eta: 0:00:01    time: 0.3129  data: 0.0039  max mem: 22448
+eval (validation): [4]  [84/85]  eta: 0:00:00    time: 0.3052  data: 0.0037  max mem: 22448
+eval (validation): [4] Total time: 0:00:30 (0.3625 s / it)
+cv: [4]  best hparam: (1, 1.0) (024) ('024_lr1.0e+00_wd1.0e+00')  loss: 2.446  acc: 0.262  f1: 0.193
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [5]  [  0/400]  eta: 0:21:59  lr: nan  time: 3.2998  data: 2.9227  max mem: 22448
+train: [5]  [ 20/400]  eta: 0:03:41  lr: 0.000300  loss: 2.6254 (2.6291)  grad: 0.2490 (0.2505)  time: 0.4467  data: 0.0056  max mem: 22448
+train: [5]  [ 40/400]  eta: 0:03:02  lr: 0.000300  loss: 2.6700 (2.6723)  grad: 0.2579 (0.2566)  time: 0.4294  data: 0.0049  max mem: 22448
+train: [5]  [ 60/400]  eta: 0:02:45  lr: 0.000300  loss: 2.6964 (2.6869)  grad: 0.2651 (0.2602)  time: 0.4465  data: 0.0052  max mem: 22448
+train: [5]  [ 80/400]  eta: 0:02:31  lr: 0.000300  loss: 2.6964 (2.6835)  grad: 0.2651 (0.2617)  time: 0.4360  data: 0.0051  max mem: 22448
+train: [5]  [100/400]  eta: 0:02:18  lr: 0.000300  loss: 2.6933 (2.6912)  grad: 0.2749 (0.2695)  time: 0.4152  data: 0.0046  max mem: 22448
+train: [5]  [120/400]  eta: 0:02:09  lr: 0.000300  loss: 2.6933 (2.6927)  grad: 0.3091 (0.2831)  time: 0.4617  data: 0.0052  max mem: 22448
+train: [5]  [140/400]  eta: 0:01:59  lr: 0.000300  loss: 2.7748 (2.7373)  grad: 0.4383 (0.3551)  time: 0.4322  data: 0.0052  max mem: 22448
+WARNING: classifier 44 (26, 1.0) diverged (loss=69.24 > 63.56) at step 1074. Freezing.
+train: [5]  [160/400]  eta: 0:01:49  lr: 0.000299  loss: 2.8681 (2.7832)  grad: 0.7122 (0.4052)  time: 0.4362  data: 0.0050  max mem: 22448
+train: [5]  [180/400]  eta: 0:01:39  lr: 0.000299  loss: 2.6900 (2.7759)  grad: 0.2406 (0.3870)  time: 0.4329  data: 0.0049  max mem: 22448
+train: [5]  [200/400]  eta: 0:01:30  lr: 0.000299  loss: 2.6900 (2.7640)  grad: 0.2406 (0.3730)  time: 0.4249  data: 0.0050  max mem: 22448
+train: [5]  [220/400]  eta: 0:01:20  lr: 0.000299  loss: 2.6513 (2.7555)  grad: 0.2440 (0.3606)  time: 0.4340  data: 0.0051  max mem: 22448
+train: [5]  [240/400]  eta: 0:01:11  lr: 0.000299  loss: 2.6464 (2.7486)  grad: 0.2392 (0.3511)  time: 0.4429  data: 0.0049  max mem: 22448
+train: [5]  [260/400]  eta: 0:01:02  lr: 0.000299  loss: 2.6464 (2.7392)  grad: 0.2392 (0.3424)  time: 0.4382  data: 0.0050  max mem: 22448
+train: [5]  [280/400]  eta: 0:00:53  lr: 0.000298  loss: 2.6534 (2.7359)  grad: 0.2445 (0.3362)  time: 0.4293  data: 0.0048  max mem: 22448
+train: [5]  [300/400]  eta: 0:00:44  lr: 0.000298  loss: 2.6683 (2.7285)  grad: 0.2466 (0.3303)  time: 0.4487  data: 0.0051  max mem: 22448
+train: [5]  [320/400]  eta: 0:00:35  lr: 0.000298  loss: 2.6619 (2.7262)  grad: 0.2499 (0.3258)  time: 0.4403  data: 0.0050  max mem: 22448
+train: [5]  [340/400]  eta: 0:00:26  lr: 0.000298  loss: 2.6766 (2.7218)  grad: 0.2561 (0.3213)  time: 0.4328  data: 0.0050  max mem: 22448
+train: [5]  [360/400]  eta: 0:00:17  lr: 0.000297  loss: 2.6603 (2.7181)  grad: 0.2520 (0.3177)  time: 0.4349  data: 0.0050  max mem: 22448
+train: [5]  [380/400]  eta: 0:00:08  lr: 0.000297  loss: 2.6341 (2.7142)  grad: 0.2516 (0.3143)  time: 0.4354  data: 0.0049  max mem: 22448
+train: [5]  [399/400]  eta: 0:00:00  lr: 0.000297  loss: 2.5976 (2.7085)  grad: 0.2454 (0.3105)  time: 0.4308  data: 0.0049  max mem: 22448
+train: [5] Total time: 0:02:57 (0.4442 s / it)
+train: [5] Summary: lr: 0.000297  loss: 2.5976 (2.7085)  grad: 0.2454 (0.3105)
+eval (validation): [5]  [ 0/85]  eta: 0:04:38    time: 3.2775  data: 2.9780  max mem: 22448
+eval (validation): [5]  [20/85]  eta: 0:00:33    time: 0.3835  data: 0.0041  max mem: 22448
+eval (validation): [5]  [40/85]  eta: 0:00:19    time: 0.3417  data: 0.0038  max mem: 22448
+eval (validation): [5]  [60/85]  eta: 0:00:10    time: 0.3333  data: 0.0039  max mem: 22448
+eval (validation): [5]  [80/85]  eta: 0:00:01    time: 0.3246  data: 0.0042  max mem: 22448
+eval (validation): [5]  [84/85]  eta: 0:00:00    time: 0.3217  data: 0.0041  max mem: 22448
+eval (validation): [5] Total time: 0:00:32 (0.3824 s / it)
+cv: [5]  best hparam: (1, 1.0) (024) ('024_lr1.0e+00_wd1.0e+00')  loss: 2.431  acc: 0.271  f1: 0.196
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-best.pth
+train: [6]  [  0/400]  eta: 0:21:29  lr: nan  time: 3.2241  data: 2.8484  max mem: 22448
+train: [6]  [ 20/400]  eta: 0:03:38  lr: 0.000296  loss: 2.5889 (2.5836)  grad: 0.2379 (0.2379)  time: 0.4413  data: 0.0039  max mem: 22448
+train: [6]  [ 40/400]  eta: 0:03:03  lr: 0.000296  loss: 2.5889 (2.6038)  grad: 0.2432 (0.2455)  time: 0.4448  data: 0.0051  max mem: 22448
+train: [6]  [ 60/400]  eta: 0:02:45  lr: 0.000296  loss: 2.5638 (2.5962)  grad: 0.2454 (0.2459)  time: 0.4365  data: 0.0048  max mem: 22448
+train: [6]  [ 80/400]  eta: 0:02:30  lr: 0.000295  loss: 2.5636 (2.5886)  grad: 0.2449 (0.2468)  time: 0.4215  data: 0.0048  max mem: 22448
+train: [6]  [100/400]  eta: 0:02:19  lr: 0.000295  loss: 2.5873 (2.5885)  grad: 0.2498 (0.2481)  time: 0.4452  data: 0.0050  max mem: 22448
+train: [6]  [120/400]  eta: 0:02:08  lr: 0.000295  loss: 2.5873 (2.5889)  grad: 0.2498 (0.2485)  time: 0.4287  data: 0.0049  max mem: 22448
+train: [6]  [140/400]  eta: 0:01:58  lr: 0.000294  loss: 2.6127 (2.5953)  grad: 0.2479 (0.2489)  time: 0.4230  data: 0.0049  max mem: 22448
+train: [6]  [160/400]  eta: 0:01:48  lr: 0.000294  loss: 2.6388 (2.6013)  grad: 0.2541 (0.2499)  time: 0.4266  data: 0.0048  max mem: 22448
+train: [6]  [180/400]  eta: 0:01:38  lr: 0.000293  loss: 2.6351 (2.6009)  grad: 0.2566 (0.2513)  time: 0.4243  data: 0.0051  max mem: 22448
+train: [6]  [200/400]  eta: 0:01:29  lr: 0.000293  loss: 2.6138 (2.6039)  grad: 0.2566 (0.2517)  time: 0.4313  data: 0.0051  max mem: 22448
+train: [6]  [220/400]  eta: 0:01:20  lr: 0.000292  loss: 2.6376 (2.6016)  grad: 0.2559 (0.2523)  time: 0.4497  data: 0.0050  max mem: 22448
+train: [6]  [240/400]  eta: 0:01:11  lr: 0.000292  loss: 2.6242 (2.6041)  grad: 0.2551 (0.2524)  time: 0.4428  data: 0.0050  max mem: 22448
+train: [6]  [260/400]  eta: 0:01:02  lr: 0.000291  loss: 2.6185 (2.6013)  grad: 0.2496 (0.2519)  time: 0.4148  data: 0.0048  max mem: 22448
+train: [6]  [280/400]  eta: 0:00:53  lr: 0.000291  loss: 2.6112 (2.6021)  grad: 0.2443 (0.2519)  time: 0.4265  data: 0.0047  max mem: 22448
+train: [6]  [300/400]  eta: 0:00:44  lr: 0.000290  loss: 2.6218 (2.6038)  grad: 0.2457 (0.2519)  time: 0.4522  data: 0.0052  max mem: 22448
+train: [6]  [320/400]  eta: 0:00:35  lr: 0.000290  loss: 2.6178 (2.6037)  grad: 0.2506 (0.2524)  time: 0.4319  data: 0.0049  max mem: 22448
+train: [6]  [340/400]  eta: 0:00:26  lr: 0.000289  loss: 2.6077 (2.6051)  grad: 0.2539 (0.2525)  time: 0.4362  data: 0.0050  max mem: 22448
+train: [6]  [360/400]  eta: 0:00:17  lr: 0.000288  loss: 2.5886 (2.6032)  grad: 0.2478 (0.2521)  time: 0.4329  data: 0.0051  max mem: 22448
+train: [6]  [380/400]  eta: 0:00:08  lr: 0.000288  loss: 2.5741 (2.6058)  grad: 0.2492 (0.2523)  time: 0.4321  data: 0.0049  max mem: 22448
+train: [6]  [399/400]  eta: 0:00:00  lr: 0.000287  loss: 2.6337 (2.6059)  grad: 0.2453 (0.2516)  time: 0.4447  data: 0.0048  max mem: 22448
+train: [6] Total time: 0:02:56 (0.4420 s / it)
+train: [6] Summary: lr: 0.000287  loss: 2.6337 (2.6059)  grad: 0.2453 (0.2516)
+eval (validation): [6]  [ 0/85]  eta: 0:04:33    time: 3.2212  data: 2.9349  max mem: 22448
+eval (validation): [6]  [20/85]  eta: 0:00:32    time: 0.3661  data: 0.0045  max mem: 22448
+eval (validation): [6]  [40/85]  eta: 0:00:18    time: 0.3320  data: 0.0043  max mem: 22448
+eval (validation): [6]  [60/85]  eta: 0:00:09    time: 0.3236  data: 0.0040  max mem: 22448
+eval (validation): [6]  [80/85]  eta: 0:00:01    time: 0.3222  data: 0.0040  max mem: 22448
+eval (validation): [6]  [84/85]  eta: 0:00:00    time: 0.3198  data: 0.0041  max mem: 22448
+eval (validation): [6] Total time: 0:00:31 (0.3730 s / it)
+cv: [6]  best hparam: (0.52, 1.0) (020) ('020_lr5.2e-01_wd1.0e+00')  loss: 2.415  acc: 0.274  f1: 0.206
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-best.pth
+train: [7]  [  0/400]  eta: 0:20:50  lr: nan  time: 3.1265  data: 2.8069  max mem: 22448
+train: [7]  [ 20/400]  eta: 0:03:30  lr: 0.000286  loss: 2.5197 (2.5044)  grad: 0.2363 (0.2454)  time: 0.4264  data: 0.0034  max mem: 22448
+train: [7]  [ 40/400]  eta: 0:02:59  lr: 0.000286  loss: 2.5197 (2.5193)  grad: 0.2449 (0.2519)  time: 0.4403  data: 0.0042  max mem: 22448
+train: [7]  [ 60/400]  eta: 0:02:42  lr: 0.000285  loss: 2.5156 (2.5064)  grad: 0.2606 (0.2558)  time: 0.4328  data: 0.0048  max mem: 22448
+train: [7]  [ 80/400]  eta: 0:02:28  lr: 0.000284  loss: 2.5169 (2.5235)  grad: 0.2582 (0.2539)  time: 0.4275  data: 0.0047  max mem: 22448
+train: [7]  [100/400]  eta: 0:02:19  lr: 0.000284  loss: 2.4995 (2.5151)  grad: 0.2484 (0.2538)  time: 0.4569  data: 0.0050  max mem: 22448
+train: [7]  [120/400]  eta: 0:02:08  lr: 0.000283  loss: 2.4809 (2.5163)  grad: 0.2552 (0.2549)  time: 0.4304  data: 0.0049  max mem: 22448
+train: [7]  [140/400]  eta: 0:01:58  lr: 0.000282  loss: 2.5095 (2.5206)  grad: 0.2566 (0.2553)  time: 0.4298  data: 0.0051  max mem: 22448
+train: [7]  [160/400]  eta: 0:01:48  lr: 0.000282  loss: 2.5274 (2.5220)  grad: 0.2562 (0.2554)  time: 0.4231  data: 0.0050  max mem: 22448
+train: [7]  [180/400]  eta: 0:01:38  lr: 0.000281  loss: 2.5705 (2.5281)  grad: 0.2617 (0.2568)  time: 0.4256  data: 0.0048  max mem: 22448
+train: [7]  [200/400]  eta: 0:01:29  lr: 0.000280  loss: 2.5705 (2.5279)  grad: 0.2630 (0.2568)  time: 0.4329  data: 0.0053  max mem: 22448
+train: [7]  [220/400]  eta: 0:01:20  lr: 0.000279  loss: 2.4959 (2.5236)  grad: 0.2531 (0.2567)  time: 0.4389  data: 0.0053  max mem: 22448
+train: [7]  [240/400]  eta: 0:01:11  lr: 0.000278  loss: 2.5249 (2.5280)  grad: 0.2568 (0.2573)  time: 0.4448  data: 0.0049  max mem: 22448
+train: [7]  [260/400]  eta: 0:01:02  lr: 0.000278  loss: 2.5424 (2.5272)  grad: 0.2561 (0.2567)  time: 0.4299  data: 0.0048  max mem: 22448
+train: [7]  [280/400]  eta: 0:00:53  lr: 0.000277  loss: 2.4930 (2.5234)  grad: 0.2510 (0.2566)  time: 0.4317  data: 0.0048  max mem: 22448
+train: [7]  [300/400]  eta: 0:00:44  lr: 0.000276  loss: 2.4584 (2.5222)  grad: 0.2513 (0.2572)  time: 0.4514  data: 0.0047  max mem: 22448
+train: [7]  [320/400]  eta: 0:00:35  lr: 0.000275  loss: 2.5011 (2.5230)  grad: 0.2623 (0.2569)  time: 0.4340  data: 0.0050  max mem: 22448
+train: [7]  [340/400]  eta: 0:00:26  lr: 0.000274  loss: 2.4969 (2.5203)  grad: 0.2459 (0.2565)  time: 0.4326  data: 0.0050  max mem: 22448
+train: [7]  [360/400]  eta: 0:00:17  lr: 0.000273  loss: 2.5058 (2.5226)  grad: 0.2508 (0.2568)  time: 0.4294  data: 0.0050  max mem: 22448
+train: [7]  [380/400]  eta: 0:00:08  lr: 0.000272  loss: 2.5317 (2.5223)  grad: 0.2581 (0.2571)  time: 0.4251  data: 0.0050  max mem: 22448
+train: [7]  [399/400]  eta: 0:00:00  lr: 0.000271  loss: 2.5383 (2.5239)  grad: 0.2611 (0.2577)  time: 0.4550  data: 0.0052  max mem: 22448
+train: [7] Total time: 0:02:56 (0.4423 s / it)
+train: [7] Summary: lr: 0.000271  loss: 2.5383 (2.5239)  grad: 0.2611 (0.2577)
+eval (validation): [7]  [ 0/85]  eta: 0:04:33    time: 3.2165  data: 2.9795  max mem: 22448
+eval (validation): [7]  [20/85]  eta: 0:00:30    time: 0.3390  data: 0.0069  max mem: 22448
+eval (validation): [7]  [40/85]  eta: 0:00:18    time: 0.3306  data: 0.0036  max mem: 22448
+eval (validation): [7]  [60/85]  eta: 0:00:09    time: 0.3222  data: 0.0035  max mem: 22448
+eval (validation): [7]  [80/85]  eta: 0:00:01    time: 0.3303  data: 0.0039  max mem: 22448
+eval (validation): [7]  [84/85]  eta: 0:00:00    time: 0.3288  data: 0.0039  max mem: 22448
+eval (validation): [7] Total time: 0:00:31 (0.3672 s / it)
+cv: [7]  best hparam: (0.61, 1.0) (021) ('021_lr6.1e-01_wd1.0e+00')  loss: 2.425  acc: 0.266  f1: 0.202
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [8]  [  0/400]  eta: 0:21:19  lr: nan  time: 3.1997  data: 2.8214  max mem: 22448
+train: [8]  [ 20/400]  eta: 0:03:42  lr: 0.000270  loss: 2.3457 (2.3958)  grad: 0.2386 (0.2449)  time: 0.4540  data: 0.0048  max mem: 22448
+train: [8]  [ 40/400]  eta: 0:03:06  lr: 0.000270  loss: 2.4368 (2.4309)  grad: 0.2502 (0.2522)  time: 0.4495  data: 0.0053  max mem: 22448
+train: [8]  [ 60/400]  eta: 0:02:46  lr: 0.000269  loss: 2.4619 (2.4417)  grad: 0.2577 (0.2541)  time: 0.4325  data: 0.0048  max mem: 22448
+train: [8]  [ 80/400]  eta: 0:02:31  lr: 0.000268  loss: 2.4893 (2.4547)  grad: 0.2577 (0.2575)  time: 0.4270  data: 0.0048  max mem: 22448
+train: [8]  [100/400]  eta: 0:02:20  lr: 0.000267  loss: 2.4588 (2.4521)  grad: 0.2689 (0.2608)  time: 0.4422  data: 0.0050  max mem: 22448
+train: [8]  [120/400]  eta: 0:02:09  lr: 0.000266  loss: 2.4518 (2.4537)  grad: 0.2731 (0.2639)  time: 0.4361  data: 0.0049  max mem: 22448
+train: [8]  [140/400]  eta: 0:01:58  lr: 0.000265  loss: 2.4720 (2.4607)  grad: 0.2703 (0.2651)  time: 0.4252  data: 0.0048  max mem: 22448
+train: [8]  [160/400]  eta: 0:01:48  lr: 0.000264  loss: 2.4617 (2.4626)  grad: 0.2727 (0.2677)  time: 0.4251  data: 0.0048  max mem: 22448
+train: [8]  [180/400]  eta: 0:01:39  lr: 0.000263  loss: 2.4479 (2.4583)  grad: 0.2688 (0.2672)  time: 0.4313  data: 0.0049  max mem: 22448
+train: [8]  [200/400]  eta: 0:01:29  lr: 0.000262  loss: 2.4536 (2.4613)  grad: 0.2653 (0.2676)  time: 0.4310  data: 0.0049  max mem: 22448
+train: [8]  [220/400]  eta: 0:01:20  lr: 0.000260  loss: 2.4703 (2.4622)  grad: 0.2653 (0.2672)  time: 0.4325  data: 0.0047  max mem: 22448
+train: [8]  [240/400]  eta: 0:01:11  lr: 0.000259  loss: 2.4563 (2.4620)  grad: 0.2603 (0.2670)  time: 0.4401  data: 0.0049  max mem: 22448
+train: [8]  [260/400]  eta: 0:01:02  lr: 0.000258  loss: 2.4668 (2.4644)  grad: 0.2620 (0.2673)  time: 0.4357  data: 0.0051  max mem: 22448
+train: [8]  [280/400]  eta: 0:00:53  lr: 0.000257  loss: 2.4660 (2.4621)  grad: 0.2620 (0.2673)  time: 0.4238  data: 0.0051  max mem: 22448
+train: [8]  [300/400]  eta: 0:00:44  lr: 0.000256  loss: 2.4596 (2.4627)  grad: 0.2637 (0.2672)  time: 0.4455  data: 0.0053  max mem: 22448
+train: [8]  [320/400]  eta: 0:00:35  lr: 0.000255  loss: 2.4768 (2.4634)  grad: 0.2616 (0.2666)  time: 0.4300  data: 0.0052  max mem: 22448
+train: [8]  [340/400]  eta: 0:00:26  lr: 0.000254  loss: 2.4768 (2.4625)  grad: 0.2614 (0.2667)  time: 0.4352  data: 0.0051  max mem: 22448
+train: [8]  [360/400]  eta: 0:00:17  lr: 0.000253  loss: 2.4256 (2.4628)  grad: 0.2614 (0.2662)  time: 0.4345  data: 0.0052  max mem: 22448
+train: [8]  [380/400]  eta: 0:00:08  lr: 0.000252  loss: 2.4360 (2.4635)  grad: 0.2597 (0.2661)  time: 0.4217  data: 0.0052  max mem: 22448
+train: [8]  [399/400]  eta: 0:00:00  lr: 0.000250  loss: 2.4665 (2.4648)  grad: 0.2608 (0.2665)  time: 0.4492  data: 0.0051  max mem: 22448
+train: [8] Total time: 0:02:57 (0.4427 s / it)
+train: [8] Summary: lr: 0.000250  loss: 2.4665 (2.4648)  grad: 0.2608 (0.2665)
+eval (validation): [8]  [ 0/85]  eta: 0:04:32    time: 3.2112  data: 2.9785  max mem: 22448
+eval (validation): [8]  [20/85]  eta: 0:00:30    time: 0.3332  data: 0.0053  max mem: 22448
+eval (validation): [8]  [40/85]  eta: 0:00:18    time: 0.3343  data: 0.0036  max mem: 22448
+eval (validation): [8]  [60/85]  eta: 0:00:09    time: 0.3380  data: 0.0042  max mem: 22448
+eval (validation): [8]  [80/85]  eta: 0:00:01    time: 0.3116  data: 0.0041  max mem: 22448
+eval (validation): [8]  [84/85]  eta: 0:00:00    time: 0.3108  data: 0.0041  max mem: 22448
+eval (validation): [8] Total time: 0:00:31 (0.3655 s / it)
+cv: [8]  best hparam: (0.38, 1.0) (018) ('018_lr3.8e-01_wd1.0e+00')  loss: 2.433  acc: 0.267  f1: 0.202
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [9]  [  0/400]  eta: 0:21:42  lr: nan  time: 3.2555  data: 2.9169  max mem: 22448
+train: [9]  [ 20/400]  eta: 0:03:35  lr: 0.000249  loss: 2.4060 (2.4104)  grad: 0.2582 (0.2693)  time: 0.4338  data: 0.0045  max mem: 22448
+train: [9]  [ 40/400]  eta: 0:03:04  lr: 0.000248  loss: 2.4508 (2.4264)  grad: 0.2591 (0.2639)  time: 0.4554  data: 0.0046  max mem: 22448
+train: [9]  [ 60/400]  eta: 0:02:46  lr: 0.000247  loss: 2.4160 (2.4128)  grad: 0.2557 (0.2603)  time: 0.4380  data: 0.0050  max mem: 22448
+train: [9]  [ 80/400]  eta: 0:02:31  lr: 0.000246  loss: 2.4002 (2.4212)  grad: 0.2557 (0.2617)  time: 0.4321  data: 0.0047  max mem: 22448
+train: [9]  [100/400]  eta: 0:02:19  lr: 0.000244  loss: 2.4036 (2.4231)  grad: 0.2664 (0.2629)  time: 0.4335  data: 0.0046  max mem: 22448
+train: [9]  [120/400]  eta: 0:02:10  lr: 0.000243  loss: 2.3919 (2.4176)  grad: 0.2602 (0.2624)  time: 0.4562  data: 0.0050  max mem: 22448
+train: [9]  [140/400]  eta: 0:02:00  lr: 0.000242  loss: 2.3919 (2.4185)  grad: 0.2607 (0.2633)  time: 0.4469  data: 0.0052  max mem: 22448
+train: [9]  [160/400]  eta: 0:01:50  lr: 0.000241  loss: 2.3866 (2.4126)  grad: 0.2647 (0.2639)  time: 0.4361  data: 0.0049  max mem: 22448
+train: [9]  [180/400]  eta: 0:01:40  lr: 0.000240  loss: 2.3952 (2.4160)  grad: 0.2618 (0.2650)  time: 0.4354  data: 0.0049  max mem: 22448
+train: [9]  [200/400]  eta: 0:01:31  lr: 0.000238  loss: 2.4008 (2.4152)  grad: 0.2703 (0.2660)  time: 0.4426  data: 0.0052  max mem: 22448
+train: [9]  [220/400]  eta: 0:01:21  lr: 0.000237  loss: 2.4079 (2.4120)  grad: 0.2703 (0.2666)  time: 0.4361  data: 0.0051  max mem: 22448
+train: [9]  [240/400]  eta: 0:01:12  lr: 0.000236  loss: 2.4377 (2.4169)  grad: 0.2701 (0.2664)  time: 0.4545  data: 0.0050  max mem: 22448
+train: [9]  [260/400]  eta: 0:01:03  lr: 0.000234  loss: 2.4444 (2.4167)  grad: 0.2646 (0.2661)  time: 0.4455  data: 0.0051  max mem: 22448
+train: [9]  [280/400]  eta: 0:00:54  lr: 0.000233  loss: 2.4205 (2.4174)  grad: 0.2670 (0.2665)  time: 0.4302  data: 0.0047  max mem: 22448
+train: [9]  [300/400]  eta: 0:00:45  lr: 0.000232  loss: 2.4112 (2.4184)  grad: 0.2727 (0.2672)  time: 0.4458  data: 0.0050  max mem: 22448
+train: [9]  [320/400]  eta: 0:00:35  lr: 0.000230  loss: 2.4112 (2.4199)  grad: 0.2715 (0.2671)  time: 0.4269  data: 0.0047  max mem: 22448
+train: [9]  [340/400]  eta: 0:00:26  lr: 0.000229  loss: 2.3910 (2.4181)  grad: 0.2671 (0.2677)  time: 0.4474  data: 0.0050  max mem: 22448
+train: [9]  [360/400]  eta: 0:00:17  lr: 0.000228  loss: 2.4466 (2.4202)  grad: 0.2760 (0.2684)  time: 0.4438  data: 0.0050  max mem: 22448
+train: [9]  [380/400]  eta: 0:00:08  lr: 0.000226  loss: 2.4080 (2.4191)  grad: 0.2679 (0.2685)  time: 0.4236  data: 0.0048  max mem: 22448
+train: [9]  [399/400]  eta: 0:00:00  lr: 0.000225  loss: 2.4172 (2.4198)  grad: 0.2691 (0.2690)  time: 0.4431  data: 0.0051  max mem: 22448
+train: [9] Total time: 0:02:59 (0.4480 s / it)
+train: [9] Summary: lr: 0.000225  loss: 2.4172 (2.4198)  grad: 0.2691 (0.2690)
+eval (validation): [9]  [ 0/85]  eta: 0:04:49    time: 3.4016  data: 3.0947  max mem: 22448
+eval (validation): [9]  [20/85]  eta: 0:00:33    time: 0.3692  data: 0.0052  max mem: 22448
+eval (validation): [9]  [40/85]  eta: 0:00:19    time: 0.3429  data: 0.0044  max mem: 22448
+eval (validation): [9]  [60/85]  eta: 0:00:09    time: 0.3329  data: 0.0041  max mem: 22448
+eval (validation): [9]  [80/85]  eta: 0:00:01    time: 0.3258  data: 0.0042  max mem: 22448
+eval (validation): [9]  [84/85]  eta: 0:00:00    time: 0.3257  data: 0.0040  max mem: 22448
+eval (validation): [9] Total time: 0:00:32 (0.3818 s / it)
+cv: [9]  best hparam: (0.38, 1.0) (018) ('018_lr3.8e-01_wd1.0e+00')  loss: 2.418  acc: 0.267  f1: 0.193
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [10]  [  0/400]  eta: 0:21:53  lr: nan  time: 3.2827  data: 2.8879  max mem: 22448
+train: [10]  [ 20/400]  eta: 0:03:46  lr: 0.000224  loss: 2.3366 (2.3420)  grad: 0.2728 (0.2732)  time: 0.4630  data: 0.0065  max mem: 22448
+train: [10]  [ 40/400]  eta: 0:03:07  lr: 0.000222  loss: 2.3284 (2.3344)  grad: 0.2681 (0.2691)  time: 0.4386  data: 0.0047  max mem: 22448
+train: [10]  [ 60/400]  eta: 0:02:47  lr: 0.000221  loss: 2.3284 (2.3566)  grad: 0.2588 (0.2669)  time: 0.4390  data: 0.0051  max mem: 22448
+train: [10]  [ 80/400]  eta: 0:02:33  lr: 0.000220  loss: 2.3518 (2.3490)  grad: 0.2587 (0.2656)  time: 0.4364  data: 0.0051  max mem: 22448
+train: [10]  [100/400]  eta: 0:02:20  lr: 0.000218  loss: 2.3510 (2.3492)  grad: 0.2685 (0.2672)  time: 0.4260  data: 0.0050  max mem: 22448
+train: [10]  [120/400]  eta: 0:02:09  lr: 0.000217  loss: 2.3696 (2.3509)  grad: 0.2711 (0.2676)  time: 0.4414  data: 0.0048  max mem: 22448
+train: [10]  [140/400]  eta: 0:01:59  lr: 0.000215  loss: 2.3597 (2.3555)  grad: 0.2673 (0.2675)  time: 0.4374  data: 0.0051  max mem: 22448
+train: [10]  [160/400]  eta: 0:01:49  lr: 0.000214  loss: 2.3597 (2.3561)  grad: 0.2613 (0.2674)  time: 0.4356  data: 0.0049  max mem: 22448
+train: [10]  [180/400]  eta: 0:01:40  lr: 0.000213  loss: 2.3417 (2.3597)  grad: 0.2630 (0.2675)  time: 0.4431  data: 0.0051  max mem: 22448
+train: [10]  [200/400]  eta: 0:01:30  lr: 0.000211  loss: 2.3908 (2.3600)  grad: 0.2682 (0.2676)  time: 0.4283  data: 0.0048  max mem: 22448
+train: [10]  [220/400]  eta: 0:01:21  lr: 0.000210  loss: 2.3760 (2.3596)  grad: 0.2608 (0.2674)  time: 0.4451  data: 0.0048  max mem: 22448
+train: [10]  [240/400]  eta: 0:01:12  lr: 0.000208  loss: 2.3594 (2.3609)  grad: 0.2639 (0.2674)  time: 0.4571  data: 0.0051  max mem: 22448
+train: [10]  [260/400]  eta: 0:01:03  lr: 0.000207  loss: 2.4081 (2.3626)  grad: 0.2639 (0.2671)  time: 0.4379  data: 0.0049  max mem: 22448
+train: [10]  [280/400]  eta: 0:00:54  lr: 0.000205  loss: 2.4081 (2.3628)  grad: 0.2646 (0.2672)  time: 0.4313  data: 0.0049  max mem: 22448
+train: [10]  [300/400]  eta: 0:00:45  lr: 0.000204  loss: 2.3193 (2.3596)  grad: 0.2594 (0.2665)  time: 0.4533  data: 0.0049  max mem: 22448
+train: [10]  [320/400]  eta: 0:00:35  lr: 0.000202  loss: 2.3182 (2.3572)  grad: 0.2602 (0.2667)  time: 0.4352  data: 0.0051  max mem: 22448
+train: [10]  [340/400]  eta: 0:00:26  lr: 0.000201  loss: 2.3419 (2.3574)  grad: 0.2724 (0.2669)  time: 0.4173  data: 0.0044  max mem: 22448
+train: [10]  [360/400]  eta: 0:00:17  lr: 0.000199  loss: 2.3205 (2.3553)  grad: 0.2626 (0.2665)  time: 0.4329  data: 0.0048  max mem: 22448
+train: [10]  [380/400]  eta: 0:00:08  lr: 0.000198  loss: 2.3032 (2.3544)  grad: 0.2618 (0.2665)  time: 0.4276  data: 0.0047  max mem: 22448
+train: [10]  [399/400]  eta: 0:00:00  lr: 0.000196  loss: 2.3335 (2.3557)  grad: 0.2691 (0.2669)  time: 0.4262  data: 0.0048  max mem: 22448
+train: [10] Total time: 0:02:58 (0.4453 s / it)
+train: [10] Summary: lr: 0.000196  loss: 2.3335 (2.3557)  grad: 0.2691 (0.2669)
+eval (validation): [10]  [ 0/85]  eta: 0:04:19    time: 3.0491  data: 2.8103  max mem: 22448
+eval (validation): [10]  [20/85]  eta: 0:00:30    time: 0.3477  data: 0.0041  max mem: 22448
+eval (validation): [10]  [40/85]  eta: 0:00:18    time: 0.3255  data: 0.0037  max mem: 22448
+eval (validation): [10]  [60/85]  eta: 0:00:09    time: 0.3302  data: 0.0038  max mem: 22448
+eval (validation): [10]  [80/85]  eta: 0:00:01    time: 0.3143  data: 0.0038  max mem: 22448
+eval (validation): [10]  [84/85]  eta: 0:00:00    time: 0.3091  data: 0.0038  max mem: 22448
+eval (validation): [10] Total time: 0:00:30 (0.3638 s / it)
+cv: [10]  best hparam: (0.32, 1.0) (017) ('017_lr3.2e-01_wd1.0e+00')  loss: 2.421  acc: 0.267  f1: 0.195
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [11]  [  0/400]  eta: 0:21:55  lr: nan  time: 3.2900  data: 2.9735  max mem: 22448
+train: [11]  [ 20/400]  eta: 0:03:33  lr: 0.000195  loss: 2.2532 (2.3028)  grad: 0.2523 (0.2616)  time: 0.4249  data: 0.0033  max mem: 22448
+train: [11]  [ 40/400]  eta: 0:02:57  lr: 0.000193  loss: 2.3034 (2.3110)  grad: 0.2639 (0.2649)  time: 0.4231  data: 0.0048  max mem: 22448
+train: [11]  [ 60/400]  eta: 0:02:41  lr: 0.000192  loss: 2.2905 (2.2917)  grad: 0.2681 (0.2669)  time: 0.4322  data: 0.0048  max mem: 22448
+train: [11]  [ 80/400]  eta: 0:02:28  lr: 0.000190  loss: 2.3033 (2.3042)  grad: 0.2710 (0.2694)  time: 0.4287  data: 0.0047  max mem: 22448
+train: [11]  [100/400]  eta: 0:02:16  lr: 0.000189  loss: 2.3186 (2.2980)  grad: 0.2710 (0.2685)  time: 0.4218  data: 0.0049  max mem: 22448
+train: [11]  [120/400]  eta: 0:02:07  lr: 0.000187  loss: 2.2011 (2.2846)  grad: 0.2660 (0.2686)  time: 0.4520  data: 0.0050  max mem: 22448
+train: [11]  [140/400]  eta: 0:01:57  lr: 0.000186  loss: 2.2591 (2.2898)  grad: 0.2660 (0.2693)  time: 0.4410  data: 0.0051  max mem: 22448
+train: [11]  [160/400]  eta: 0:01:48  lr: 0.000184  loss: 2.3172 (2.2957)  grad: 0.2744 (0.2698)  time: 0.4387  data: 0.0050  max mem: 22448
+train: [11]  [180/400]  eta: 0:01:38  lr: 0.000183  loss: 2.2912 (2.2955)  grad: 0.2821 (0.2719)  time: 0.4348  data: 0.0052  max mem: 22448
+train: [11]  [200/400]  eta: 0:01:29  lr: 0.000181  loss: 2.2962 (2.2995)  grad: 0.2856 (0.2726)  time: 0.4376  data: 0.0046  max mem: 22448
+train: [11]  [220/400]  eta: 0:01:20  lr: 0.000180  loss: 2.3475 (2.3054)  grad: 0.2765 (0.2723)  time: 0.4291  data: 0.0053  max mem: 22448
+train: [11]  [240/400]  eta: 0:01:11  lr: 0.000178  loss: 2.3508 (2.3078)  grad: 0.2614 (0.2721)  time: 0.4466  data: 0.0051  max mem: 22448
+train: [11]  [260/400]  eta: 0:01:02  lr: 0.000177  loss: 2.3220 (2.3087)  grad: 0.2658 (0.2727)  time: 0.4479  data: 0.0051  max mem: 22448
+train: [11]  [280/400]  eta: 0:00:53  lr: 0.000175  loss: 2.3135 (2.3102)  grad: 0.2709 (0.2729)  time: 0.4295  data: 0.0049  max mem: 22448
+train: [11]  [300/400]  eta: 0:00:44  lr: 0.000174  loss: 2.3461 (2.3137)  grad: 0.2730 (0.2731)  time: 0.4285  data: 0.0048  max mem: 22448
+train: [11]  [320/400]  eta: 0:00:35  lr: 0.000172  loss: 2.3461 (2.3148)  grad: 0.2753 (0.2738)  time: 0.4400  data: 0.0047  max mem: 22448
+train: [11]  [340/400]  eta: 0:00:26  lr: 0.000170  loss: 2.2901 (2.3154)  grad: 0.2753 (0.2744)  time: 0.4400  data: 0.0049  max mem: 22448
+train: [11]  [360/400]  eta: 0:00:17  lr: 0.000169  loss: 2.2953 (2.3142)  grad: 0.2684 (0.2748)  time: 0.4384  data: 0.0051  max mem: 22448
+train: [11]  [380/400]  eta: 0:00:08  lr: 0.000167  loss: 2.2778 (2.3110)  grad: 0.2684 (0.2745)  time: 0.4408  data: 0.0050  max mem: 22448
+train: [11]  [399/400]  eta: 0:00:00  lr: 0.000166  loss: 2.2979 (2.3136)  grad: 0.2628 (0.2739)  time: 0.4320  data: 0.0051  max mem: 22448
+train: [11] Total time: 0:02:57 (0.4431 s / it)
+train: [11] Summary: lr: 0.000166  loss: 2.2979 (2.3136)  grad: 0.2628 (0.2739)
+eval (validation): [11]  [ 0/85]  eta: 0:04:40    time: 3.2953  data: 3.0454  max mem: 22448
+eval (validation): [11]  [20/85]  eta: 0:00:32    time: 0.3594  data: 0.0062  max mem: 22448
+eval (validation): [11]  [40/85]  eta: 0:00:19    time: 0.3502  data: 0.0037  max mem: 22448
+eval (validation): [11]  [60/85]  eta: 0:00:09    time: 0.3166  data: 0.0039  max mem: 22448
+eval (validation): [11]  [80/85]  eta: 0:00:01    time: 0.3145  data: 0.0039  max mem: 22448
+eval (validation): [11]  [84/85]  eta: 0:00:00    time: 0.3140  data: 0.0039  max mem: 22448
+eval (validation): [11] Total time: 0:00:31 (0.3722 s / it)
+cv: [11]  best hparam: (0.32, 1.0) (017) ('017_lr3.2e-01_wd1.0e+00')  loss: 2.428  acc: 0.265  f1: 0.193
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [12]  [  0/400]  eta: 0:23:06  lr: nan  time: 3.4671  data: 3.0893  max mem: 22448
+train: [12]  [ 20/400]  eta: 0:03:44  lr: 0.000164  loss: 2.2093 (2.2211)  grad: 0.2607 (0.2612)  time: 0.4458  data: 0.0041  max mem: 22448
+train: [12]  [ 40/400]  eta: 0:03:05  lr: 0.000163  loss: 2.2416 (2.2402)  grad: 0.2625 (0.2662)  time: 0.4389  data: 0.0051  max mem: 22448
+train: [12]  [ 60/400]  eta: 0:02:47  lr: 0.000161  loss: 2.2452 (2.2466)  grad: 0.2652 (0.2651)  time: 0.4415  data: 0.0051  max mem: 22448
+train: [12]  [ 80/400]  eta: 0:02:32  lr: 0.000160  loss: 2.2551 (2.2514)  grad: 0.2619 (0.2641)  time: 0.4328  data: 0.0050  max mem: 22448
+train: [12]  [100/400]  eta: 0:02:21  lr: 0.000158  loss: 2.2579 (2.2509)  grad: 0.2659 (0.2648)  time: 0.4438  data: 0.0050  max mem: 22448
+train: [12]  [120/400]  eta: 0:02:09  lr: 0.000156  loss: 2.2561 (2.2548)  grad: 0.2651 (0.2643)  time: 0.4262  data: 0.0050  max mem: 22448
+train: [12]  [140/400]  eta: 0:02:00  lr: 0.000155  loss: 2.2513 (2.2519)  grad: 0.2692 (0.2665)  time: 0.4526  data: 0.0053  max mem: 22448
+train: [12]  [160/400]  eta: 0:01:50  lr: 0.000153  loss: 2.2253 (2.2512)  grad: 0.2774 (0.2680)  time: 0.4407  data: 0.0052  max mem: 22448
+train: [12]  [180/400]  eta: 0:01:40  lr: 0.000152  loss: 2.2468 (2.2521)  grad: 0.2774 (0.2688)  time: 0.4360  data: 0.0051  max mem: 22448
+train: [12]  [200/400]  eta: 0:01:30  lr: 0.000150  loss: 2.2939 (2.2586)  grad: 0.2712 (0.2693)  time: 0.4401  data: 0.0051  max mem: 22448
+train: [12]  [220/400]  eta: 0:01:21  lr: 0.000149  loss: 2.3136 (2.2639)  grad: 0.2667 (0.2689)  time: 0.4484  data: 0.0053  max mem: 22448
+train: [12]  [240/400]  eta: 0:01:12  lr: 0.000147  loss: 2.2598 (2.2630)  grad: 0.2740 (0.2702)  time: 0.4418  data: 0.0051  max mem: 22448
+train: [12]  [260/400]  eta: 0:01:03  lr: 0.000145  loss: 2.2438 (2.2621)  grad: 0.2753 (0.2703)  time: 0.4550  data: 0.0051  max mem: 22448
+train: [12]  [280/400]  eta: 0:00:54  lr: 0.000144  loss: 2.2416 (2.2574)  grad: 0.2654 (0.2702)  time: 0.4404  data: 0.0051  max mem: 22448
+train: [12]  [300/400]  eta: 0:00:45  lr: 0.000142  loss: 2.2533 (2.2610)  grad: 0.2693 (0.2711)  time: 0.4399  data: 0.0050  max mem: 22448
+train: [12]  [320/400]  eta: 0:00:36  lr: 0.000141  loss: 2.2617 (2.2620)  grad: 0.2694 (0.2709)  time: 0.4506  data: 0.0053  max mem: 22448
+train: [12]  [340/400]  eta: 0:00:27  lr: 0.000139  loss: 2.2617 (2.2617)  grad: 0.2673 (0.2707)  time: 0.4476  data: 0.0050  max mem: 22448
+train: [12]  [360/400]  eta: 0:00:18  lr: 0.000138  loss: 2.2277 (2.2602)  grad: 0.2635 (0.2704)  time: 0.4448  data: 0.0051  max mem: 22448
+train: [12]  [380/400]  eta: 0:00:09  lr: 0.000136  loss: 2.2327 (2.2601)  grad: 0.2639 (0.2702)  time: 0.4412  data: 0.0050  max mem: 22448
+train: [12]  [399/400]  eta: 0:00:00  lr: 0.000134  loss: 2.2335 (2.2589)  grad: 0.2691 (0.2706)  time: 0.4359  data: 0.0052  max mem: 22448
+train: [12] Total time: 0:03:00 (0.4504 s / it)
+train: [12] Summary: lr: 0.000134  loss: 2.2335 (2.2589)  grad: 0.2691 (0.2706)
+eval (validation): [12]  [ 0/85]  eta: 0:04:39    time: 3.2833  data: 3.0432  max mem: 22448
+eval (validation): [12]  [20/85]  eta: 0:00:32    time: 0.3592  data: 0.0039  max mem: 22448
+eval (validation): [12]  [40/85]  eta: 0:00:19    time: 0.3439  data: 0.0041  max mem: 22448
+eval (validation): [12]  [60/85]  eta: 0:00:09    time: 0.3498  data: 0.0046  max mem: 22448
+eval (validation): [12]  [80/85]  eta: 0:00:01    time: 0.3285  data: 0.0038  max mem: 22448
+eval (validation): [12]  [84/85]  eta: 0:00:00    time: 0.3252  data: 0.0036  max mem: 22448
+eval (validation): [12] Total time: 0:00:32 (0.3816 s / it)
+cv: [12]  best hparam: (0.32, 1.0) (017) ('017_lr3.2e-01_wd1.0e+00')  loss: 2.443  acc: 0.264  f1: 0.195
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [13]  [  0/400]  eta: 0:22:23  lr: nan  time: 3.3597  data: 2.9843  max mem: 22448
+train: [13]  [ 20/400]  eta: 0:03:39  lr: 0.000133  loss: 2.2115 (2.2137)  grad: 0.2693 (0.2699)  time: 0.4381  data: 0.0045  max mem: 22448
+train: [13]  [ 40/400]  eta: 0:03:03  lr: 0.000131  loss: 2.2199 (2.2106)  grad: 0.2693 (0.2707)  time: 0.4416  data: 0.0049  max mem: 22448
+train: [13]  [ 60/400]  eta: 0:02:44  lr: 0.000130  loss: 2.2094 (2.2071)  grad: 0.2710 (0.2711)  time: 0.4265  data: 0.0048  max mem: 22448
+train: [13]  [ 80/400]  eta: 0:02:32  lr: 0.000128  loss: 2.1794 (2.2013)  grad: 0.2706 (0.2713)  time: 0.4500  data: 0.0051  max mem: 22448
+train: [13]  [100/400]  eta: 0:02:20  lr: 0.000127  loss: 2.1653 (2.1991)  grad: 0.2594 (0.2699)  time: 0.4361  data: 0.0052  max mem: 22448
+train: [13]  [120/400]  eta: 0:02:09  lr: 0.000125  loss: 2.1899 (2.1953)  grad: 0.2594 (0.2701)  time: 0.4383  data: 0.0050  max mem: 22448
+train: [13]  [140/400]  eta: 0:01:59  lr: 0.000124  loss: 2.1701 (2.2048)  grad: 0.2773 (0.2720)  time: 0.4456  data: 0.0050  max mem: 22448
+train: [13]  [160/400]  eta: 0:01:50  lr: 0.000122  loss: 2.1956 (2.2042)  grad: 0.2837 (0.2736)  time: 0.4463  data: 0.0053  max mem: 22448
+train: [13]  [180/400]  eta: 0:01:40  lr: 0.000120  loss: 2.2223 (2.2130)  grad: 0.2837 (0.2743)  time: 0.4456  data: 0.0052  max mem: 22448
+train: [13]  [200/400]  eta: 0:01:31  lr: 0.000119  loss: 2.2049 (2.2084)  grad: 0.2764 (0.2743)  time: 0.4434  data: 0.0050  max mem: 22448
+train: [13]  [220/400]  eta: 0:01:22  lr: 0.000117  loss: 2.2077 (2.2100)  grad: 0.2741 (0.2750)  time: 0.4589  data: 0.0050  max mem: 22448
+train: [13]  [240/400]  eta: 0:01:13  lr: 0.000116  loss: 2.2248 (2.2092)  grad: 0.2741 (0.2750)  time: 0.4604  data: 0.0052  max mem: 22448
+train: [13]  [260/400]  eta: 0:01:03  lr: 0.000114  loss: 2.1933 (2.2108)  grad: 0.2699 (0.2744)  time: 0.4590  data: 0.0053  max mem: 22448
+train: [13]  [280/400]  eta: 0:00:54  lr: 0.000113  loss: 2.1670 (2.2074)  grad: 0.2645 (0.2733)  time: 0.4401  data: 0.0050  max mem: 22448
+train: [13]  [300/400]  eta: 0:00:45  lr: 0.000111  loss: 2.2017 (2.2076)  grad: 0.2610 (0.2721)  time: 0.4468  data: 0.0051  max mem: 22448
+train: [13]  [320/400]  eta: 0:00:36  lr: 0.000110  loss: 2.2212 (2.2092)  grad: 0.2636 (0.2723)  time: 0.4503  data: 0.0053  max mem: 22448
+train: [13]  [340/400]  eta: 0:00:27  lr: 0.000108  loss: 2.2123 (2.2102)  grad: 0.2696 (0.2720)  time: 0.4409  data: 0.0052  max mem: 22448
+train: [13]  [360/400]  eta: 0:00:18  lr: 0.000107  loss: 2.2060 (2.2105)  grad: 0.2734 (0.2724)  time: 0.4363  data: 0.0052  max mem: 22448
+train: [13]  [380/400]  eta: 0:00:09  lr: 0.000105  loss: 2.2226 (2.2107)  grad: 0.2764 (0.2727)  time: 0.4467  data: 0.0053  max mem: 22448
+train: [13]  [399/400]  eta: 0:00:00  lr: 0.000104  loss: 2.2250 (2.2121)  grad: 0.2728 (0.2728)  time: 0.4472  data: 0.0051  max mem: 22448
+train: [13] Total time: 0:03:01 (0.4528 s / it)
+train: [13] Summary: lr: 0.000104  loss: 2.2250 (2.2121)  grad: 0.2728 (0.2728)
+eval (validation): [13]  [ 0/85]  eta: 0:04:27    time: 3.1466  data: 2.8638  max mem: 22448
+eval (validation): [13]  [20/85]  eta: 0:00:34    time: 0.4065  data: 0.0340  max mem: 22448
+eval (validation): [13]  [40/85]  eta: 0:00:20    time: 0.3511  data: 0.0042  max mem: 22448
+eval (validation): [13]  [60/85]  eta: 0:00:10    time: 0.3506  data: 0.0046  max mem: 22448
+eval (validation): [13]  [80/85]  eta: 0:00:01    time: 0.3233  data: 0.0042  max mem: 22448
+eval (validation): [13]  [84/85]  eta: 0:00:00    time: 0.3167  data: 0.0041  max mem: 22448
+eval (validation): [13] Total time: 0:00:33 (0.3926 s / it)
+cv: [13]  best hparam: (0.27, 1.0) (016) ('016_lr2.7e-01_wd1.0e+00')  loss: 2.429  acc: 0.270  f1: 0.198
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [14]  [  0/400]  eta: 0:22:12  lr: nan  time: 3.3307  data: 2.9967  max mem: 22448
+train: [14]  [ 20/400]  eta: 0:03:43  lr: 0.000102  loss: 2.1245 (2.1219)  grad: 0.2563 (0.2580)  time: 0.4511  data: 0.0042  max mem: 22448
+train: [14]  [ 40/400]  eta: 0:03:08  lr: 0.000101  loss: 2.1318 (2.1277)  grad: 0.2558 (0.2589)  time: 0.4549  data: 0.0046  max mem: 22448
+train: [14]  [ 60/400]  eta: 0:02:47  lr: 0.000099  loss: 2.1673 (2.1404)  grad: 0.2664 (0.2622)  time: 0.4285  data: 0.0050  max mem: 22448
+train: [14]  [ 80/400]  eta: 0:02:33  lr: 0.000098  loss: 2.1652 (2.1504)  grad: 0.2619 (0.2614)  time: 0.4459  data: 0.0051  max mem: 22448
+train: [14]  [100/400]  eta: 0:02:21  lr: 0.000096  loss: 2.1565 (2.1572)  grad: 0.2572 (0.2625)  time: 0.4274  data: 0.0049  max mem: 22448
+train: [14]  [120/400]  eta: 0:02:10  lr: 0.000095  loss: 2.1632 (2.1533)  grad: 0.2680 (0.2640)  time: 0.4466  data: 0.0049  max mem: 22448
+train: [14]  [140/400]  eta: 0:02:00  lr: 0.000093  loss: 2.1588 (2.1541)  grad: 0.2702 (0.2654)  time: 0.4393  data: 0.0051  max mem: 22448
+train: [14]  [160/400]  eta: 0:01:50  lr: 0.000092  loss: 2.1128 (2.1490)  grad: 0.2722 (0.2659)  time: 0.4485  data: 0.0050  max mem: 22448
+train: [14]  [180/400]  eta: 0:01:40  lr: 0.000090  loss: 2.1100 (2.1453)  grad: 0.2685 (0.2663)  time: 0.4427  data: 0.0051  max mem: 22448
+train: [14]  [200/400]  eta: 0:01:31  lr: 0.000089  loss: 2.1343 (2.1479)  grad: 0.2734 (0.2675)  time: 0.4431  data: 0.0053  max mem: 22448
+train: [14]  [220/400]  eta: 0:01:22  lr: 0.000088  loss: 2.1699 (2.1504)  grad: 0.2822 (0.2681)  time: 0.4416  data: 0.0053  max mem: 22448
+train: [14]  [240/400]  eta: 0:01:12  lr: 0.000086  loss: 2.1746 (2.1549)  grad: 0.2707 (0.2682)  time: 0.4505  data: 0.0051  max mem: 22448
+train: [14]  [260/400]  eta: 0:01:03  lr: 0.000085  loss: 2.2083 (2.1575)  grad: 0.2711 (0.2686)  time: 0.4528  data: 0.0052  max mem: 22448
+train: [14]  [280/400]  eta: 0:00:54  lr: 0.000083  loss: 2.2083 (2.1596)  grad: 0.2705 (0.2682)  time: 0.4330  data: 0.0050  max mem: 22448
+train: [14]  [300/400]  eta: 0:00:45  lr: 0.000082  loss: 2.2189 (2.1648)  grad: 0.2651 (0.2682)  time: 0.4677  data: 0.0053  max mem: 22448
+train: [14]  [320/400]  eta: 0:00:36  lr: 0.000081  loss: 2.2080 (2.1659)  grad: 0.2711 (0.2687)  time: 0.4409  data: 0.0052  max mem: 22448
+train: [14]  [340/400]  eta: 0:00:27  lr: 0.000079  loss: 2.1886 (2.1662)  grad: 0.2738 (0.2690)  time: 0.4360  data: 0.0052  max mem: 22448
+train: [14]  [360/400]  eta: 0:00:18  lr: 0.000078  loss: 2.1486 (2.1661)  grad: 0.2666 (0.2689)  time: 0.4332  data: 0.0051  max mem: 22448
+train: [14]  [380/400]  eta: 0:00:09  lr: 0.000076  loss: 2.1375 (2.1639)  grad: 0.2672 (0.2689)  time: 0.4427  data: 0.0049  max mem: 22448
+train: [14]  [399/400]  eta: 0:00:00  lr: 0.000075  loss: 2.1298 (2.1644)  grad: 0.2761 (0.2695)  time: 0.4366  data: 0.0051  max mem: 22448
+train: [14] Total time: 0:03:00 (0.4510 s / it)
+train: [14] Summary: lr: 0.000075  loss: 2.1298 (2.1644)  grad: 0.2761 (0.2695)
+eval (validation): [14]  [ 0/85]  eta: 0:04:27    time: 3.1498  data: 2.9163  max mem: 22448
+eval (validation): [14]  [20/85]  eta: 0:00:31    time: 0.3448  data: 0.0041  max mem: 22448
+eval (validation): [14]  [40/85]  eta: 0:00:18    time: 0.3526  data: 0.0042  max mem: 22448
+eval (validation): [14]  [60/85]  eta: 0:00:09    time: 0.3539  data: 0.0043  max mem: 22448
+eval (validation): [14]  [80/85]  eta: 0:00:01    time: 0.3191  data: 0.0041  max mem: 22448
+eval (validation): [14]  [84/85]  eta: 0:00:00    time: 0.3138  data: 0.0042  max mem: 22448
+eval (validation): [14] Total time: 0:00:32 (0.3772 s / it)
+cv: [14]  best hparam: (0.27, 1.0) (016) ('016_lr2.7e-01_wd1.0e+00')  loss: 2.424  acc: 0.272  f1: 0.201
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [15]  [  0/400]  eta: 0:22:30  lr: nan  time: 3.3752  data: 2.9929  max mem: 22448
+train: [15]  [ 20/400]  eta: 0:03:41  lr: 0.000074  loss: 2.1213 (2.1263)  grad: 0.2582 (0.2638)  time: 0.4445  data: 0.0047  max mem: 22448
+train: [15]  [ 40/400]  eta: 0:03:04  lr: 0.000072  loss: 2.1213 (2.1260)  grad: 0.2582 (0.2620)  time: 0.4384  data: 0.0048  max mem: 22448
+train: [15]  [ 60/400]  eta: 0:02:46  lr: 0.000071  loss: 2.0979 (2.1146)  grad: 0.2632 (0.2634)  time: 0.4409  data: 0.0049  max mem: 22448
+train: [15]  [ 80/400]  eta: 0:02:32  lr: 0.000070  loss: 2.0834 (2.1039)  grad: 0.2632 (0.2627)  time: 0.4416  data: 0.0052  max mem: 22448
+train: [15]  [100/400]  eta: 0:02:20  lr: 0.000068  loss: 2.0599 (2.0983)  grad: 0.2599 (0.2635)  time: 0.4292  data: 0.0050  max mem: 22448
+train: [15]  [120/400]  eta: 0:02:10  lr: 0.000067  loss: 2.1045 (2.1064)  grad: 0.2736 (0.2654)  time: 0.4588  data: 0.0051  max mem: 22448
+train: [15]  [140/400]  eta: 0:02:00  lr: 0.000066  loss: 2.1342 (2.1142)  grad: 0.2744 (0.2679)  time: 0.4361  data: 0.0049  max mem: 22448
+train: [15]  [160/400]  eta: 0:01:50  lr: 0.000064  loss: 2.1161 (2.1140)  grad: 0.2721 (0.2675)  time: 0.4387  data: 0.0043  max mem: 22448
+train: [15]  [180/400]  eta: 0:01:40  lr: 0.000063  loss: 2.1684 (2.1252)  grad: 0.2727 (0.2687)  time: 0.4481  data: 0.0053  max mem: 22448
+train: [15]  [200/400]  eta: 0:01:31  lr: 0.000062  loss: 2.1684 (2.1259)  grad: 0.2658 (0.2682)  time: 0.4487  data: 0.0051  max mem: 22448
+train: [15]  [220/400]  eta: 0:01:21  lr: 0.000061  loss: 2.1272 (2.1242)  grad: 0.2625 (0.2683)  time: 0.4333  data: 0.0051  max mem: 22448
+train: [15]  [240/400]  eta: 0:01:12  lr: 0.000059  loss: 2.1272 (2.1259)  grad: 0.2741 (0.2687)  time: 0.4557  data: 0.0054  max mem: 22448
+train: [15]  [260/400]  eta: 0:01:03  lr: 0.000058  loss: 2.1445 (2.1301)  grad: 0.2701 (0.2687)  time: 0.4627  data: 0.0052  max mem: 22448
+train: [15]  [280/400]  eta: 0:00:54  lr: 0.000057  loss: 2.1445 (2.1280)  grad: 0.2583 (0.2679)  time: 0.4279  data: 0.0047  max mem: 22448
+train: [15]  [300/400]  eta: 0:00:45  lr: 0.000056  loss: 2.1152 (2.1280)  grad: 0.2582 (0.2678)  time: 0.4589  data: 0.0052  max mem: 22448
+train: [15]  [320/400]  eta: 0:00:36  lr: 0.000054  loss: 2.1242 (2.1284)  grad: 0.2696 (0.2677)  time: 0.4535  data: 0.0051  max mem: 22448
+train: [15]  [340/400]  eta: 0:00:27  lr: 0.000053  loss: 2.1466 (2.1297)  grad: 0.2660 (0.2676)  time: 0.4468  data: 0.0051  max mem: 22448
+train: [15]  [360/400]  eta: 0:00:18  lr: 0.000052  loss: 2.1243 (2.1293)  grad: 0.2660 (0.2678)  time: 0.4372  data: 0.0052  max mem: 22448
+train: [15]  [380/400]  eta: 0:00:09  lr: 0.000051  loss: 2.0983 (2.1275)  grad: 0.2589 (0.2673)  time: 0.4481  data: 0.0051  max mem: 22448
+train: [15]  [399/400]  eta: 0:00:00  lr: 0.000050  loss: 2.1103 (2.1271)  grad: 0.2530 (0.2669)  time: 0.4419  data: 0.0051  max mem: 22448
+train: [15] Total time: 0:03:01 (0.4526 s / it)
+train: [15] Summary: lr: 0.000050  loss: 2.1103 (2.1271)  grad: 0.2530 (0.2669)
+eval (validation): [15]  [ 0/85]  eta: 0:04:46    time: 3.3666  data: 3.0691  max mem: 22448
+eval (validation): [15]  [20/85]  eta: 0:00:31    time: 0.3453  data: 0.0045  max mem: 22448
+eval (validation): [15]  [40/85]  eta: 0:00:18    time: 0.3488  data: 0.0040  max mem: 22448
+eval (validation): [15]  [60/85]  eta: 0:00:09    time: 0.3427  data: 0.0043  max mem: 22448
+eval (validation): [15]  [80/85]  eta: 0:00:01    time: 0.3283  data: 0.0042  max mem: 22448
+eval (validation): [15]  [84/85]  eta: 0:00:00    time: 0.3226  data: 0.0042  max mem: 22448
+eval (validation): [15] Total time: 0:00:32 (0.3793 s / it)
+cv: [15]  best hparam: (0.27, 1.0) (016) ('016_lr2.7e-01_wd1.0e+00')  loss: 2.426  acc: 0.269  f1: 0.198
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [16]  [  0/400]  eta: 0:22:45  lr: nan  time: 3.4149  data: 3.0298  max mem: 22448
+train: [16]  [ 20/400]  eta: 0:03:36  lr: 0.000048  loss: 2.0669 (2.0892)  grad: 0.2458 (0.2493)  time: 0.4275  data: 0.0045  max mem: 22448
+train: [16]  [ 40/400]  eta: 0:03:04  lr: 0.000047  loss: 2.0669 (2.0831)  grad: 0.2515 (0.2518)  time: 0.4511  data: 0.0047  max mem: 22448
+train: [16]  [ 60/400]  eta: 0:02:46  lr: 0.000046  loss: 2.0665 (2.0802)  grad: 0.2564 (0.2555)  time: 0.4451  data: 0.0054  max mem: 22448
+train: [16]  [ 80/400]  eta: 0:02:33  lr: 0.000045  loss: 2.1132 (2.0917)  grad: 0.2658 (0.2596)  time: 0.4533  data: 0.0051  max mem: 22448
+train: [16]  [100/400]  eta: 0:02:20  lr: 0.000044  loss: 2.1047 (2.0883)  grad: 0.2694 (0.2607)  time: 0.4255  data: 0.0049  max mem: 22448
+train: [16]  [120/400]  eta: 0:02:10  lr: 0.000043  loss: 2.0633 (2.0883)  grad: 0.2642 (0.2609)  time: 0.4529  data: 0.0051  max mem: 22448
+train: [16]  [140/400]  eta: 0:02:00  lr: 0.000042  loss: 2.0917 (2.0890)  grad: 0.2602 (0.2608)  time: 0.4431  data: 0.0050  max mem: 22448
+train: [16]  [160/400]  eta: 0:01:50  lr: 0.000041  loss: 2.0949 (2.0937)  grad: 0.2651 (0.2619)  time: 0.4231  data: 0.0048  max mem: 22448
+train: [16]  [180/400]  eta: 0:01:40  lr: 0.000040  loss: 2.1094 (2.0956)  grad: 0.2651 (0.2616)  time: 0.4531  data: 0.0053  max mem: 22448
+train: [16]  [200/400]  eta: 0:01:31  lr: 0.000039  loss: 2.0844 (2.0913)  grad: 0.2562 (0.2608)  time: 0.4501  data: 0.0049  max mem: 22448
+train: [16]  [220/400]  eta: 0:01:21  lr: 0.000038  loss: 2.0830 (2.0906)  grad: 0.2534 (0.2606)  time: 0.4361  data: 0.0049  max mem: 22448
+train: [16]  [240/400]  eta: 0:01:12  lr: 0.000036  loss: 2.0764 (2.0896)  grad: 0.2576 (0.2613)  time: 0.4602  data: 0.0052  max mem: 22448
+train: [16]  [260/400]  eta: 0:01:03  lr: 0.000035  loss: 2.0764 (2.0922)  grad: 0.2715 (0.2621)  time: 0.4538  data: 0.0053  max mem: 22448
+train: [16]  [280/400]  eta: 0:00:54  lr: 0.000034  loss: 2.1062 (2.0931)  grad: 0.2650 (0.2621)  time: 0.4334  data: 0.0050  max mem: 22448
+train: [16]  [300/400]  eta: 0:00:45  lr: 0.000033  loss: 2.1086 (2.0945)  grad: 0.2619 (0.2625)  time: 0.4517  data: 0.0052  max mem: 22448
+train: [16]  [320/400]  eta: 0:00:36  lr: 0.000032  loss: 2.1190 (2.0976)  grad: 0.2691 (0.2629)  time: 0.4468  data: 0.0051  max mem: 22448
+train: [16]  [340/400]  eta: 0:00:27  lr: 0.000031  loss: 2.0696 (2.0963)  grad: 0.2680 (0.2631)  time: 0.4538  data: 0.0050  max mem: 22448
+train: [16]  [360/400]  eta: 0:00:18  lr: 0.000031  loss: 2.0627 (2.0966)  grad: 0.2621 (0.2633)  time: 0.4522  data: 0.0049  max mem: 22448
+train: [16]  [380/400]  eta: 0:00:09  lr: 0.000030  loss: 2.0819 (2.0963)  grad: 0.2630 (0.2638)  time: 0.4277  data: 0.0050  max mem: 22448
+train: [16]  [399/400]  eta: 0:00:00  lr: 0.000029  loss: 2.0938 (2.0975)  grad: 0.2698 (0.2642)  time: 0.4515  data: 0.0049  max mem: 22448
+train: [16] Total time: 0:03:01 (0.4525 s / it)
+train: [16] Summary: lr: 0.000029  loss: 2.0938 (2.0975)  grad: 0.2698 (0.2642)
+eval (validation): [16]  [ 0/85]  eta: 0:04:42    time: 3.3287  data: 3.0722  max mem: 22448
+eval (validation): [16]  [20/85]  eta: 0:00:33    time: 0.3679  data: 0.0049  max mem: 22448
+eval (validation): [16]  [40/85]  eta: 0:00:19    time: 0.3500  data: 0.0037  max mem: 22448
+eval (validation): [16]  [60/85]  eta: 0:00:10    time: 0.3438  data: 0.0043  max mem: 22448
+eval (validation): [16]  [80/85]  eta: 0:00:01    time: 0.3226  data: 0.0039  max mem: 22448
+eval (validation): [16]  [84/85]  eta: 0:00:00    time: 0.3182  data: 0.0039  max mem: 22448
+eval (validation): [16] Total time: 0:00:32 (0.3829 s / it)
+cv: [16]  best hparam: (0.27, 1.0) (016) ('016_lr2.7e-01_wd1.0e+00')  loss: 2.424  acc: 0.272  f1: 0.200
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [17]  [  0/400]  eta: 0:22:11  lr: nan  time: 3.3279  data: 2.9813  max mem: 22448
+train: [17]  [ 20/400]  eta: 0:03:43  lr: 0.000028  loss: 2.0036 (2.0288)  grad: 0.2365 (0.2524)  time: 0.4506  data: 0.0248  max mem: 22448
+train: [17]  [ 40/400]  eta: 0:03:07  lr: 0.000027  loss: 2.0672 (2.0622)  grad: 0.2541 (0.2556)  time: 0.4481  data: 0.0044  max mem: 22448
+train: [17]  [ 60/400]  eta: 0:02:48  lr: 0.000026  loss: 2.0902 (2.0794)  grad: 0.2515 (0.2546)  time: 0.4445  data: 0.0050  max mem: 22448
+train: [17]  [ 80/400]  eta: 0:02:33  lr: 0.000025  loss: 2.0462 (2.0664)  grad: 0.2509 (0.2544)  time: 0.4366  data: 0.0052  max mem: 22448
+train: [17]  [100/400]  eta: 0:02:21  lr: 0.000024  loss: 2.0343 (2.0702)  grad: 0.2473 (0.2532)  time: 0.4342  data: 0.0049  max mem: 22448
+train: [17]  [120/400]  eta: 0:02:11  lr: 0.000023  loss: 2.0787 (2.0693)  grad: 0.2476 (0.2538)  time: 0.4557  data: 0.0053  max mem: 22448
+train: [17]  [140/400]  eta: 0:02:00  lr: 0.000023  loss: 2.0787 (2.0693)  grad: 0.2538 (0.2539)  time: 0.4399  data: 0.0052  max mem: 22448
+train: [17]  [160/400]  eta: 0:01:51  lr: 0.000022  loss: 2.0837 (2.0701)  grad: 0.2562 (0.2550)  time: 0.4494  data: 0.0051  max mem: 22448
+train: [17]  [180/400]  eta: 0:01:41  lr: 0.000021  loss: 2.0666 (2.0686)  grad: 0.2529 (0.2545)  time: 0.4469  data: 0.0051  max mem: 22448
+train: [17]  [200/400]  eta: 0:01:31  lr: 0.000020  loss: 2.0608 (2.0678)  grad: 0.2528 (0.2544)  time: 0.4459  data: 0.0050  max mem: 22448
+train: [17]  [220/400]  eta: 0:01:22  lr: 0.000019  loss: 2.0644 (2.0683)  grad: 0.2568 (0.2552)  time: 0.4469  data: 0.0050  max mem: 22448
+train: [17]  [240/400]  eta: 0:01:13  lr: 0.000019  loss: 2.0644 (2.0659)  grad: 0.2535 (0.2555)  time: 0.4544  data: 0.0051  max mem: 22448
+train: [17]  [260/400]  eta: 0:01:04  lr: 0.000018  loss: 2.0179 (2.0648)  grad: 0.2540 (0.2559)  time: 0.4527  data: 0.0050  max mem: 22448
+train: [17]  [280/400]  eta: 0:00:54  lr: 0.000017  loss: 2.0501 (2.0655)  grad: 0.2600 (0.2561)  time: 0.4382  data: 0.0050  max mem: 22448
+train: [17]  [300/400]  eta: 0:00:45  lr: 0.000016  loss: 2.0746 (2.0674)  grad: 0.2593 (0.2562)  time: 0.4528  data: 0.0052  max mem: 22448
+train: [17]  [320/400]  eta: 0:00:36  lr: 0.000016  loss: 2.0664 (2.0670)  grad: 0.2530 (0.2555)  time: 0.4470  data: 0.0049  max mem: 22448
+train: [17]  [340/400]  eta: 0:00:27  lr: 0.000015  loss: 2.0664 (2.0673)  grad: 0.2508 (0.2559)  time: 0.4456  data: 0.0050  max mem: 22448
+train: [17]  [360/400]  eta: 0:00:18  lr: 0.000014  loss: 2.0937 (2.0681)  grad: 0.2533 (0.2560)  time: 0.4524  data: 0.0052  max mem: 22448
+train: [17]  [380/400]  eta: 0:00:09  lr: 0.000014  loss: 2.0494 (2.0673)  grad: 0.2589 (0.2567)  time: 0.4356  data: 0.0049  max mem: 22448
+train: [17]  [399/400]  eta: 0:00:00  lr: 0.000013  loss: 2.0188 (2.0663)  grad: 0.2597 (0.2565)  time: 0.4462  data: 0.0049  max mem: 22448
+train: [17] Total time: 0:03:01 (0.4541 s / it)
+train: [17] Summary: lr: 0.000013  loss: 2.0188 (2.0663)  grad: 0.2597 (0.2565)
+eval (validation): [17]  [ 0/85]  eta: 0:04:47    time: 3.3877  data: 3.1394  max mem: 22448
+eval (validation): [17]  [20/85]  eta: 0:00:32    time: 0.3495  data: 0.0049  max mem: 22448
+eval (validation): [17]  [40/85]  eta: 0:00:19    time: 0.3523  data: 0.0035  max mem: 22448
+eval (validation): [17]  [60/85]  eta: 0:00:10    time: 0.3573  data: 0.0044  max mem: 22448
+eval (validation): [17]  [80/85]  eta: 0:00:01    time: 0.3234  data: 0.0041  max mem: 22448
+eval (validation): [17]  [84/85]  eta: 0:00:00    time: 0.3108  data: 0.0039  max mem: 22448
+eval (validation): [17] Total time: 0:00:32 (0.3825 s / it)
+cv: [17]  best hparam: (0.27, 1.0) (016) ('016_lr2.7e-01_wd1.0e+00')  loss: 2.420  acc: 0.272  f1: 0.201
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [18]  [  0/400]  eta: 0:22:57  lr: nan  time: 3.4442  data: 3.0506  max mem: 22448
+train: [18]  [ 20/400]  eta: 0:03:48  lr: 0.000012  loss: 2.1094 (2.1033)  grad: 0.2486 (0.2546)  time: 0.4590  data: 0.0049  max mem: 22448
+train: [18]  [ 40/400]  eta: 0:03:07  lr: 0.000012  loss: 2.0405 (2.0512)  grad: 0.2510 (0.2527)  time: 0.4356  data: 0.0048  max mem: 22448
+train: [18]  [ 60/400]  eta: 0:02:47  lr: 0.000011  loss: 1.9924 (2.0426)  grad: 0.2571 (0.2570)  time: 0.4322  data: 0.0050  max mem: 22448
+train: [18]  [ 80/400]  eta: 0:02:33  lr: 0.000011  loss: 2.0018 (2.0411)  grad: 0.2567 (0.2549)  time: 0.4389  data: 0.0052  max mem: 22448
+train: [18]  [100/400]  eta: 0:02:21  lr: 0.000010  loss: 2.0655 (2.0411)  grad: 0.2504 (0.2543)  time: 0.4371  data: 0.0048  max mem: 22448
+train: [18]  [120/400]  eta: 0:02:10  lr: 0.000009  loss: 2.0057 (2.0343)  grad: 0.2487 (0.2536)  time: 0.4485  data: 0.0053  max mem: 22448
+train: [18]  [140/400]  eta: 0:02:00  lr: 0.000009  loss: 2.0278 (2.0387)  grad: 0.2529 (0.2542)  time: 0.4537  data: 0.0051  max mem: 22448
+train: [18]  [160/400]  eta: 0:01:50  lr: 0.000008  loss: 2.0371 (2.0404)  grad: 0.2529 (0.2534)  time: 0.4433  data: 0.0047  max mem: 22448
+train: [18]  [180/400]  eta: 0:01:41  lr: 0.000008  loss: 2.0258 (2.0383)  grad: 0.2548 (0.2542)  time: 0.4534  data: 0.0054  max mem: 22448
+train: [18]  [200/400]  eta: 0:01:31  lr: 0.000007  loss: 2.0258 (2.0406)  grad: 0.2548 (0.2541)  time: 0.4443  data: 0.0050  max mem: 22448
+train: [18]  [220/400]  eta: 0:01:22  lr: 0.000007  loss: 2.0573 (2.0406)  grad: 0.2528 (0.2543)  time: 0.4647  data: 0.0051  max mem: 22448
+train: [18]  [240/400]  eta: 0:01:13  lr: 0.000006  loss: 2.0575 (2.0440)  grad: 0.2546 (0.2545)  time: 0.4563  data: 0.0049  max mem: 22448
+train: [18]  [260/400]  eta: 0:01:04  lr: 0.000006  loss: 2.0709 (2.0454)  grad: 0.2523 (0.2544)  time: 0.4354  data: 0.0047  max mem: 22448
+train: [18]  [280/400]  eta: 0:00:54  lr: 0.000006  loss: 2.0580 (2.0451)  grad: 0.2519 (0.2544)  time: 0.4414  data: 0.0049  max mem: 22448
+train: [18]  [300/400]  eta: 0:00:45  lr: 0.000005  loss: 2.0558 (2.0458)  grad: 0.2517 (0.2542)  time: 0.4579  data: 0.0052  max mem: 22448
+train: [18]  [320/400]  eta: 0:00:36  lr: 0.000005  loss: 2.0841 (2.0488)  grad: 0.2517 (0.2541)  time: 0.4484  data: 0.0051  max mem: 22448
+train: [18]  [340/400]  eta: 0:00:27  lr: 0.000004  loss: 2.0363 (2.0464)  grad: 0.2524 (0.2537)  time: 0.4492  data: 0.0051  max mem: 22448
+train: [18]  [360/400]  eta: 0:00:18  lr: 0.000004  loss: 2.0343 (2.0458)  grad: 0.2516 (0.2536)  time: 0.4461  data: 0.0049  max mem: 22448
+train: [18]  [380/400]  eta: 0:00:09  lr: 0.000004  loss: 2.0392 (2.0448)  grad: 0.2516 (0.2536)  time: 0.4432  data: 0.0050  max mem: 22448
+train: [18]  [399/400]  eta: 0:00:00  lr: 0.000003  loss: 2.0269 (2.0429)  grad: 0.2498 (0.2538)  time: 0.4491  data: 0.0050  max mem: 22448
+train: [18] Total time: 0:03:02 (0.4550 s / it)
+train: [18] Summary: lr: 0.000003  loss: 2.0269 (2.0429)  grad: 0.2498 (0.2538)
+eval (validation): [18]  [ 0/85]  eta: 0:04:47    time: 3.3823  data: 3.1419  max mem: 22448
+eval (validation): [18]  [20/85]  eta: 0:00:34    time: 0.3855  data: 0.0043  max mem: 22448
+eval (validation): [18]  [40/85]  eta: 0:00:19    time: 0.3360  data: 0.0040  max mem: 22448
+eval (validation): [18]  [60/85]  eta: 0:00:10    time: 0.3480  data: 0.0043  max mem: 22448
+eval (validation): [18]  [80/85]  eta: 0:00:01    time: 0.3219  data: 0.0039  max mem: 22448
+eval (validation): [18]  [84/85]  eta: 0:00:00    time: 0.3204  data: 0.0039  max mem: 22448
+eval (validation): [18] Total time: 0:00:32 (0.3858 s / it)
+cv: [18]  best hparam: (0.27, 1.0) (016) ('016_lr2.7e-01_wd1.0e+00')  loss: 2.423  acc: 0.271  f1: 0.200
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [19]  [  0/400]  eta: 0:27:35  lr: nan  time: 4.1396  data: 3.8006  max mem: 22448
+train: [19]  [ 20/400]  eta: 0:03:54  lr: 0.000003  loss: 2.0197 (2.0603)  grad: 0.2396 (0.2472)  time: 0.4422  data: 0.0027  max mem: 22448
+train: [19]  [ 40/400]  eta: 0:03:11  lr: 0.000003  loss: 2.0197 (2.0297)  grad: 0.2472 (0.2495)  time: 0.4405  data: 0.0049  max mem: 22448
+train: [19]  [ 60/400]  eta: 0:02:50  lr: 0.000002  loss: 2.0304 (2.0459)  grad: 0.2467 (0.2495)  time: 0.4430  data: 0.0051  max mem: 22448
+train: [19]  [ 80/400]  eta: 0:02:36  lr: 0.000002  loss: 2.0552 (2.0471)  grad: 0.2461 (0.2485)  time: 0.4476  data: 0.0049  max mem: 22448
+train: [19]  [100/400]  eta: 0:02:24  lr: 0.000002  loss: 2.0300 (2.0443)  grad: 0.2471 (0.2492)  time: 0.4479  data: 0.0051  max mem: 22448
+train: [19]  [120/400]  eta: 0:02:12  lr: 0.000002  loss: 2.0361 (2.0484)  grad: 0.2471 (0.2503)  time: 0.4358  data: 0.0050  max mem: 22448
+train: [19]  [140/400]  eta: 0:02:02  lr: 0.000001  loss: 2.0409 (2.0456)  grad: 0.2438 (0.2495)  time: 0.4517  data: 0.0051  max mem: 22448
+train: [19]  [160/400]  eta: 0:01:51  lr: 0.000001  loss: 2.0248 (2.0426)  grad: 0.2413 (0.2486)  time: 0.4360  data: 0.0052  max mem: 22448
+train: [19]  [180/400]  eta: 0:01:41  lr: 0.000001  loss: 2.0129 (2.0404)  grad: 0.2502 (0.2500)  time: 0.4438  data: 0.0053  max mem: 22448
+train: [19]  [200/400]  eta: 0:01:32  lr: 0.000001  loss: 2.0089 (2.0385)  grad: 0.2512 (0.2496)  time: 0.4468  data: 0.0054  max mem: 22448
+train: [19]  [220/400]  eta: 0:01:23  lr: 0.000001  loss: 2.0168 (2.0378)  grad: 0.2442 (0.2490)  time: 0.4548  data: 0.0053  max mem: 22448
+train: [19]  [240/400]  eta: 0:01:13  lr: 0.000001  loss: 2.0334 (2.0364)  grad: 0.2453 (0.2496)  time: 0.4532  data: 0.0051  max mem: 22448
+train: [19]  [260/400]  eta: 0:01:04  lr: 0.000000  loss: 2.0209 (2.0342)  grad: 0.2490 (0.2495)  time: 0.4384  data: 0.0051  max mem: 22448
+train: [19]  [280/400]  eta: 0:00:54  lr: 0.000000  loss: 2.0409 (2.0378)  grad: 0.2500 (0.2497)  time: 0.4401  data: 0.0051  max mem: 22448
+train: [19]  [300/400]  eta: 0:00:45  lr: 0.000000  loss: 2.0573 (2.0371)  grad: 0.2516 (0.2496)  time: 0.4467  data: 0.0052  max mem: 22448
+train: [19]  [320/400]  eta: 0:00:36  lr: 0.000000  loss: 2.0566 (2.0399)  grad: 0.2535 (0.2501)  time: 0.4439  data: 0.0051  max mem: 22448
+train: [19]  [340/400]  eta: 0:00:27  lr: 0.000000  loss: 2.0707 (2.0420)  grad: 0.2482 (0.2500)  time: 0.4310  data: 0.0048  max mem: 22448
+train: [19]  [360/400]  eta: 0:00:18  lr: 0.000000  loss: 2.0841 (2.0430)  grad: 0.2438 (0.2497)  time: 0.4477  data: 0.0051  max mem: 22448
+train: [19]  [380/400]  eta: 0:00:09  lr: 0.000000  loss: 2.0280 (2.0426)  grad: 0.2457 (0.2496)  time: 0.4390  data: 0.0050  max mem: 22448
+train: [19]  [399/400]  eta: 0:00:00  lr: 0.000000  loss: 2.0331 (2.0438)  grad: 0.2460 (0.2496)  time: 0.4389  data: 0.0050  max mem: 22448
+train: [19] Total time: 0:03:01 (0.4534 s / it)
+train: [19] Summary: lr: 0.000000  loss: 2.0331 (2.0438)  grad: 0.2460 (0.2496)
+eval (validation): [19]  [ 0/85]  eta: 0:04:34    time: 3.2287  data: 2.9470  max mem: 22448
+eval (validation): [19]  [20/85]  eta: 0:00:32    time: 0.3701  data: 0.0320  max mem: 22448
+eval (validation): [19]  [40/85]  eta: 0:00:19    time: 0.3642  data: 0.0031  max mem: 22448
+eval (validation): [19]  [60/85]  eta: 0:00:10    time: 0.3731  data: 0.0040  max mem: 22448
+eval (validation): [19]  [80/85]  eta: 0:00:01    time: 0.3204  data: 0.0038  max mem: 22448
+eval (validation): [19]  [84/85]  eta: 0:00:00    time: 0.3121  data: 0.0037  max mem: 22448
+eval (validation): [19] Total time: 0:00:33 (0.3920 s / it)
+cv: [19]  best hparam: (0.27, 1.0) (016) ('016_lr2.7e-01_wd1.0e+00')  loss: 2.422  acc: 0.272  f1: 0.201
+saving checkpoint experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+evaluating last checkpoint: experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+eval model info:
+{"score": 0.27150239940937615, "hparam": [0.27, 1.0], "hparam_id": 16, "epoch": 19, "is_best": false, "best_score": 0.2737172388335179}
+eval (train): [20]  [  0/509]  eta: 0:26:29    time: 3.1236  data: 2.8258  max mem: 22448
+eval (train): [20]  [ 20/509]  eta: 0:04:17    time: 0.3978  data: 0.0049  max mem: 22448
+eval (train): [20]  [ 40/509]  eta: 0:03:31    time: 0.3713  data: 0.0047  max mem: 22448
+eval (train): [20]  [ 60/509]  eta: 0:03:07    time: 0.3497  data: 0.0043  max mem: 22448
+eval (train): [20]  [ 80/509]  eta: 0:02:53    time: 0.3595  data: 0.0045  max mem: 22448
+eval (train): [20]  [100/509]  eta: 0:02:38    time: 0.3258  data: 0.0041  max mem: 22448
+eval (train): [20]  [120/509]  eta: 0:02:28    time: 0.3511  data: 0.0041  max mem: 22448
+eval (train): [20]  [140/509]  eta: 0:02:19    time: 0.3634  data: 0.0046  max mem: 22448
+eval (train): [20]  [160/509]  eta: 0:02:10    time: 0.3408  data: 0.0045  max mem: 22448
+eval (train): [20]  [180/509]  eta: 0:02:02    time: 0.3422  data: 0.0042  max mem: 22448
+eval (train): [20]  [200/509]  eta: 0:01:53    time: 0.3460  data: 0.0043  max mem: 22448
+eval (train): [20]  [220/509]  eta: 0:01:45    time: 0.3393  data: 0.0041  max mem: 22448
+eval (train): [20]  [240/509]  eta: 0:01:37    time: 0.3406  data: 0.0043  max mem: 22448
+eval (train): [20]  [260/509]  eta: 0:01:29    time: 0.3286  data: 0.0040  max mem: 22448
+eval (train): [20]  [280/509]  eta: 0:01:22    time: 0.3787  data: 0.0048  max mem: 22448
+eval (train): [20]  [300/509]  eta: 0:01:15    time: 0.3507  data: 0.0040  max mem: 22448
+eval (train): [20]  [320/509]  eta: 0:01:08    time: 0.3415  data: 0.0040  max mem: 22448
+eval (train): [20]  [340/509]  eta: 0:01:00    time: 0.3265  data: 0.0041  max mem: 22448
+eval (train): [20]  [360/509]  eta: 0:00:53    time: 0.3529  data: 0.0044  max mem: 22448
+eval (train): [20]  [380/509]  eta: 0:00:46    time: 0.3458  data: 0.0044  max mem: 22448
+eval (train): [20]  [400/509]  eta: 0:00:38    time: 0.3579  data: 0.0047  max mem: 22448
+eval (train): [20]  [420/509]  eta: 0:00:31    time: 0.3497  data: 0.0043  max mem: 22448
+eval (train): [20]  [440/509]  eta: 0:00:24    time: 0.3243  data: 0.0041  max mem: 22448
+eval (train): [20]  [460/509]  eta: 0:00:17    time: 0.3473  data: 0.0044  max mem: 22448
+eval (train): [20]  [480/509]  eta: 0:00:10    time: 0.3505  data: 0.0045  max mem: 22448
+eval (train): [20]  [500/509]  eta: 0:00:03    time: 0.3430  data: 0.0040  max mem: 22448
+eval (train): [20]  [508/509]  eta: 0:00:00    time: 0.3245  data: 0.0039  max mem: 22448
+eval (train): [20] Total time: 0:03:01 (0.3561 s / it)
+eval (validation): [20]  [ 0/85]  eta: 0:04:24    time: 3.1064  data: 2.8701  max mem: 22448
+eval (validation): [20]  [20/85]  eta: 0:00:32    time: 0.3622  data: 0.0034  max mem: 22448
+eval (validation): [20]  [40/85]  eta: 0:00:19    time: 0.3628  data: 0.0040  max mem: 22448
+eval (validation): [20]  [60/85]  eta: 0:00:10    time: 0.3532  data: 0.0042  max mem: 22448
+eval (validation): [20]  [80/85]  eta: 0:00:01    time: 0.3243  data: 0.0040  max mem: 22448
+eval (validation): [20]  [84/85]  eta: 0:00:00    time: 0.3198  data: 0.0039  max mem: 22448
+eval (validation): [20] Total time: 0:00:32 (0.3846 s / it)
+eval (test): [20]  [ 0/85]  eta: 0:04:31    time: 3.1907  data: 2.8804  max mem: 22448
+eval (test): [20]  [20/85]  eta: 0:00:32    time: 0.3629  data: 0.0053  max mem: 22448
+eval (test): [20]  [40/85]  eta: 0:00:20    time: 0.3981  data: 0.0044  max mem: 22448
+eval (test): [20]  [60/85]  eta: 0:00:10    time: 0.3551  data: 0.0044  max mem: 22448
+eval (test): [20]  [80/85]  eta: 0:00:01    time: 0.3316  data: 0.0043  max mem: 22448
+eval (test): [20]  [84/85]  eta: 0:00:00    time: 0.3150  data: 0.0040  max mem: 22448
+eval (test): [20] Total time: 0:00:33 (0.3947 s / it)
+eval (testid): [20]  [ 0/82]  eta: 0:04:03    time: 2.9740  data: 2.6964  max mem: 22448
+eval (testid): [20]  [20/82]  eta: 0:00:31    time: 0.3797  data: 0.0047  max mem: 22448
+eval (testid): [20]  [40/82]  eta: 0:00:18    time: 0.3722  data: 0.0044  max mem: 22448
+eval (testid): [20]  [60/82]  eta: 0:00:09    time: 0.3683  data: 0.0049  max mem: 22448
+eval (testid): [20]  [80/82]  eta: 0:00:00    time: 0.3390  data: 0.0050  max mem: 22448
+eval (testid): [20]  [81/82]  eta: 0:00:00    time: 0.3242  data: 0.0047  max mem: 22448
+eval (testid): [20] Total time: 0:00:32 (0.3980 s / it)
+evaluating best checkpoint: experiments/data_scaling/output/data_scaling/n200_2/eval_v2/nsd_cococlip__patch__attn/checkpoint-best.pth
+eval model info:
+{"score": 0.2737172388335179, "hparam": [0.52, 1.0], "hparam_id": 20, "epoch": 6, "is_best": true, "best_score": 0.2737172388335179}
+eval (train): [20]  [  0/509]  eta: 0:26:20    time: 3.1046  data: 2.8597  max mem: 22448
+eval (train): [20]  [ 20/509]  eta: 0:04:12    time: 0.3876  data: 0.0145  max mem: 22448
+eval (train): [20]  [ 40/509]  eta: 0:03:22    time: 0.3426  data: 0.0034  max mem: 22448
+eval (train): [20]  [ 60/509]  eta: 0:03:01    time: 0.3502  data: 0.0041  max mem: 22448
+eval (train): [20]  [ 80/509]  eta: 0:02:48    time: 0.3534  data: 0.0044  max mem: 22448
+eval (train): [20]  [100/509]  eta: 0:02:41    time: 0.4094  data: 0.0048  max mem: 22448
+eval (train): [20]  [120/509]  eta: 0:02:34    time: 0.4093  data: 0.0050  max mem: 22448
+eval (train): [20]  [140/509]  eta: 0:02:24    time: 0.3532  data: 0.0046  max mem: 22448
+eval (train): [20]  [160/509]  eta: 0:02:15    time: 0.3644  data: 0.0046  max mem: 22448
+eval (train): [20]  [180/509]  eta: 0:02:06    time: 0.3588  data: 0.0046  max mem: 22448
+eval (train): [20]  [200/509]  eta: 0:01:58    time: 0.3683  data: 0.0046  max mem: 22448
+eval (train): [20]  [220/509]  eta: 0:01:50    time: 0.3745  data: 0.0046  max mem: 22448
+eval (train): [20]  [240/509]  eta: 0:01:42    time: 0.3753  data: 0.0046  max mem: 22448
+eval (train): [20]  [260/509]  eta: 0:01:34    time: 0.3575  data: 0.0044  max mem: 22448
+eval (train): [20]  [280/509]  eta: 0:01:26    time: 0.3507  data: 0.0044  max mem: 22448
+eval (train): [20]  [300/509]  eta: 0:01:18    time: 0.3545  data: 0.0046  max mem: 22448
+eval (train): [20]  [320/509]  eta: 0:01:10    time: 0.3409  data: 0.0039  max mem: 22448
+eval (train): [20]  [340/509]  eta: 0:01:03    time: 0.3613  data: 0.0042  max mem: 22448
+eval (train): [20]  [360/509]  eta: 0:00:55    time: 0.3600  data: 0.0042  max mem: 22448
+eval (train): [20]  [380/509]  eta: 0:00:48    time: 0.3876  data: 0.0046  max mem: 22448
+eval (train): [20]  [400/509]  eta: 0:00:40    time: 0.3679  data: 0.0045  max mem: 22448
+eval (train): [20]  [420/509]  eta: 0:00:33    time: 0.3357  data: 0.0041  max mem: 22448
+eval (train): [20]  [440/509]  eta: 0:00:25    time: 0.3744  data: 0.0043  max mem: 22448
+eval (train): [20]  [460/509]  eta: 0:00:18    time: 0.3457  data: 0.0037  max mem: 22448
+eval (train): [20]  [480/509]  eta: 0:00:10    time: 0.3630  data: 0.0043  max mem: 22448
+eval (train): [20]  [500/509]  eta: 0:00:03    time: 0.3646  data: 0.0045  max mem: 22448
+eval (train): [20]  [508/509]  eta: 0:00:00    time: 0.3427  data: 0.0038  max mem: 22448
+eval (train): [20] Total time: 0:03:08 (0.3713 s / it)
+eval (validation): [20]  [ 0/85]  eta: 0:04:00    time: 2.8269  data: 2.5972  max mem: 22448
+eval (validation): [20]  [20/85]  eta: 0:00:33    time: 0.3997  data: 0.0245  max mem: 22448
+eval (validation): [20]  [40/85]  eta: 0:00:19    time: 0.3635  data: 0.0041  max mem: 22448
+eval (validation): [20]  [60/85]  eta: 0:00:10    time: 0.3604  data: 0.0045  max mem: 22448
+eval (validation): [20]  [80/85]  eta: 0:00:01    time: 0.3304  data: 0.0039  max mem: 22448
+eval (validation): [20]  [84/85]  eta: 0:00:00    time: 0.3126  data: 0.0036  max mem: 22448
+eval (validation): [20] Total time: 0:00:33 (0.3932 s / it)
+eval (test): [20]  [ 0/85]  eta: 0:04:19    time: 3.0568  data: 2.7767  max mem: 22448
+eval (test): [20]  [20/85]  eta: 0:00:35    time: 0.4136  data: 0.0050  max mem: 22448
+eval (test): [20]  [40/85]  eta: 0:00:20    time: 0.3830  data: 0.0048  max mem: 22448
+eval (test): [20]  [60/85]  eta: 0:00:10    time: 0.3669  data: 0.0044  max mem: 22448
+eval (test): [20]  [80/85]  eta: 0:00:02    time: 0.3755  data: 0.0046  max mem: 22448
+eval (test): [20]  [84/85]  eta: 0:00:00    time: 0.3503  data: 0.0042  max mem: 22448
+eval (test): [20] Total time: 0:00:35 (0.4167 s / it)
+eval (testid): [20]  [ 0/82]  eta: 0:04:18    time: 3.1475  data: 2.8597  max mem: 22448
+eval (testid): [20]  [20/82]  eta: 0:00:32    time: 0.3848  data: 0.0056  max mem: 22448
+eval (testid): [20]  [40/82]  eta: 0:00:18    time: 0.3527  data: 0.0037  max mem: 22448
+eval (testid): [20]  [60/82]  eta: 0:00:09    time: 0.3655  data: 0.0044  max mem: 22448
+eval (testid): [20]  [80/82]  eta: 0:00:00    time: 0.3360  data: 0.0043  max mem: 22448
+eval (testid): [20]  [81/82]  eta: 0:00:00    time: 0.3240  data: 0.0041  max mem: 22448
+eval (testid): [20] Total time: 0:00:32 (0.3951 s / it)
+eval results:
+
+| model    | repr   | clf   | dataset      | ckpt   |   epoch |       lr |   wd |   hparam_id | hparam      | split      |   loss |     acc |   acc_std |      f1 |    f1_std |
+|:---------|:-------|:------|:-------------|:-------|--------:|---------:|-----:|------------:|:------------|:-----------|-------:|--------:|----------:|--------:|----------:|
+| flat_mae | patch  | attn  | nsd_cococlip | best   |       6 | 0.000156 | 0.05 |          20 | [0.52, 1.0] | train      | 2.2521 | 0.32767 | 0.0022201 | 0.26225 | 0.0022267 |
+| flat_mae | patch  | attn  | nsd_cococlip | best   |       6 | 0.000156 | 0.05 |          20 | [0.52, 1.0] | validation | 2.4153 | 0.27372 | 0.0051447 | 0.20614 | 0.0046339 |
+| flat_mae | patch  | attn  | nsd_cococlip | best   |       6 | 0.000156 | 0.05 |          20 | [0.52, 1.0] | test       | 2.3897 | 0.27959 | 0.0053009 | 0.202   | 0.0048196 |
+| flat_mae | patch  | attn  | nsd_cococlip | best   |       6 | 0.000156 | 0.05 |          20 | [0.52, 1.0] | testid     | 2.3752 | 0.27916 | 0.0055198 | 0.21536 | 0.0049827 |
+
+
+done! total time: 1:23:07

data_scaling/n200_2/eval_v2/ppmi_dx__patch__logistic/log.txt

@@ -0,0 +1,247 @@
+fMRI foundation model logistic probe eval
+version: 0.1.dev66+g7ddd3aa04
+sha: 58906bf7243fb545e1349221e6921a1797e2e666, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-02-26 17:21:48
+config:
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n200_2; eval v2 (ppmi_dx patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n200_2/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n200_2/eval_v2/ppmi_dx__patch__logistic
+model: flat_mae
+representation: patch
+dataset: ppmi_dx
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n200_2/eval_v2/ppmi_dx__patch__logistic
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=True, reg_tokens=0, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: ppmi_dx (flat)
+train (n=463):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'dir', 'sex', 'age', 'age_bin', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 463
+}),
+    labels=['PD' 'Prodromal'],
+    counts=[178 285]
+)
+
+validation (n=99):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'dir', 'sex', 'age', 'age_bin', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 99
+}),
+    labels=['PD' 'Prodromal'],
+    counts=[39 60]
+)
+
+test (n=100):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'dir', 'sex', 'age', 'age_bin', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 100
+}),
+    labels=['PD' 'Prodromal'],
+    counts=[37 63]
+)
+
+extracting features for all splits
+extract (train)  [  0/232]  eta: 0:14:38    time: 3.7856  data: 2.9313  max mem: 2698
+extract (train)  [ 20/232]  eta: 0:01:16    time: 0.1890  data: 0.0601  max mem: 2851
+extract (train)  [ 40/232]  eta: 0:00:51    time: 0.1671  data: 0.0498  max mem: 2851
+extract (train)  [ 60/232]  eta: 0:00:40    time: 0.1758  data: 0.0546  max mem: 2851
+extract (train)  [ 80/232]  eta: 0:00:33    time: 0.1709  data: 0.0519  max mem: 2851
+extract (train)  [100/232]  eta: 0:00:27    time: 0.1754  data: 0.0571  max mem: 2851
+extract (train)  [120/232]  eta: 0:00:22    time: 0.1531  data: 0.0443  max mem: 2851
+extract (train)  [140/232]  eta: 0:00:18    time: 0.1666  data: 0.0528  max mem: 2851
+extract (train)  [160/232]  eta: 0:00:13    time: 0.1595  data: 0.0470  max mem: 2851
+extract (train)  [180/232]  eta: 0:00:09    time: 0.1664  data: 0.0525  max mem: 2851
+extract (train)  [200/232]  eta: 0:00:05    time: 0.1620  data: 0.0506  max mem: 2851
+extract (train)  [220/232]  eta: 0:00:02    time: 0.1453  data: 0.0412  max mem: 2851
+extract (train)  [231/232]  eta: 0:00:00    time: 0.1378  data: 0.0391  max mem: 2851
+extract (train) Total time: 0:00:42 (0.1819 s / it)
+extract (validation)  [ 0/50]  eta: 0:02:20    time: 2.8169  data: 2.6676  max mem: 2851
+extract (validation)  [20/50]  eta: 0:00:09    time: 0.2090  data: 0.0740  max mem: 2851
+extract (validation)  [40/50]  eta: 0:00:02    time: 0.1357  data: 0.0350  max mem: 2851
+extract (validation)  [49/50]  eta: 0:00:00    time: 0.1339  data: 0.0350  max mem: 2851
+extract (validation) Total time: 0:00:11 (0.2245 s / it)
+extract (test)  [ 0/50]  eta: 0:02:29    time: 2.9931  data: 2.8634  max mem: 2851
+extract (test)  [20/50]  eta: 0:00:09    time: 0.1930  data: 0.0628  max mem: 2851
+extract (test)  [40/50]  eta: 0:00:02    time: 0.1538  data: 0.0456  max mem: 2851
+extract (test)  [49/50]  eta: 0:00:00    time: 0.1464  data: 0.0428  max mem: 2851
+extract (test) Total time: 0:00:11 (0.2312 s / it)
+feature extraction time: 0:01:05
+train features: (463, 768)
+validation features: (99, 768)
+test features: (100, 768)
+evaluating fixed splits
+eval results (fixed splits):
+
+| model    | repr   | clf      | dataset   | trial   |         C | split   |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:----------|:--------|----------:|:--------|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | ppmi_dx   |         | 0.0059948 | train   | 0.70819 |  0.018364 | 0.66683 | 0.022029 | 0.6623  |   0.0203   |
+| flat_mae | patch  | logistic | ppmi_dx   |         | 0.0059948 | test    | 0.62    |  0.038342 | 0.52877 | 0.049457 | 0.54226 |   0.041356 |
+
+
+evaluating random splits (n=100)
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 1, "C": 0.000774263682681127, "split": "test", "acc": 0.68, "acc_std": 0.030736629613540897, "f1": 0.5733333333333333, "f1_std": 0.05189001976482154, "bacc": 0.5942275042444822, "bacc_std": 0.03683512850271709}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 2, "C": 0.046415888336127774, "split": "test", "acc": 0.66, "acc_std": 0.04277136892829127, "f1": 0.6155585707824514, "f1_std": 0.04936262247944019, "bacc": 0.6137521222410866, "bacc_std": 0.04616418432116035}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 3, "C": 0.3593813663804626, "split": "test", "acc": 0.64, "acc_std": 0.04848987935641828, "f1": 0.6179966044142615, "f1_std": 0.05050006009842691, "bacc": 0.6179966044142615, "bacc_std": 0.050408197170613095}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 4, "C": 0.005994842503189409, "split": "test", "acc": 0.65, "acc_std": 0.044444464222217824, "f1": 0.6178622120318812, "f1_std": 0.0483860205983203, "bacc": 0.615874363327674, "bacc_std": 0.04680937673319964}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 5, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.04224012784071563, "f1": 0.5324918186068257, "f1_std": 0.04990448951082515, "bacc": 0.5398981324278438, "bacc_std": 0.044283867469402716}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 6, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.04451380010738243, "f1": 0.5324918186068257, "f1_std": 0.05261052105748719, "bacc": 0.5398981324278438, "bacc_std": 0.046844995667879634}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 7, "C": 2.782559402207126, "split": "test", "acc": 0.47, "acc_std": 0.05139017804989589, "f1": 0.4403970013726111, "f1_std": 0.051665368136301855, "bacc": 0.4401528013582343, "bacc_std": 0.05200238970440739}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 8, "C": 0.046415888336127774, "split": "test", "acc": 0.64, "acc_std": 0.044051220187413656, "f1": 0.5863970588235294, "f1_std": 0.052056350631273174, "bacc": 0.5874363327674024, "bacc_std": 0.04724368307900424}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 9, "C": 0.005994842503189409, "split": "test", "acc": 0.73, "acc_std": 0.03768944679880564, "f1": 0.6754417598269022, "f1_std": 0.050577116921077185, "bacc": 0.6702037351443124, "bacc_std": 0.04406554199505446}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 10, "C": 0.046415888336127774, "split": "test", "acc": 0.64, "acc_std": 0.04443790724145322, "f1": 0.6043956043956044, "f1_std": 0.049265180585515815, "bacc": 0.6027164685908319, "bacc_std": 0.04750955857849844}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 11, "C": 0.005994842503189409, "split": "test", "acc": 0.69, "acc_std": 0.040212689539497345, "f1": 0.6408295678368672, "f1_std": 0.0509312152451608, "bacc": 0.6379456706281834, "bacc_std": 0.04584673033014523}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 12, "C": 0.3593813663804626, "split": "test", "acc": 0.59, "acc_std": 0.04825188493727473, "f1": 0.5577607593571352, "f1_std": 0.0518659397048125, "bacc": 0.5573005093378608, "bacc_std": 0.050764171992111}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 13, "C": 0.005994842503189409, "split": "test", "acc": 0.65, "acc_std": 0.040304893003207436, "f1": 0.5872154735228211, "f1_std": 0.04987334119805604, "bacc": 0.5904074702886248, "bacc_std": 0.04374388828264086}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 14, "C": 0.046415888336127774, "split": "test", "acc": 0.62, "acc_std": 0.04498719373332815, "f1": 0.5766488413547237, "f1_std": 0.05078639753526829, "bacc": 0.5764006791171477, "bacc_std": 0.04804089903406483}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 15, "C": 0.046415888336127774, "split": "test", "acc": 0.67, "acc_std": 0.04330683086996784, "f1": 0.6349153667441089, "f1_std": 0.048321722942161256, "bacc": 0.6320033955857385, "bacc_std": 0.04625283487390942}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 16, "C": 2.782559402207126, "split": "test", "acc": 0.59, "acc_std": 0.05059563617546478, "f1": 0.5777983729790959, "f1_std": 0.05142044196964082, "bacc": 0.58276740237691, "bacc_std": 0.0526037340494783}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 17, "C": 0.005994842503189409, "split": "test", "acc": 0.64, "acc_std": 0.03856329342781812, "f1": 0.5628946090335114, "f1_std": 0.04917020115973005, "bacc": 0.5721561969439728, "bacc_std": 0.04170609521593963}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 18, "C": 0.005994842503189409, "split": "test", "acc": 0.59, "acc_std": 0.042491617996965, "f1": 0.5071523019593701, "f1_std": 0.05090138759945219, "bacc": 0.5216468590831919, "bacc_std": 0.044021137704484334}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 19, "C": 21.54434690031882, "split": "test", "acc": 0.52, "acc_std": 0.05258934873146843, "f1": 0.5073891625615763, "f1_std": 0.05291226820519527, "bacc": 0.5110356536502546, "bacc_std": 0.05422756704125392}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 20, "C": 0.005994842503189409, "split": "test", "acc": 0.62, "acc_std": 0.04369902058399021, "f1": 0.5634191176470589, "f1_std": 0.05118711544378891, "bacc": 0.566213921901528, "bacc_std": 0.046604896158455617}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 21, "C": 0.046415888336127774, "split": "test", "acc": 0.66, "acc_std": 0.04264118197236094, "f1": 0.609375, "f1_std": 0.050920675764732984, "bacc": 0.6086587436332768, "bacc_std": 0.04633493208437767}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 22, "C": 0.046415888336127774, "split": "test", "acc": 0.61, "acc_std": 0.046937486085217646, "f1": 0.5623386825272135, "f1_std": 0.05279943125632097, "bacc": 0.5632427843803056, "bacc_std": 0.04940461508775626}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 23, "C": 0.046415888336127774, "split": "test", "acc": 0.58, "acc_std": 0.050275640224665456, "f1": 0.5384615384615385, "f1_std": 0.05410545452032071, "bacc": 0.5390492359932089, "bacc_std": 0.05216279865791108}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 24, "C": 0.005994842503189409, "split": "test", "acc": 0.66, "acc_std": 0.039359242878896944, "f1": 0.587178241864983, "f1_std": 0.0518995362724847, "bacc": 0.5933786078098472, "bacc_std": 0.043844258623806844}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 25, "C": 0.046415888336127774, "split": "test", "acc": 0.65, "acc_std": 0.03874296839427769, "f1": 0.5872154735228211, "f1_std": 0.04856378598021259, "bacc": 0.5904074702886248, "bacc_std": 0.04255879245264194}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 26, "C": 0.005994842503189409, "split": "test", "acc": 0.69, "acc_std": 0.0442093021885666, "f1": 0.6570417081535569, "f1_std": 0.04929526034755114, "bacc": 0.6532258064516129, "bacc_std": 0.04723120308719199}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 27, "C": 0.005994842503189409, "split": "test", "acc": 0.7, "acc_std": 0.04159736049318513, "f1": 0.6553308823529411, "f1_std": 0.05078575885749312, "bacc": 0.6511035653650254, "bacc_std": 0.04659728183366576}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 28, "C": 0.000774263682681127, "split": "test", "acc": 0.67, "acc_std": 0.03067796603427286, "f1": 0.553993782943641, "f1_std": 0.051359156581592735, "bacc": 0.5810696095076401, "bacc_std": 0.03623461740012693}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 29, "C": 0.005994842503189409, "split": "test", "acc": 0.66, "acc_std": 0.042535161925164926, "f1": 0.5952380952380952, "f1_std": 0.05439996012862267, "bacc": 0.5984719864176571, "bacc_std": 0.04721323439084075}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 30, "C": 0.005994842503189409, "split": "test", "acc": 0.67, "acc_std": 0.037467185642906235, "f1": 0.5862068965517242, "f1_std": 0.05314493459309705, "bacc": 0.5963497453310695, "bacc_std": 0.04261170123691426}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 31, "C": 0.000774263682681127, "split": "test", "acc": 0.68, "acc_std": 0.03890572708483931, "f1": 0.6114618746964546, "f1_std": 0.05213949409250786, "bacc": 0.6146010186757216, "bacc_std": 0.0438386360921965}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 32, "C": 0.046415888336127774, "split": "test", "acc": 0.58, "acc_std": 0.04585163464915945, "f1": 0.5384615384615385, "f1_std": 0.05025710180174014, "bacc": 0.5390492359932089, "bacc_std": 0.04839080570630248}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 33, "C": 0.005994842503189409, "split": "test", "acc": 0.7, "acc_std": 0.04169225827416883, "f1": 0.66078697421981, "f1_std": 0.04900294942351862, "bacc": 0.6561969439728353, "bacc_std": 0.04594011927019174}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 34, "C": 0.005994842503189409, "split": "test", "acc": 0.69, "acc_std": 0.037143241646361444, "f1": 0.627359057579036, "f1_std": 0.04924439057761353, "bacc": 0.6277589134125636, "bacc_std": 0.04238580014377241}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 35, "C": 0.3593813663804626, "split": "test", "acc": 0.52, "acc_std": 0.04902917906716367, "f1": 0.5104039167686658, "f1_std": 0.04935047167202559, "bacc": 0.5161290322580645, "bacc_std": 0.05097404658807995}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 36, "C": 0.3593813663804626, "split": "test", "acc": 0.64, "acc_std": 0.0467050318488276, "f1": 0.6043956043956044, "f1_std": 0.05175280086752021, "bacc": 0.6027164685908319, "bacc_std": 0.04994603542896979}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 37, "C": 0.005994842503189409, "split": "test", "acc": 0.64, "acc_std": 0.040771382120305906, "f1": 0.5792426367461431, "f1_std": 0.049714054848068435, "bacc": 0.5823429541595926, "bacc_std": 0.04407053423483324}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 38, "C": 0.000774263682681127, "split": "test", "acc": 0.6, "acc_std": 0.039408202191929545, "f1": 0.5238095238095238, "f1_std": 0.04776189835226844, "bacc": 0.5348047538200339, "bacc_std": 0.04162718831390681}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 39, "C": 0.046415888336127774, "split": "test", "acc": 0.69, "acc_std": 0.04189844388518504, "f1": 0.6521153630344518, "f1_std": 0.04886775212455069, "bacc": 0.6481324278438031, "bacc_std": 0.04623444912076014}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 40, "C": 0.005994842503189409, "split": "test", "acc": 0.65, "acc_std": 0.040883106535585086, "f1": 0.5944849959448499, "f1_std": 0.04888557286750138, "bacc": 0.5955008488964346, "bacc_std": 0.04405137536803788}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 41, "C": 2.782559402207126, "split": "test", "acc": 0.56, "acc_std": 0.04783058017628471, "f1": 0.537620849096259, "f1_std": 0.04993789794468509, "bacc": 0.5382003395585738, "bacc_std": 0.050065650091611735}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 42, "C": 0.005994842503189409, "split": "test", "acc": 0.7, "acc_std": 0.03861988606922604, "f1": 0.6428571428571428, "f1_std": 0.05049855592888827, "bacc": 0.6409168081494058, "bacc_std": 0.0441068985962559}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 43, "C": 0.3593813663804626, "split": "test", "acc": 0.63, "acc_std": 0.04438772803377078, "f1": 0.5960257670051315, "f1_std": 0.048085910881573234, "bacc": 0.5946519524617997, "bacc_std": 0.04668779065305063}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 44, "C": 0.005994842503189409, "split": "test", "acc": 0.7, "acc_std": 0.039695133202950705, "f1": 0.6493688639551192, "f1_std": 0.05053104375039742, "bacc": 0.6460101867572157, "bacc_std": 0.04523567723721884}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 45, "C": 0.005994842503189409, "split": "test", "acc": 0.77, "acc_std": 0.03939796441442122, "f1": 0.7335187116209014, "f1_std": 0.04950828575556397, "bacc": 0.7228353140916808, "bacc_std": 0.04632578261857936}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 46, "C": 0.005994842503189409, "split": "test", "acc": 0.62, "acc_std": 0.042296453752058226, "f1": 0.5634191176470589, "f1_std": 0.05003778234165454, "bacc": 0.566213921901528, "bacc_std": 0.04529927427406702}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 47, "C": 21.54434690031882, "split": "test", "acc": 0.58, "acc_std": 0.049160353945023626, "f1": 0.565936337329475, "f1_std": 0.050333300825817016, "bacc": 0.5696095076400679, "bacc_std": 0.051501506448355466}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 48, "C": 0.005994842503189409, "split": "test", "acc": 0.69, "acc_std": 0.042002338030162074, "f1": 0.6521153630344518, "f1_std": 0.04890805221489297, "bacc": 0.6481324278438031, "bacc_std": 0.04612783881202852}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 49, "C": 0.046415888336127774, "split": "test", "acc": 0.64, "acc_std": 0.04315617684642605, "f1": 0.592944369063772, "f1_std": 0.05008514708316441, "bacc": 0.5925297113752122, "bacc_std": 0.04653431799624525}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 50, "C": 0.005994842503189409, "split": "test", "acc": 0.65, "acc_std": 0.04186960233868958, "f1": 0.5872154735228211, "f1_std": 0.052089441479292606, "bacc": 0.5904074702886248, "bacc_std": 0.0457184701150498}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 51, "C": 0.046415888336127774, "split": "test", "acc": 0.58, "acc_std": 0.04394858814569587, "f1": 0.525101763907734, "f1_std": 0.04925723519474069, "bacc": 0.5288624787775891, "bacc_std": 0.04576065646434931}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 52, "C": 0.3593813663804626, "split": "test", "acc": 0.64, "acc_std": 0.04594526743855128, "f1": 0.6043956043956044, "f1_std": 0.05108141746199053, "bacc": 0.6027164685908319, "bacc_std": 0.04921371355729635}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 53, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.046641423648941076, "f1": 0.5755517826825127, "f1_std": 0.04922807942378391, "bacc": 0.5755517826825127, "bacc_std": 0.04909963904002723}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 54, "C": 0.3593813663804626, "split": "test", "acc": 0.5, "acc_std": 0.04654825882887565, "f1": 0.47456914670029426, "f1_std": 0.04571767923692783, "bacc": 0.47453310696095075, "bacc_std": 0.045906918194169734}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 55, "C": 0.046415888336127774, "split": "test", "acc": 0.63, "acc_std": 0.04858992076552502, "f1": 0.5960257670051315, "f1_std": 0.053327716404404285, "bacc": 0.5946519524617997, "bacc_std": 0.05170670299797897}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 56, "C": 0.005994842503189409, "split": "test", "acc": 0.61, "acc_std": 0.04263674940705495, "f1": 0.5481404240528328, "f1_std": 0.050460640504954794, "bacc": 0.5530560271646858, "bacc_std": 0.045366017278115796}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 57, "C": 166.81005372000556, "split": "test", "acc": 0.59, "acc_std": 0.047769714254954465, "f1": 0.5710848415106182, "f1_std": 0.04885226079628632, "bacc": 0.5725806451612903, "bacc_std": 0.04926112876344077}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 58, "C": 0.046415888336127774, "split": "test", "acc": 0.64, "acc_std": 0.04806986165987999, "f1": 0.6216897856242118, "f1_std": 0.05020571911418118, "bacc": 0.6230899830220713, "bacc_std": 0.05049275220263842}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 59, "C": 0.005994842503189409, "split": "test", "acc": 0.71, "acc_std": 0.038603191577899355, "f1": 0.6514004087029691, "f1_std": 0.05191555313728436, "bacc": 0.648981324278438, "bacc_std": 0.04456835347795613}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 60, "C": 0.3593813663804626, "split": "test", "acc": 0.56, "acc_std": 0.04742544043021635, "f1": 0.5225694444444444, "f1_std": 0.051029139827030244, "bacc": 0.5229202037351443, "bacc_std": 0.049639849423037244}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 61, "C": 0.005994842503189409, "split": "test", "acc": 0.71, "acc_std": 0.040202666578225875, "f1": 0.6695156695156695, "f1_std": 0.04840326159540572, "bacc": 0.6642614601018676, "bacc_std": 0.04503985354922326}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 62, "C": 0.005994842503189409, "split": "test", "acc": 0.66, "acc_std": 0.036247460600709665, "f1": 0.5783730158730158, "f1_std": 0.05049094614126628, "bacc": 0.5882852292020373, "bacc_std": 0.04075973323452815}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 63, "C": 21.54434690031882, "split": "test", "acc": 0.55, "acc_std": 0.04771942581381297, "f1": 0.5248653785239151, "f1_std": 0.04885202447274437, "bacc": 0.5250424448217317, "bacc_std": 0.04900030924280954}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 64, "C": 0.046415888336127774, "split": "test", "acc": 0.63, "acc_std": 0.042978743583311045, "f1": 0.5713127099988413, "f1_std": 0.05176255931612501, "bacc": 0.5742784380305602, "bacc_std": 0.046450711647883274}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 65, "C": 0.3593813663804626, "split": "test", "acc": 0.58, "acc_std": 0.046169682693299935, "f1": 0.565936337329475, "f1_std": 0.0469895086396627, "bacc": 0.5696095076400679, "bacc_std": 0.04818929973274407}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 66, "C": 0.005994842503189409, "split": "test", "acc": 0.62, "acc_std": 0.043131686727972976, "f1": 0.5558672276764843, "f1_std": 0.0523967815617926, "bacc": 0.5611205432937181, "bacc_std": 0.04641474866162193}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 67, "C": 0.046415888336127774, "split": "test", "acc": 0.63, "acc_std": 0.04479919195699852, "f1": 0.5847828526540231, "f1_std": 0.05143843886985339, "bacc": 0.5844651952461799, "bacc_std": 0.04847250386261593}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 68, "C": 0.005994842503189409, "split": "test", "acc": 0.65, "acc_std": 0.04474319613080853, "f1": 0.6011396011396011, "f1_std": 0.05305899889450785, "bacc": 0.6005942275042444, "bacc_std": 0.048770019244813355}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 69, "C": 0.005994842503189409, "split": "test", "acc": 0.65, "acc_std": 0.04067628301602791, "f1": 0.5944849959448499, "f1_std": 0.04928671416219566, "bacc": 0.5955008488964346, "bacc_std": 0.0444172881789824}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 70, "C": 0.000774263682681127, "split": "test", "acc": 0.67, "acc_std": 0.03445126412775008, "f1": 0.5764343473238351, "f1_std": 0.050888489064860036, "bacc": 0.5912563667232598, "bacc_std": 0.03929187207403891}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 71, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.04122567646503814, "f1": 0.5143273433705683, "f1_std": 0.04924436446425617, "bacc": 0.5297113752122241, "bacc_std": 0.04259209948312931}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 72, "C": 0.005994842503189409, "split": "test", "acc": 0.68, "acc_std": 0.04190751722543343, "f1": 0.6323529411764706, "f1_std": 0.049305480374412375, "bacc": 0.6298811544991512, "bacc_std": 0.045275786603852244}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 73, "C": 1291.5496650148827, "split": "test", "acc": 0.57, "acc_std": 0.05017910322036454, "f1": 0.5664885573142454, "f1_std": 0.04985104033358967, "bacc": 0.581918505942275, "bacc_std": 0.05150845717111757}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 74, "C": 0.005994842503189409, "split": "test", "acc": 0.57, "acc_std": 0.043248024232327646, "f1": 0.50997150997151, "f1_std": 0.04921508989038315, "bacc": 0.515704584040747, "bacc_std": 0.04522025283875993}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 75, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.046275743105864874, "f1": 0.554367201426025, "f1_std": 0.05114936117160911, "bacc": 0.5551782682512734, "bacc_std": 0.04860033491428074}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 76, "C": 0.000774263682681127, "split": "test", "acc": 0.69, "acc_std": 0.035506866941480475, "f1": 0.6112852664576802, "f1_std": 0.05146837049809206, "bacc": 0.6175721561969439, "bacc_std": 0.041218261308588126}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 77, "C": 0.005994842503189409, "split": "test", "acc": 0.64, "acc_std": 0.04395857595509664, "f1": 0.5863970588235294, "f1_std": 0.051798751158720534, "bacc": 0.5874363327674024, "bacc_std": 0.04739767414166249}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 78, "C": 0.005994842503189409, "split": "test", "acc": 0.67, "acc_std": 0.03937946673077226, "f1": 0.6033177064551027, "f1_std": 0.049849766014654365, "bacc": 0.6065365025466893, "bacc_std": 0.04306558442341393}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 79, "C": 0.005994842503189409, "split": "test", "acc": 0.67, "acc_std": 0.04121473522904156, "f1": 0.6108031607500884, "f1_std": 0.051745442356282666, "bacc": 0.6116298811544991, "bacc_std": 0.04574801539268233}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 80, "C": 0.005994842503189409, "split": "test", "acc": 0.66, "acc_std": 0.0399095777978169, "f1": 0.6026180458158018, "f1_std": 0.049791065432573195, "bacc": 0.6035653650254669, "bacc_std": 0.044132056201035434}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 81, "C": 0.005994842503189409, "split": "test", "acc": 0.64, "acc_std": 0.04566519462347665, "f1": 0.5989304812834224, "f1_std": 0.05039282945270579, "bacc": 0.597623089983022, "bacc_std": 0.04779268998359048}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 82, "C": 0.005994842503189409, "split": "test", "acc": 0.65, "acc_std": 0.039005763676667075, "f1": 0.5792763553311696, "f1_std": 0.050154649859558226, "bacc": 0.5853140916808149, "bacc_std": 0.04275125937926583}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 83, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.049621995123130626, "f1": 0.5796553173602353, "f1_std": 0.05164147892762746, "bacc": 0.5806451612903225, "bacc_std": 0.052138683193298986}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 84, "C": 0.005994842503189409, "split": "test", "acc": 0.63, "acc_std": 0.04225444828654139, "f1": 0.5783475783475784, "f1_std": 0.04927786317854687, "bacc": 0.5793718166383701, "bacc_std": 0.04561249419529883}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 85, "C": 0.005994842503189409, "split": "test", "acc": 0.64, "acc_std": 0.03957314240744599, "f1": 0.5714285714285714, "f1_std": 0.0492308765102415, "bacc": 0.5772495755517827, "bacc_std": 0.04288495908842678}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 86, "C": 166.81005372000556, "split": "test", "acc": 0.56, "acc_std": 0.04400586779055719, "f1": 0.5098039215686274, "f1_std": 0.04914286928388947, "bacc": 0.5127334465195246, "bacc_std": 0.046033859951775294}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 87, "C": 0.005994842503189409, "split": "test", "acc": 0.61, "acc_std": 0.04589185548656755, "f1": 0.5555555555555556, "f1_std": 0.05325900343796785, "bacc": 0.5581494057724957, "bacc_std": 0.048876738658343286}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 88, "C": 0.046415888336127774, "split": "test", "acc": 0.68, "acc_std": 0.044693247812169566, "f1": 0.64349376114082, "f1_std": 0.050921199274246945, "bacc": 0.6400679117147707, "bacc_std": 0.04834830926838689}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 89, "C": 0.005994842503189409, "split": "test", "acc": 0.74, "acc_std": 0.03629377908126958, "f1": 0.6843127731908694, "f1_std": 0.05108378237636309, "bacc": 0.6782682512733447, "bacc_std": 0.04379747710619452}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 90, "C": 0.005994842503189409, "split": "test", "acc": 0.59, "acc_std": 0.04499066569856462, "f1": 0.5523528769516323, "f1_std": 0.04785661818701442, "bacc": 0.5522071307300509, "bacc_std": 0.04626587123696536}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 91, "C": 0.000774263682681127, "split": "test", "acc": 0.67, "acc_std": 0.03502371196775121, "f1": 0.5764343473238351, "f1_std": 0.05233653378290198, "bacc": 0.5912563667232598, "bacc_std": 0.04002633085863833}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 92, "C": 1291.5496650148827, "split": "test", "acc": 0.63, "acc_std": 0.05081862257086471, "f1": 0.6009060511271707, "f1_std": 0.053810753861437015, "bacc": 0.5997453310696095, "bacc_std": 0.05276452961869997}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 93, "C": 0.005994842503189409, "split": "test", "acc": 0.59, "acc_std": 0.04602699642601068, "f1": 0.5577607593571352, "f1_std": 0.04998317101197874, "bacc": 0.5573005093378608, "bacc_std": 0.0490519133303839}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 94, "C": 0.005994842503189409, "split": "test", "acc": 0.69, "acc_std": 0.03923965341335215, "f1": 0.6343908479773559, "f1_std": 0.050090822011478915, "bacc": 0.6328522920203735, "bacc_std": 0.04416126339782681}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 95, "C": 0.005994842503189409, "split": "test", "acc": 0.66, "acc_std": 0.03705590911042393, "f1": 0.587178241864983, "f1_std": 0.04962565319413015, "bacc": 0.5933786078098472, "bacc_std": 0.041286367439623155}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 96, "C": 0.046415888336127774, "split": "test", "acc": 0.56, "acc_std": 0.044387457687955045, "f1": 0.5024875621890548, "f1_std": 0.048658932682766505, "bacc": 0.5076400679117148, "bacc_std": 0.045479593773532924}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 97, "C": 21.54434690031882, "split": "test", "acc": 0.56, "acc_std": 0.04806538879484904, "f1": 0.5452666391070691, "f1_std": 0.04831039666654839, "bacc": 0.5483870967741935, "bacc_std": 0.04918423056231293}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 98, "C": 0.3593813663804626, "split": "test", "acc": 0.66, "acc_std": 0.04396021383023517, "f1": 0.6263736263736264, "f1_std": 0.04962542459903583, "bacc": 0.6239388794567062, "bacc_std": 0.048158213756542334}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 99, "C": 166.81005372000556, "split": "test", "acc": 0.55, "acc_std": 0.04653043735019047, "f1": 0.529239460194581, "f1_std": 0.048117163702210176, "bacc": 0.5301358234295416, "bacc_std": 0.04879885908818956}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 100, "C": 0.3593813663804626, "split": "test", "acc": 0.62, "acc_std": 0.04733426243219598, "f1": 0.5876736111111112, "f1_std": 0.05216983789690622, "bacc": 0.5865874363327674, "bacc_std": 0.05078974637886583}
+eval results (random splits):
+
+| model    | repr   | clf      | dataset   | split   |   n_trials |      C |   C_std |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:----------|:--------|-----------:|-------:|--------:|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | ppmi_dx   | train   |        100 | 31.834 |  183.13 | 0.79861 |  0.10142  | 0.76515 | 0.1232   | 0.76035 |   0.12253  |
+| flat_mae | patch  | logistic | ppmi_dx   | test    |        100 | 31.834 |  183.13 | 0.6326  |  0.052486 | 0.58374 | 0.048046 | 0.58595 |   0.045347 |
+
+
+done! total time: 0:05:11

data_scaling/n200_2/pretrain/config.yaml

@@ -0,0 +1,109 @@
+name: data_scaling/n200_2/pretrain
+notes: data scaling experiment n200_2 (seed=3472)
+output_dir: experiments/data_scaling/output/data_scaling/n200_2/pretrain
+input_space: flat
+patch_size: 16
+num_frames: 16
+t_patch_size: 4
+mask_ratio: 0.9
+pred_mask_ratio: null
+masking: tube
+masking_kwargs: {}
+mask_patch_size: null
+model: mae_vit_base
+model_kwargs:
+  decoding: attn
+  pos_embed: sep
+  target_norm: null
+  pca_norm_nc: 2
+  t_pred_stride: 2
+  no_decode_pos: true
+  mask_drop_scale: false
+  pred_edge_pad: 0
+  gauss_sigma: null
+  class_token: true
+  reg_tokens: 0
+  no_embed_class: true
+  head_init_scale: 0.0
+  decoder_depth: 4
+  drop_path_rate: 0.0
+datasets:
+  hcp-train:
+    type: wds
+    url: /data/fmri-datasets/pretrain/hcpya-all.flat.wds/hcpya-all-flat-{00800..00999}.tar
+    clipping: random
+    clipping_kwargs:
+      oversample: 4.0
+    shuffle: true
+    buffer_size: 2000
+    samples_per_epoch: 200000
+  hcp-train-subset:
+    type: arrow
+    root: s3://medarc/fmri-datasets/eval/hcpya-clips.${input_space}.arrow/train
+    split_range:
+    - 0
+    - 2000
+    shuffle: false
+  hcp-val:
+    type: arrow
+    root: s3://medarc/fmri-datasets/eval/hcpya-clips.${input_space}.arrow/test
+    split_range:
+    - 0
+    - 2000
+    shuffle: false
+  nsd-val:
+    type: arrow
+    root: s3://medarc/fmri-datasets/eval/nsd-cococlip.${input_space}.arrow/testid
+    split_range:
+    - 0
+    - 2000
+    shuffle: false
+train_dataset: hcp-train
+eval_datasets:
+- hcp-train-subset
+- hcp-val
+- nsd-val
+val_dataset: hcp-val
+clip_vmax: 3.0
+normalize: frame
+tr_scale: null
+crop_scale: null
+crop_aspect: null
+gray_jitter: null
+num_workers: 16
+epochs: 100
+batch_size: 32
+accum_iter: 1
+base_lr: 0.001
+min_lr: 0.0
+warmup_epochs: 5
+weight_decay: 0.05
+betas:
+- 0.9
+- 0.95
+clip_grad: 1.0
+amp: true
+amp_dtype: float16
+ckpt: null
+resume: true
+auto_resume: true
+start_epoch: 0
+max_checkpoints: 20
+checkpoint_period: 5
+plot_period: 5
+device: cuda
+presend_cuda: false
+seed: 3472
+debug: false
+wandb: true
+wandb_entity: null
+wandb_project: fMRI-foundation-model
+rank: 0
+world_size: 1
+gpu: 0
+distributed: true
+dist_backend: nccl
+in_chans: 1
+img_size:
+- 224
+- 560

data_scaling/n200_2/pretrain/log.json

@@ -0,0 +1,100 @@
+{"epoch": 0, "train/lr": 1.2502400076802458e-05, "train/grad": 0.06074886535599828, "train/loss": 0.9929967404460907, "eval/hcp-train-subset/loss": 0.9907431679387246, "eval/hcp-val/loss": 0.9907587920465777, "eval/nsd-val/loss": 0.9919356838349374}
+{"epoch": 1, "train/lr": 3.750320010240327e-05, "train/grad": 0.09351523154258728, "train/loss": 0.986495768699646, "eval/hcp-train-subset/loss": 0.9877115382302192, "eval/hcp-val/loss": 0.9871609307104542, "eval/nsd-val/loss": 0.9874657959707321}
+{"epoch": 2, "train/lr": 6.250400012800409e-05, "train/grad": 0.14525675498695612, "train/loss": 0.9820738704299927, "eval/hcp-train-subset/loss": 0.9808463252359821, "eval/hcp-val/loss": 0.980232952102538, "eval/nsd-val/loss": 0.9813966395393494}
+{"epoch": 3, "train/lr": 8.75048001536049e-05, "train/grad": 0.22075390657766333, "train/loss": 0.9706496034145355, "eval/hcp-train-subset/loss": 0.9609758171343035, "eval/hcp-val/loss": 0.9621732955978762, "eval/nsd-val/loss": 0.946446277441517}
+{"epoch": 4, "train/lr": 0.00011250559953918529, "train/grad": 0.25123899827846036, "train/loss": 0.9345017951488495, "eval/hcp-train-subset/loss": 0.9174794506642127, "eval/hcp-val/loss": 0.9159393512433575, "eval/nsd-val/loss": 0.884861292377595}
+{"epoch": 5, "train/lr": 0.00012498860637884563, "train/grad": 0.17747241761091137, "train/loss": 0.8922654366493226, "eval/hcp-train-subset/loss": 0.8803248453524805, "eval/hcp-val/loss": 0.8790773566692106, "eval/nsd-val/loss": 0.8424928832438684}
+{"epoch": 6, "train/lr": 0.0001249202705377922, "train/grad": 0.12238330470347984, "train/loss": 0.867705241613388, "eval/hcp-train-subset/loss": 0.8673066727576717, "eval/hcp-val/loss": 0.8655408351652084, "eval/nsd-val/loss": 0.8303181577113367}
+{"epoch": 7, "train/lr": 0.0001247836790473516, "train/grad": 0.1037894368042884, "train/loss": 0.8538292635154724, "eval/hcp-train-subset/loss": 0.8628166587122025, "eval/hcp-val/loss": 0.8607290762086068, "eval/nsd-val/loss": 0.8269688756235184}
+{"epoch": 8, "train/lr": 0.000124578981268311, "train/grad": 0.09880859165890059, "train/loss": 0.8468218855571746, "eval/hcp-train-subset/loss": 0.8601642731697329, "eval/hcp-val/loss": 0.8583529245468878, "eval/nsd-val/loss": 0.8248861797394291}
+{"epoch": 9, "train/lr": 0.00012430640103468907, "train/grad": 0.10141315121380771, "train/loss": 0.8383862003993988, "eval/hcp-train-subset/loss": 0.858305658063581, "eval/hcp-val/loss": 0.8568743332739799, "eval/nsd-val/loss": 0.8243883915485875}
+{"epoch": 10, "train/lr": 0.00012396623640896796, "train/grad": 0.10129307399125066, "train/loss": 0.8328803121948242, "eval/hcp-train-subset/loss": 0.8574367694316372, "eval/hcp-val/loss": 0.8551280931118996, "eval/nsd-val/loss": 0.825301754859186}
+{"epoch": 11, "train/lr": 0.0001235588593561712, "train/grad": 0.09790804282322087, "train/loss": 0.8301758789920807, "eval/hcp-train-subset/loss": 0.857351605930636, "eval/hcp-val/loss": 0.8552730717966633, "eval/nsd-val/loss": 0.8271028716717997}
+{"epoch": 12, "train/lr": 0.00012308471533712604, "train/grad": 0.10205107842358581, "train/loss": 0.8263192761421203, "eval/hcp-train-subset/loss": 0.855712209017046, "eval/hcp-val/loss": 0.8536175806676188, "eval/nsd-val/loss": 0.8211562056695262}
+{"epoch": 13, "train/lr": 0.00012254432282135565, "train/grad": 0.10727709414742687, "train/loss": 0.8184426688480377, "eval/hcp-train-subset/loss": 0.8572292212517031, "eval/hcp-val/loss": 0.8547284324322978, "eval/nsd-val/loss": 0.8278160325942501}
+{"epoch": 14, "train/lr": 0.00012193827272014171, "train/grad": 0.11076085550755872, "train/loss": 0.8117992891883851, "eval/hcp-train-subset/loss": 0.8562669840551191, "eval/hcp-val/loss": 0.8550113699128551, "eval/nsd-val/loss": 0.8245106733614399}
+{"epoch": 15, "train/lr": 0.00012126722774037197, "train/grad": 0.1141433127931696, "train/loss": 0.8045218962097168, "eval/hcp-train-subset/loss": 0.8558692653332988, "eval/hcp-val/loss": 0.854046288036531, "eval/nsd-val/loss": 0.8244515176742307}
+{"epoch": 16, "train/lr": 0.00012053192165988122, "train/grad": 0.11490780322975866, "train/loss": 0.8031361094474793, "eval/hcp-train-subset/loss": 0.8552336000627087, "eval/hcp-val/loss": 0.8533100553097264, "eval/nsd-val/loss": 0.8275022872032658}
+{"epoch": 17, "train/lr": 0.00011973315852507104, "train/grad": 0.11752134790218702, "train/loss": 0.7948549908924103, "eval/hcp-train-subset/loss": 0.8574261367321014, "eval/hcp-val/loss": 0.8549363670810577, "eval/nsd-val/loss": 0.8386709565116514}
+{"epoch": 18, "train/lr": 0.00011887181177170142, "train/grad": 0.11806112254488386, "train/loss": 0.7936267003154754, "eval/hcp-train-subset/loss": 0.8547361589247181, "eval/hcp-val/loss": 0.853979435659224, "eval/nsd-val/loss": 0.8342916850120791}
+{"epoch": 19, "train/lr": 0.00011794882326980209, "train/grad": 0.12073047730651959, "train/loss": 0.7897758152580261, "eval/hcp-train-subset/loss": 0.8555082063521108, "eval/hcp-val/loss": 0.8539005767914557, "eval/nsd-val/loss": 0.8277239213066716}
+{"epoch": 20, "train/lr": 0.00011696520229374954, "train/grad": 0.11888711814780092, "train/loss": 0.7894636356163025, "eval/hcp-train-subset/loss": 0.8570529222488403, "eval/hcp-val/loss": 0.854197072405969, "eval/nsd-val/loss": 0.8252206761990825}
+{"epoch": 21, "train/lr": 0.00011592202441863837, "train/grad": 0.12250903351859849, "train/loss": 0.7834287502002716, "eval/hcp-train-subset/loss": 0.8577828772606388, "eval/hcp-val/loss": 0.8554161806260386, "eval/nsd-val/loss": 0.8294835311751212}
+{"epoch": 22, "train/lr": 0.00011482043034415979, "train/grad": 0.12603224357551313, "train/loss": 0.7777083412647248, "eval/hcp-train-subset/loss": 0.8552338903950106, "eval/hcp-val/loss": 0.8539669282974736, "eval/nsd-val/loss": 0.8302801559048314}
+{"epoch": 23, "train/lr": 0.00011366162464726024, "train/grad": 0.1259937200483016, "train/loss": 0.777787874212265, "eval/hcp-train-subset/loss": 0.8568591398577536, "eval/hcp-val/loss": 0.8549495243257091, "eval/nsd-val/loss": 0.8314102638152338}
+{"epoch": 24, "train/lr": 0.0001124468744649569, "train/grad": 0.12925501137246853, "train/loss": 0.7733322998523712, "eval/hcp-train-subset/loss": 0.8569119726457903, "eval/hcp-val/loss": 0.8549876761051917, "eval/nsd-val/loss": 0.8321533443466309}
+{"epoch": 25, "train/lr": 0.0001111775081087387, "train/grad": 0.1301940694759089, "train/loss": 0.7696429798603058, "eval/hcp-train-subset/loss": 0.8567405496874163, "eval/hcp-val/loss": 0.8540384827121612, "eval/nsd-val/loss": 0.8328149049512802}
+{"epoch": 26, "train/lr": 0.0001098549136120796, "train/grad": 0.1288295753405128, "train/loss": 0.7737991156959534, "eval/hcp-train-subset/loss": 0.8570112778294471, "eval/hcp-val/loss": 0.8556579293743256, "eval/nsd-val/loss": 0.8308307278540826}
+{"epoch": 27, "train/lr": 0.00010848053721264312, "train/grad": 0.13354393598598532, "train/loss": 0.7642915752506256, "eval/hcp-train-subset/loss": 0.8561610098808042, "eval/hcp-val/loss": 0.8544871143756374, "eval/nsd-val/loss": 0.8286346418242301}
+{"epoch": 28, "train/lr": 0.00010705588177084458, "train/grad": 0.13351944186257347, "train/loss": 0.7627652764129639, "eval/hcp-train-subset/loss": 0.8574909615901208, "eval/hcp-val/loss": 0.8555521897731289, "eval/nsd-val/loss": 0.832959400069329}
+{"epoch": 29, "train/lr": 0.00010558250512649171, "train/grad": 0.13470303897802707, "train/loss": 0.7668015116500855, "eval/hcp-train-subset/loss": 0.856824601850202, "eval/hcp-val/loss": 0.8546492236275827, "eval/nsd-val/loss": 0.8305300464553218}
+{"epoch": 30, "train/lr": 0.00010406201839531515, "train/grad": 0.1373441127124874, "train/loss": 0.7615819299697876, "eval/hcp-train-subset/loss": 0.8593883033721678, "eval/hcp-val/loss": 0.8568122329250458, "eval/nsd-val/loss": 0.8302586838122337}
+{"epoch": 31, "train/lr": 0.00010249608420723018, "train/grad": 0.1374305129592592, "train/loss": 0.7591948459815979, "eval/hcp-train-subset/loss": 0.8580583333969116, "eval/hcp-val/loss": 0.8546640949864541, "eval/nsd-val/loss": 0.8360447777855781}
+{"epoch": 32, "train/lr": 0.00010088641488828097, "train/grad": 0.13892441383690318, "train/loss": 0.7577868027496338, "eval/hcp-train-subset/loss": 0.8576850814204062, "eval/hcp-val/loss": 0.855865998614219, "eval/nsd-val/loss": 0.8296882241002975}
+{"epoch": 33, "train/lr": 9.923477058823526e-05, "train/grad": 0.1419349339823281, "train/loss": 0.7534558692550659, "eval/hcp-train-subset/loss": 0.8592259355129734, "eval/hcp-val/loss": 0.855732322700562, "eval/nsd-val/loss": 0.8300859274402741}
+{"epoch": 34, "train/lr": 9.754295735588547e-05, "train/grad": 0.143895031402774, "train/loss": 0.7519091509437561, "eval/hcp-train-subset/loss": 0.856386472140589, "eval/hcp-val/loss": 0.8546181513417151, "eval/nsd-val/loss": 0.830078270166151}
+{"epoch": 35, "train/lr": 9.581282516416285e-05, "train/grad": 0.14372638611673888, "train/loss": 0.7531642698287964, "eval/hcp-train-subset/loss": 0.8588362432295277, "eval/hcp-val/loss": 0.8562913344752404, "eval/nsd-val/loss": 0.8313780330842541}
+{"epoch": 36, "train/lr": 9.404626588721676e-05, "train/grad": 0.14384142661087987, "train/loss": 0.750057115983963, "eval/hcp-train-subset/loss": 0.8600366577025382, "eval/hcp-val/loss": 0.8589841233145806, "eval/nsd-val/loss": 0.8332039425449986}
+{"epoch": 37, "train/lr": 9.224521123168153e-05, "train/grad": 0.14616839765963383, "train/loss": 0.7520381041717529, "eval/hcp-train-subset/loss": 0.857975838645812, "eval/hcp-val/loss": 0.8556927115686478, "eval/nsd-val/loss": 0.8268031708655819}
+{"epoch": 38, "train/lr": 9.041163062437843e-05, "train/grad": 0.1505200284176653, "train/loss": 0.7466013880825043, "eval/hcp-train-subset/loss": 0.8612049894948159, "eval/hcp-val/loss": 0.8601584280690839, "eval/nsd-val/loss": 0.8306797665934409}
+{"epoch": 39, "train/lr": 8.85475290587822e-05, "train/grad": 0.15134367883548958, "train/loss": 0.7425296100139618, "eval/hcp-train-subset/loss": 0.862225366215552, "eval/hcp-val/loss": 0.8605233094384593, "eval/nsd-val/loss": 0.8346813744114291}
+{"epoch": 40, "train/lr": 8.665494490258622e-05, "train/grad": 0.15440325881026581, "train/loss": 0.7406586199569702, "eval/hcp-train-subset/loss": 0.8603614318755365, "eval/hcp-val/loss": 0.8590858299886027, "eval/nsd-val/loss": 0.8317492700392201}
+{"epoch": 41, "train/lr": 8.473594766877838e-05, "train/grad": 0.15227809418636767, "train/loss": 0.7426287366676331, "eval/hcp-train-subset/loss": 0.8601657411744518, "eval/hcp-val/loss": 0.8584105603156551, "eval/nsd-val/loss": 0.8297559445904147}
+{"epoch": 42, "train/lr": 8.279263575265999e-05, "train/grad": 0.1562015736516131, "train/loss": 0.741456170244217, "eval/hcp-train-subset/loss": 0.8620805855720274, "eval/hcp-val/loss": 0.8601814921825163, "eval/nsd-val/loss": 0.8320642584754575}
+{"epoch": 43, "train/lr": 8.082713413727944e-05, "train/grad": 0.15440719326138402, "train/loss": 0.7408219782924652, "eval/hcp-train-subset/loss": 0.8630445263078136, "eval/hcp-val/loss": 0.8605247668681606, "eval/nsd-val/loss": 0.8352487500636808}
+{"epoch": 44, "train/lr": 7.884159206979602e-05, "train/grad": 0.15839646718966824, "train/loss": 0.7380263511753082, "eval/hcp-train-subset/loss": 0.8628127123078992, "eval/hcp-val/loss": 0.8615153076187256, "eval/nsd-val/loss": 0.8343590786380153}
+{"epoch": 45, "train/lr": 7.683818071130916e-05, "train/grad": 0.15937190388762196, "train/loss": 0.7351642228794097, "eval/hcp-train-subset/loss": 0.8627195829345334, "eval/hcp-val/loss": 0.8612916882961027, "eval/nsd-val/loss": 0.8365419987709292}
+{"epoch": 46, "train/lr": 7.481909076272522e-05, "train/grad": 0.16109139839289818, "train/loss": 0.7339894597244263, "eval/hcp-train-subset/loss": 0.8622432293430451, "eval/hcp-val/loss": 0.8592312259058799, "eval/nsd-val/loss": 0.8326808339165103}
+{"epoch": 47, "train/lr": 7.278653006925963e-05, "train/grad": 0.16124788359661377, "train/loss": 0.7399765062332153, "eval/hcp-train-subset/loss": 0.8648987047133907, "eval/hcp-val/loss": 0.8632527647479888, "eval/nsd-val/loss": 0.8369248759362006}
+{"epoch": 48, "train/lr": 7.074272120618864e-05, "train/grad": 0.1616229197050942, "train/loss": 0.735530675201416, "eval/hcp-train-subset/loss": 0.8628119037997338, "eval/hcp-val/loss": 0.8614514393191184, "eval/nsd-val/loss": 0.8336973747899455}
+{"epoch": 49, "train/lr": 6.868989904849677e-05, "train/grad": 0.16531834598944933, "train/loss": 0.7309800907564163, "eval/hcp-train-subset/loss": 0.8645809100520226, "eval/hcp-val/loss": 0.862122526091914, "eval/nsd-val/loss": 0.8330993479298007}
+{"epoch": 50, "train/lr": 6.6630308327075e-05, "train/grad": 0.16626524612088134, "train/loss": 0.7342979122543335, "eval/hcp-train-subset/loss": 0.863880826580909, "eval/hcp-val/loss": 0.862056827352893, "eval/nsd-val/loss": 0.8326942872616553}
+{"epoch": 51, "train/lr": 6.456620117413798e-05, "train/grad": 0.16825580714539462, "train/loss": 0.7308739564037323, "eval/hcp-train-subset/loss": 0.8655327645040327, "eval/hcp-val/loss": 0.8624400054254839, "eval/nsd-val/loss": 0.834556172932348}
+{"epoch": 52, "train/lr": 6.249983466055255e-05, "train/grad": 0.16845951939458026, "train/loss": 0.7318064731693268, "eval/hcp-train-subset/loss": 0.8651785216023845, "eval/hcp-val/loss": 0.8621255361264751, "eval/nsd-val/loss": 0.833506760097319}
+{"epoch": 53, "train/lr": 6.0433468327763305e-05, "train/grad": 0.1715095791341859, "train/loss": 0.7311238735961914, "eval/hcp-train-subset/loss": 0.8672118927201917, "eval/hcp-val/loss": 0.8654560456352849, "eval/nsd-val/loss": 0.8398959213687528}
+{"epoch": 54, "train/lr": 5.83693617170174e-05, "train/grad": 0.17294580892269698, "train/loss": 0.7287598222446442, "eval/hcp-train-subset/loss": 0.8661532527016055, "eval/hcp-val/loss": 0.8639100501614232, "eval/nsd-val/loss": 0.837516192466982}
+{"epoch": 55, "train/lr": 5.6309771898588165e-05, "train/grad": 0.17358474203081534, "train/loss": 0.7297283543205261, "eval/hcp-train-subset/loss": 0.8651725111469146, "eval/hcp-val/loss": 0.8636563747159897, "eval/nsd-val/loss": 0.8346559107303619}
+{"epoch": 56, "train/lr": 5.4256951003704155e-05, "train/grad": 0.17487595142217094, "train/loss": 0.7304398733901978, "eval/hcp-train-subset/loss": 0.868542968265472, "eval/hcp-val/loss": 0.8662177362749653, "eval/nsd-val/loss": 0.8387845687327846}
+{"epoch": 57, "train/lr": 5.221314376187425e-05, "train/grad": 0.1788934983346473, "train/loss": 0.7268931568813324, "eval/hcp-train-subset/loss": 0.8664373063272045, "eval/hcp-val/loss": 0.864177510623009, "eval/nsd-val/loss": 0.8342412960144782}
+{"epoch": 58, "train/lr": 5.018058504631059e-05, "train/grad": 0.1820407150501401, "train/loss": 0.724999898147583, "eval/hcp-train-subset/loss": 0.8662707978679288, "eval/hcp-val/loss": 0.8643075093146293, "eval/nsd-val/loss": 0.8361720519681131}
+{"epoch": 59, "train/lr": 4.816149743012713e-05, "train/grad": 0.18057002734140076, "train/loss": 0.7259486714363098, "eval/hcp-train-subset/loss": 0.8676299289349587, "eval/hcp-val/loss": 0.864532700469417, "eval/nsd-val/loss": 0.8378282564301645}
+{"epoch": 60, "train/lr": 4.615808875598772e-05, "train/grad": 0.1837256980408013, "train/loss": 0.7241565581226349, "eval/hcp-train-subset/loss": 0.8668681950338425, "eval/hcp-val/loss": 0.8644066981730922, "eval/nsd-val/loss": 0.8358535439737381}
+{"epoch": 61, "train/lr": 4.417254972186445e-05, "train/grad": 0.18283198898832256, "train/loss": 0.7249583155345917, "eval/hcp-train-subset/loss": 0.8678699529940083, "eval/hcp-val/loss": 0.8661888828200679, "eval/nsd-val/loss": 0.8391736457424779}
+{"epoch": 62, "train/lr": 4.220705148553925e-05, "train/grad": 0.18456242503534226, "train/loss": 0.7249735138511658, "eval/hcp-train-subset/loss": 0.871081055172028, "eval/hcp-val/loss": 0.8678340171614001, "eval/nsd-val/loss": 0.8378261837267107}
+{"epoch": 63, "train/lr": 4.026374329047657e-05, "train/grad": 0.18991707668497942, "train/loss": 0.7197164055347443, "eval/hcp-train-subset/loss": 0.8694705203656228, "eval/hcp-val/loss": 0.8660907774202286, "eval/nsd-val/loss": 0.8371644452694924}
+{"epoch": 64, "train/lr": 3.834475011565652e-05, "train/grad": 0.18921554228589035, "train/loss": 0.7204726678657531, "eval/hcp-train-subset/loss": 0.8705275683633743, "eval/hcp-val/loss": 0.8681311367019531, "eval/nsd-val/loss": 0.83871970253606}
+{"epoch": 65, "train/lr": 3.6452170351940815e-05, "train/grad": 0.1895571667879877, "train/loss": 0.7221404563713074, "eval/hcp-train-subset/loss": 0.8718270919015331, "eval/hcp-val/loss": 0.8685564177651559, "eval/nsd-val/loss": 0.8460862732702686}
+{"epoch": 66, "train/lr": 3.458807350751516e-05, "train/grad": 0.19051936196423255, "train/loss": 0.7215649707508087, "eval/hcp-train-subset/loss": 0.8701012240302178, "eval/hcp-val/loss": 0.8664481437975361, "eval/nsd-val/loss": 0.8381536699110462}
+{"epoch": 67, "train/lr": 3.2754497944910164e-05, "train/grad": 0.19591290835009723, "train/loss": 0.7148278578662872, "eval/hcp-train-subset/loss": 0.8713854628224527, "eval/hcp-val/loss": 0.8682702299087278, "eval/nsd-val/loss": 0.8397899981467954}
+{"epoch": 68, "train/lr": 3.0953448652083367e-05, "train/grad": 0.19391465681472678, "train/loss": 0.7174831116771698, "eval/hcp-train-subset/loss": 0.8710751966122658, "eval/hcp-val/loss": 0.8685679368434414, "eval/nsd-val/loss": 0.8402931286442664}
+{"epoch": 69, "train/lr": 2.9186895049993948e-05, "train/grad": 0.19518698169096824, "train/loss": 0.7163275263023376, "eval/hcp-train-subset/loss": 0.870512455701828, "eval/hcp-val/loss": 0.867338785240727, "eval/nsd-val/loss": 0.8393718744477918}
+{"epoch": 70, "train/lr": 2.7456768839068717e-05, "train/grad": 0.19657022177829633, "train/loss": 0.7163009504032135, "eval/hcp-train-subset/loss": 0.8719363049153359, "eval/hcp-val/loss": 0.8699121225264764, "eval/nsd-val/loss": 0.8437373215152372}
+{"epoch": 71, "train/lr": 2.5764961886919063e-05, "train/grad": 0.19868654596038574, "train/loss": 0.7157794020652771, "eval/hcp-train-subset/loss": 0.8731889157525955, "eval/hcp-val/loss": 0.8693707037356592, "eval/nsd-val/loss": 0.8405706324884968}
+{"epoch": 72, "train/lr": 2.411332415960724e-05, "train/grad": 0.19881315912669117, "train/loss": 0.7153051083278656, "eval/hcp-train-subset/loss": 0.8710730969905853, "eval/hcp-val/loss": 0.868619813073066, "eval/nsd-val/loss": 0.8423014427385023}
+{"epoch": 73, "train/lr": 2.2503661698739544e-05, "train/grad": 0.20231571043608226, "train/loss": 0.7095257316207886, "eval/hcp-train-subset/loss": 0.8719962383470228, "eval/hcp-val/loss": 0.8693943331318517, "eval/nsd-val/loss": 0.8413551940071967}
+{"epoch": 74, "train/lr": 2.0937734646583902e-05, "train/grad": 0.20074898637396027, "train/loss": 0.7185145177268982, "eval/hcp-train-subset/loss": 0.8739026602237455, "eval/hcp-val/loss": 0.8716033639446381, "eval/nsd-val/loss": 0.8412466251081036}
+{"epoch": 75, "train/lr": 1.9417255321381202e-05, "train/grad": 0.20581496126111343, "train/loss": 0.7118735445785522, "eval/hcp-train-subset/loss": 0.8721510864073231, "eval/hcp-val/loss": 0.8696432094420156, "eval/nsd-val/loss": 0.8432149733266523}
+{"epoch": 76, "train/lr": 1.7943886344950134e-05, "train/grad": 0.2044664387055762, "train/loss": 0.7115664150524139, "eval/hcp-train-subset/loss": 0.8722263901464401, "eval/hcp-val/loss": 0.8695706196369664, "eval/nsd-val/loss": 0.8433404141856778}
+{"epoch": 77, "train/lr": 1.651923882463461e-05, "train/grad": 0.20386270371307885, "train/loss": 0.7118117702579498, "eval/hcp-train-subset/loss": 0.8739733119164744, "eval/hcp-val/loss": 0.8728330366073116, "eval/nsd-val/loss": 0.8437683851488175}
+{"epoch": 78, "train/lr": 1.5144870591581508e-05, "train/grad": 0.20352018478442635, "train/loss": 0.7160713409042359, "eval/hcp-train-subset/loss": 0.8724139102043644, "eval/hcp-val/loss": 0.8698414112291029, "eval/nsd-val/loss": 0.8450444465683352}
+{"epoch": 79, "train/lr": 1.3822284497275662e-05, "train/grad": 0.20562076799906545, "train/loss": 0.7119615220165253, "eval/hcp-train-subset/loss": 0.8747318137076593, "eval/hcp-val/loss": 0.8716151637415732, "eval/nsd-val/loss": 0.844193696975708}
+{"epoch": 80, "train/lr": 1.2552926770192975e-05, "train/grad": 0.20649914063687436, "train/loss": 0.7110919276618958, "eval/hcp-train-subset/loss": 0.8739719967688283, "eval/hcp-val/loss": 0.8717914041011564, "eval/nsd-val/loss": 0.843709175625155}
+{"epoch": 81, "train/lr": 1.1338185434371453e-05, "train/grad": 0.20632213215068193, "train/loss": 0.7137980181980133, "eval/hcp-train-subset/loss": 0.874406042598909, "eval/hcp-val/loss": 0.8710521055806068, "eval/nsd-val/loss": 0.8447390910117857}
+{"epoch": 82, "train/lr": 1.0179388791627326e-05, "train/grad": 0.20756376241470917, "train/loss": 0.7112266501331329, "eval/hcp-train-subset/loss": 0.8749512558983218, "eval/hcp-val/loss": 0.8725365534905465, "eval/nsd-val/loss": 0.8450656956241976}
+{"epoch": 83, "train/lr": 9.07780396907607e-06, "train/grad": 0.20902254831386372, "train/loss": 0.7091298444461822, "eval/hcp-train-subset/loss": 0.874635495485798, "eval/hcp-val/loss": 0.8719231563229715, "eval/nsd-val/loss": 0.8439989013056601}
+{"epoch": 84, "train/lr": 8.034635533547902e-06, "train/grad": 0.20842173441765077, "train/loss": 0.7131246441555024, "eval/hcp-train-subset/loss": 0.8733978136893241, "eval/hcp-val/loss": 0.8721965022625462, "eval/nsd-val/loss": 0.8446195760080891}
+{"epoch": 85, "train/lr": 7.051024174411275e-06, "train/grad": 0.20892425406063092, "train/loss": 0.7118115135288239, "eval/hcp-train-subset/loss": 0.8749923196531111, "eval/hcp-val/loss": 0.8721272830040224, "eval/nsd-val/loss": 0.8438200095007496}
+{"epoch": 86, "train/lr": 6.1280454562463606e-06, "train/grad": 0.21078284344160106, "train/loss": 0.7100567783260345, "eval/hcp-train-subset/loss": 0.8739077996823096, "eval/hcp-val/loss": 0.871945938756389, "eval/nsd-val/loss": 0.8460658077270754}
+{"epoch": 87, "train/lr": 5.266708642730326e-06, "train/grad": 0.21001632502075956, "train/loss": 0.7114262472438813, "eval/hcp-train-subset/loss": 0.8740562565865055, "eval/hcp-val/loss": 0.8712447535607123, "eval/nsd-val/loss": 0.8460225443686208}
+{"epoch": 88, "train/lr": 4.467955593022733e-06, "train/grad": 0.20902271229104077, "train/loss": 0.7131079891490937, "eval/hcp-train-subset/loss": 0.8759983720317963, "eval/hcp-val/loss": 0.8730242540759425, "eval/nsd-val/loss": 0.8468314092005452}
+{"epoch": 89, "train/lr": 3.732659731856291e-06, "train/grad": 0.21100131321397925, "train/loss": 0.7082633395290375, "eval/hcp-train-subset/loss": 0.8761817668714831, "eval/hcp-val/loss": 0.8725108206272125, "eval/nsd-val/loss": 0.8460712115610799}
+{"epoch": 90, "train/lr": 3.0616250944596583e-06, "train/grad": 0.21295957351711586, "train/loss": 0.7054266449785233, "eval/hcp-train-subset/loss": 0.8761070361060481, "eval/hcp-val/loss": 0.8739212607183764, "eval/nsd-val/loss": 0.8470285813654622}
+{"epoch": 91, "train/lr": 2.4555854473568305e-06, "train/grad": 0.21217555239169297, "train/loss": 0.7091550577688217, "eval/hcp-train-subset/loss": 0.8751362014201379, "eval/hcp-val/loss": 0.872117583790133, "eval/nsd-val/loss": 0.8472357536515882}
+{"epoch": 92, "train/lr": 1.915203486004091e-06, "train/grad": 0.2120047974902254, "train/loss": 0.7074762072086335, "eval/hcp-train-subset/loss": 0.8745175619279185, "eval/hcp-val/loss": 0.872350683135371, "eval/nsd-val/loss": 0.847239847144773}
+{"epoch": 93, "train/lr": 1.4410701101423926e-06, "train/grad": 0.21079717397394085, "train/loss": 0.7093232870435715, "eval/hcp-train-subset/loss": 0.8745136626305119, "eval/hcp-val/loss": 0.8721338539354263, "eval/nsd-val/loss": 0.8467171124873623}
+{"epoch": 94, "train/lr": 1.0337037776570775e-06, "train/grad": 0.2118202142452007, "train/loss": 0.7079074987983703, "eval/hcp-train-subset/loss": 0.8746398187452747, "eval/hcp-val/loss": 0.8725971262301168, "eval/nsd-val/loss": 0.8466979649759108}
+{"epoch": 95, "train/lr": 6.935499376518293e-07, "train/grad": 0.21210052014790473, "train/loss": 0.7090149361419678, "eval/hcp-train-subset/loss": 0.8753533920934123, "eval/hcp-val/loss": 0.872240723140778, "eval/nsd-val/loss": 0.8467560920023149}
+{"epoch": 96, "train/lr": 4.209805433566085e-07, "train/grad": 0.2098160737817785, "train/loss": 0.7156093966960907, "eval/hcp-train-subset/loss": 0.8753257903360552, "eval/hcp-val/loss": 0.8723211663384591, "eval/nsd-val/loss": 0.8467711863979217}
+{"epoch": 97, "train/lr": 2.1629364540224422e-07, "train/grad": 0.21401807334025383, "train/loss": 0.7060976057100296, "eval/hcp-train-subset/loss": 0.8748501702662437, "eval/hcp-val/loss": 0.8720592654520466, "eval/nsd-val/loss": 0.8469197596273115}
+{"epoch": 98, "train/lr": 7.971306590647406e-08, "train/grad": 0.211841072680321, "train/loss": 0.7058749144601821, "eval/hcp-train-subset/loss": 0.8750185783832304, "eval/hcp-val/loss": 0.8720644250992806, "eval/nsd-val/loss": 0.8466727656702842}
+{"epoch": 99, "train/lr": 1.1388153727718725e-08, "train/grad": 0.21050761424233988, "train/loss": 0.7094936176300048, "eval/hcp-train-subset/loss": 0.8747062904219474, "eval/hcp-val/loss": 0.8730447455759971, "eval/nsd-val/loss": 0.8471881987587098}

data_scaling/n400_1/eval_v2/aabc_age__patch__logistic/config.yaml

@@ -0,0 +1,30 @@
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (aabc_age patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_1/eval_v2/aabc_age__patch__logistic
+model: flat_mae
+representation: patch
+dataset: aabc_age
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/aabc_age__patch__logistic
+remote_dir: null

data_scaling/n400_1/eval_v2/aabc_age__patch__logistic/eval_table.csv

@@ -0,0 +1,203 @@
+model,repr,clf,dataset,trial,C,split,acc,acc_std,f1,f1_std,bacc,bacc_std
+flat_mae,patch,logistic,aabc_age,,0.046415888336127774,train,0.8366141732283464,0.017304224075997805,0.8374166745166707,0.017270180496607626,0.8371057381214022,0.01731051358221176
+flat_mae,patch,logistic,aabc_age,,0.046415888336127774,test,0.38461538461538464,0.057949427043539926,0.37859649122807015,0.05797698426232789,0.3740842490842491,0.057542111571559786
+flat_mae,patch,logistic,aabc_age,1,0.046415888336127774,train,0.8523622047244095,0.015160432783316071,0.8522637481254558,0.015262048372313798,0.8529298661843653,0.015172538605698267
+flat_mae,patch,logistic,aabc_age,1,0.046415888336127774,test,0.5192307692307693,0.0623871762729618,0.5219915848527349,0.06017802432524102,0.5157967032967032,0.062236807096345596
+flat_mae,patch,logistic,aabc_age,2,9.999999999999999e-05,train,0.4881889763779528,0.020160661965932112,0.4694745631175553,0.020762451132304603,0.48674022658705884,0.020040536296096464
+flat_mae,patch,logistic,aabc_age,2,9.999999999999999e-05,test,0.46153846153846156,0.06388899165300599,0.426585173193946,0.06119357747452577,0.4535256410256411,0.06262066287810407
+flat_mae,patch,logistic,aabc_age,3,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_age,3,21.54434690031882,test,0.46153846153846156,0.06524192575131152,0.45466570466570466,0.06482470264517816,0.45970695970695974,0.06498110562727226
+flat_mae,patch,logistic,aabc_age,4,0.3593813663804626,train,0.984251968503937,0.005526555083017095,0.9846078279860113,0.005399985646330674,0.9844236980450298,0.005440807435428798
+flat_mae,patch,logistic,aabc_age,4,0.3593813663804626,test,0.5192307692307693,0.06443055065718564,0.5116483516483517,0.0650331388161421,0.5201465201465201,0.06462726599464377
+flat_mae,patch,logistic,aabc_age,5,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_age,5,2.782559402207126,test,0.4807692307692308,0.0617952694422784,0.47609978588239454,0.06279175548211148,0.48031135531135527,0.06206483155694821
+flat_mae,patch,logistic,aabc_age,6,0.3593813663804626,train,0.984251968503937,0.005289350664728219,0.9845581997411132,0.005198571345366485,0.9849088204776089,0.005092115159400898
+flat_mae,patch,logistic,aabc_age,6,0.3593813663804626,test,0.5192307692307693,0.06371162843338048,0.5122053872053872,0.06416665991261807,0.5249542124542124,0.0642360784865812
+flat_mae,patch,logistic,aabc_age,7,0.000774263682681127,train,0.5610236220472441,0.02042458671130176,0.5546119670779828,0.02109033117091042,0.5624866941956965,0.020413143068824684
+flat_mae,patch,logistic,aabc_age,7,0.000774263682681127,test,0.4807692307692308,0.0615353845384577,0.4376142142062609,0.06121649682855208,0.4741300366300366,0.06066880513261205
+flat_mae,patch,logistic,aabc_age,8,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_age,8,21.54434690031882,test,0.46153846153846156,0.06576022205508227,0.46042572463768117,0.06487791290788249,0.46863553113553114,0.06635234674296442
+flat_mae,patch,logistic,aabc_age,9,0.005994842503189409,train,0.6614173228346457,0.019985298629850965,0.6588975092055919,0.0204787729543169,0.6637481803413574,0.019887646047387998
+flat_mae,patch,logistic,aabc_age,9,0.005994842503189409,test,0.5384615384615384,0.06190242130412678,0.5288539553752536,0.0679387437598958,0.538003663003663,0.06219816353545317
+flat_mae,patch,logistic,aabc_age,10,0.3593813663804626,train,0.9862204724409449,0.005153421496006914,0.986480137602754,0.005044066121638992,0.9864398270772878,0.005083290707373919
+flat_mae,patch,logistic,aabc_age,10,0.3593813663804626,test,0.5,0.07216594125294201,0.5118248992386923,0.07041445830854252,0.5027472527472527,0.0723557845570276
+flat_mae,patch,logistic,aabc_age,11,0.046415888336127774,train,0.8346456692913385,0.016644918281764242,0.8343608133266487,0.016803556748954913,0.8354550297378732,0.016667297805469383
+flat_mae,patch,logistic,aabc_age,11,0.046415888336127774,test,0.5576923076923077,0.06693654815786321,0.5756060606060607,0.06473787371741364,0.5606684981684982,0.06697772160222111
+flat_mae,patch,logistic,aabc_age,12,0.005994842503189409,train,0.6751968503937008,0.02010432945581935,0.671980968951026,0.02045283363922507,0.6779110702373852,0.019933471815921006
+flat_mae,patch,logistic,aabc_age,12,0.005994842503189409,test,0.46153846153846156,0.06553687024070302,0.45795074812824,0.06634004442730149,0.4597069597069597,0.0654509001386346
+flat_mae,patch,logistic,aabc_age,13,0.046415888336127774,train,0.8464566929133859,0.016422120185045026,0.8458267804361161,0.016663516074116976,0.8465139511957482,0.016489442129094176
+flat_mae,patch,logistic,aabc_age,13,0.046415888336127774,test,0.38461538461538464,0.06320731129531956,0.38568376068376065,0.06044551634336766,0.3882783882783883,0.06412114375397251
+flat_mae,patch,logistic,aabc_age,14,0.3593813663804626,train,0.9822834645669292,0.005997802468176894,0.9824005858708567,0.005943817699825465,0.9820900277911502,0.0060531211192861294
+flat_mae,patch,logistic,aabc_age,14,0.3593813663804626,test,0.5,0.06341728060465807,0.5011733094491715,0.06418396716573467,0.5057234432234432,0.06352568745341663
+flat_mae,patch,logistic,aabc_age,15,0.3593813663804626,train,0.984251968503937,0.0055891980174501655,0.984601940608802,0.00544737054833931,0.9841561434936295,0.005622061272098373
+flat_mae,patch,logistic,aabc_age,15,0.3593813663804626,test,0.5769230769230769,0.06423075771533752,0.5815322580645161,0.06388026233559048,0.5771520146520146,0.06442034689101897
+flat_mae,patch,logistic,aabc_age,16,0.000774263682681127,train,0.5551181102362205,0.02136047341562562,0.549334969982032,0.021723616908749223,0.5557356167851644,0.02123994782391769
+flat_mae,patch,logistic,aabc_age,16,0.000774263682681127,test,0.4807692307692308,0.0633639809579562,0.47022546419098143,0.06343769505593227,0.47870879120879123,0.06334870908274497
+flat_mae,patch,logistic,aabc_age,17,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_age,17,2.782559402207126,test,0.4230769230769231,0.06329348377801144,0.4166269841269841,0.06275823129090746,0.4237637362637363,0.06342433434426233
+flat_mae,patch,logistic,aabc_age,18,0.3593813663804626,train,0.9822834645669292,0.005918154446202299,0.9823429874669921,0.005889874389855083,0.9824075690127716,0.005875951780187102
+flat_mae,patch,logistic,aabc_age,18,0.3593813663804626,test,0.40384615384615385,0.06913209848847436,0.4096962629796213,0.06846346696398492,0.40796703296703296,0.06957443001465052
+flat_mae,patch,logistic,aabc_age,19,0.046415888336127774,train,0.8326771653543307,0.016970758203841932,0.832627450502353,0.01709774244295392,0.833874036467973,0.016850955002157306
+flat_mae,patch,logistic,aabc_age,19,0.046415888336127774,test,0.4423076923076923,0.06807665145534687,0.4446703296703297,0.06880843796884462,0.44505494505494503,0.06839798620861753
+flat_mae,patch,logistic,aabc_age,20,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_age,20,2.782559402207126,test,0.5,0.06821011525049057,0.4909688013136289,0.06848508840388282,0.49793956043956045,0.06819440187499051
+flat_mae,patch,logistic,aabc_age,21,0.3593813663804626,train,0.9862204724409449,0.005027636129857651,0.9863201337071931,0.004992184818086114,0.9856871500933084,0.005222947866718278
+flat_mae,patch,logistic,aabc_age,21,0.3593813663804626,test,0.5192307692307693,0.05233137190258917,0.48503637566137564,0.04975114814772035,0.5258699633699634,0.053755919961926545
+flat_mae,patch,logistic,aabc_age,22,0.3593813663804626,train,0.9862204724409449,0.005628908008593731,0.9864319197338205,0.005533432613905819,0.9864398270772878,0.0055499154431457275
+flat_mae,patch,logistic,aabc_age,22,0.3593813663804626,test,0.4423076923076923,0.06771820964230939,0.45055858120374254,0.06617821920266428,0.4466575091575092,0.06814361301590238
+flat_mae,patch,logistic,aabc_age,23,0.046415888336127774,train,0.844488188976378,0.01598435669644914,0.8447804141166625,0.016050928210342395,0.8454180803584272,0.01593752219359285
+flat_mae,patch,logistic,aabc_age,23,0.046415888336127774,test,0.4230769230769231,0.05982469459044682,0.41280786099865047,0.05780069976389986,0.41941391941391937,0.05946350142948324
+flat_mae,patch,logistic,aabc_age,24,0.046415888336127774,train,0.8523622047244095,0.015164611779762157,0.853063538114029,0.015114696670533573,0.853297394076208,0.01507044461471897
+flat_mae,patch,logistic,aabc_age,24,0.046415888336127774,test,0.36538461538461536,0.060143750382389456,0.35896993505689156,0.05873125603552922,0.36744505494505497,0.060525387632811414
+flat_mae,patch,logistic,aabc_age,25,0.046415888336127774,train,0.84251968503937,0.015523026442311447,0.842369971808399,0.015616946248626171,0.8432343701152115,0.015436272940504822
+flat_mae,patch,logistic,aabc_age,25,0.046415888336127774,test,0.3076923076923077,0.06092573566459571,0.32169055082848186,0.058994593669884424,0.3067765567765568,0.060883066869048846
+flat_mae,patch,logistic,aabc_age,26,0.000774263682681127,train,0.5728346456692913,0.022061568743532965,0.5661268359175173,0.02240147970384463,0.5732780611021713,0.021969989529761317
+flat_mae,patch,logistic,aabc_age,26,0.000774263682681127,test,0.40384615384615385,0.06300440249480833,0.39202551834130783,0.06404430005433341,0.4001831501831502,0.06261948844984895
+flat_mae,patch,logistic,aabc_age,27,0.046415888336127774,train,0.8543307086614174,0.015088127975438565,0.8543510571706774,0.01521560509407616,0.8552635364382448,0.01505773716552186
+flat_mae,patch,logistic,aabc_age,27,0.046415888336127774,test,0.38461538461538464,0.06725628806307757,0.38421481899742765,0.06760645536317131,0.38278388278388276,0.06702383451020436
+flat_mae,patch,logistic,aabc_age,28,0.005994842503189409,train,0.6594488188976378,0.020846307613989645,0.6564926844695819,0.021270798784434195,0.6606118464332773,0.020708451110238335
+flat_mae,patch,logistic,aabc_age,28,0.005994842503189409,test,0.5192307692307693,0.06609267073027351,0.5105042016806722,0.0700139892398969,0.5247252747252747,0.06648703176882743
+flat_mae,patch,logistic,aabc_age,29,0.000774263682681127,train,0.5433070866141733,0.0209330468012876,0.5333680345685327,0.021495813505817234,0.5442091940950039,0.02082994310841855
+flat_mae,patch,logistic,aabc_age,29,0.000774263682681127,test,0.46153846153846156,0.06464829747846347,0.45146520146520147,0.06577861543241968,0.459478021978022,0.06465919222965302
+flat_mae,patch,logistic,aabc_age,30,0.046415888336127774,train,0.8385826771653543,0.016592299572657513,0.8387770345596433,0.016711163108045823,0.8396372478130532,0.01655672913119235
+flat_mae,patch,logistic,aabc_age,30,0.046415888336127774,test,0.5769230769230769,0.069163332540063,0.5751719576719576,0.06955112487931323,0.5828754578754578,0.06896529519250544
+flat_mae,patch,logistic,aabc_age,31,0.005994842503189409,train,0.6515748031496063,0.021155094825504196,0.6482721884400435,0.02159592489564974,0.6534176018289607,0.02107357917020877
+flat_mae,patch,logistic,aabc_age,31,0.005994842503189409,test,0.5576923076923077,0.06704819206489909,0.5408740176232436,0.07130820324176729,0.5558608058608059,0.06677979720913636
+flat_mae,patch,logistic,aabc_age,32,0.3593813663804626,train,0.9822834645669292,0.005918058849558781,0.982441234915656,0.005866505931011071,0.9824575556829929,0.005868523624928659
+flat_mae,patch,logistic,aabc_age,32,0.3593813663804626,test,0.4423076923076923,0.06118968470904961,0.4282176157176157,0.06127410700816197,0.44299450549450553,0.061556466042595244
+flat_mae,patch,logistic,aabc_age,33,0.046415888336127774,train,0.8523622047244095,0.015436896420444835,0.8526464040514825,0.015503765630147612,0.8530298395248077,0.015386746705399844
+flat_mae,patch,logistic,aabc_age,33,0.046415888336127774,test,0.5576923076923077,0.0652738428178778,0.5582919254658385,0.06614300974257094,0.5590659340659341,0.06556453681880652
+flat_mae,patch,logistic,aabc_age,34,0.046415888336127774,train,0.8523622047244095,0.016020514218143933,0.8525601553415326,0.01605002066379656,0.8531974207357654,0.01597909810083145
+flat_mae,patch,logistic,aabc_age,34,0.046415888336127774,test,0.4423076923076923,0.06530171214805516,0.4492713737875028,0.06484840239157982,0.4432234432234432,0.06558736264401135
+flat_mae,patch,logistic,aabc_age,35,0.005994842503189409,train,0.6732283464566929,0.020152280025564857,0.6700672417604137,0.020667987492229593,0.6754098187725479,0.020115328320113115
+flat_mae,patch,logistic,aabc_age,35,0.005994842503189409,test,0.4230769230769231,0.06423058498361414,0.4195623985522118,0.06469782093245457,0.4223901098901099,0.06419945702259432
+flat_mae,patch,logistic,aabc_age,36,0.3593813663804626,train,0.9803149606299213,0.0061001253202748115,0.9806555716522977,0.0059971673986943036,0.9810441436240505,0.005888779349119141
+flat_mae,patch,logistic,aabc_age,36,0.3593813663804626,test,0.4807692307692308,0.06299219215259419,0.4533880237300705,0.0646290857410857,0.4860347985347986,0.06399464873685771
+flat_mae,patch,logistic,aabc_age,37,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_age,37,166.81005372000556,test,0.38461538461538464,0.06413009152537719,0.38902116402116405,0.06428738273190888,0.3841575091575091,0.06433070209897496
+flat_mae,patch,logistic,aabc_age,38,0.046415888336127774,train,0.8385826771653543,0.01647649977847511,0.8377610383862729,0.016661077285063258,0.8388345841588526,0.016516882295181088
+flat_mae,patch,logistic,aabc_age,38,0.046415888336127774,test,0.5769230769230769,0.06492617564606488,0.5681697612732095,0.0663409114958182,0.575091575091575,0.06485475122448725
+flat_mae,patch,logistic,aabc_age,39,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_age,39,2.782559402207126,test,0.4807692307692308,0.06719661163618755,0.4769137866963954,0.06720643835893196,0.4775641025641026,0.06707026983024775
+flat_mae,patch,logistic,aabc_age,40,0.046415888336127774,train,0.84251968503937,0.01566507431482874,0.8435775778757,0.015617190481106267,0.8437018713474969,0.015598986669402659
+flat_mae,patch,logistic,aabc_age,40,0.046415888336127774,test,0.5384615384615384,0.06683995722363087,0.5308862433862434,0.06828493723383323,0.5396062271062272,0.06709156417340942
+flat_mae,patch,logistic,aabc_age,41,0.3593813663804626,train,0.9862204724409449,0.005136096500161814,0.9863440390498199,0.005103478978239385,0.986707381628688,0.004985251452857188
+flat_mae,patch,logistic,aabc_age,41,0.3593813663804626,test,0.5,0.06252781629523685,0.5004417434140073,0.060324350117111,0.5022893772893773,0.06288170989868064
+flat_mae,patch,logistic,aabc_age,42,0.000774263682681127,train,0.5511811023622047,0.020533885683935285,0.5423376508885052,0.02128698100624967,0.5517209799209419,0.020568341902777322
+flat_mae,patch,logistic,aabc_age,42,0.000774263682681127,test,0.4230769230769231,0.04863485491317966,0.38369963369963367,0.05961748395225562,0.4251373626373626,0.04911938812421867
+flat_mae,patch,logistic,aabc_age,43,0.046415888336127774,train,0.84251968503937,0.016228804908193888,0.8427866300898963,0.016307922810576123,0.8432343701152114,0.01618598864758266
+flat_mae,patch,logistic,aabc_age,43,0.046415888336127774,test,0.5769230769230769,0.06724338686386251,0.5769047619047619,0.06881765818112363,0.5771520146520146,0.06738253467740424
+flat_mae,patch,logistic,aabc_age,44,0.046415888336127774,train,0.8346456692913385,0.016055720042162307,0.8342399098639675,0.016161231241252728,0.8353374351971369,0.015980090269044893
+flat_mae,patch,logistic,aabc_age,44,0.046415888336127774,test,0.5,0.07110550332815524,0.5025000000000001,0.07129035176151845,0.4983974358974359,0.07138900894804541
+flat_mae,patch,logistic,aabc_age,45,0.046415888336127774,train,0.8346456692913385,0.015948952527930285,0.8345031401663541,0.01608157409722121,0.8360401255108949,0.015917524101886488
+flat_mae,patch,logistic,aabc_age,45,0.046415888336127774,test,0.4807692307692308,0.06487214325435997,0.48148693510387663,0.06353068449861626,0.4862637362637363,0.06574287996009155
+flat_mae,patch,logistic,aabc_age,46,0.000774263682681127,train,0.5393700787401575,0.020262969468364294,0.5311671495256034,0.02082010109510627,0.5402769093709303,0.02022785668812309
+flat_mae,patch,logistic,aabc_age,46,0.000774263682681127,test,0.5769230769230769,0.06438516680216036,0.5743686061381075,0.07027834743891563,0.5753205128205128,0.0645452214856435
+flat_mae,patch,logistic,aabc_age,47,0.000774263682681127,train,0.5728346456692913,0.019582211351785648,0.567200617381456,0.01997371747270983,0.5735456156535715,0.019565219470533926
+flat_mae,patch,logistic,aabc_age,47,0.000774263682681127,test,0.4230769230769231,0.06323548316697633,0.4089279895731508,0.06411366667151348,0.41941391941391937,0.06257990943290731
+flat_mae,patch,logistic,aabc_age,48,0.046415888336127774,train,0.8366141732283464,0.01642375862827067,0.8365362923424959,0.01655581959045081,0.8377886999917528,0.01635772447292449
+flat_mae,patch,logistic,aabc_age,48,0.046415888336127774,test,0.40384615384615385,0.06832149678598182,0.4122980481676134,0.06558963688299861,0.40613553113553114,0.0686484628371634
+flat_mae,patch,logistic,aabc_age,49,0.005994842503189409,train,0.6496062992125984,0.019538563701330332,0.6453295970679597,0.020132377964804066,0.6516190406778816,0.019537979257332902
+flat_mae,patch,logistic,aabc_age,49,0.005994842503189409,test,0.40384615384615385,0.0649302197128014,0.4029971988795518,0.06477845406577831,0.40041208791208793,0.0647196560149341
+flat_mae,patch,logistic,aabc_age,50,0.3593813663804626,train,0.9881889763779528,0.005031383619805011,0.9883273189342456,0.004956999800576488,0.988238388228367,0.005018578953962119
+flat_mae,patch,logistic,aabc_age,50,0.3593813663804626,test,0.5384615384615384,0.06745526679030689,0.5433214882943144,0.06734358620754474,0.5384615384615385,0.06762030133758463
+flat_mae,patch,logistic,aabc_age,51,0.046415888336127774,train,0.8287401574803149,0.016786053483269472,0.8293113738509714,0.016753576480814814,0.8298741438733843,0.016685835255586135
+flat_mae,patch,logistic,aabc_age,51,0.046415888336127774,test,0.5769230769230769,0.06488036327319337,0.5727609427609428,0.06440225643183178,0.5737179487179487,0.06493301202675035
+flat_mae,patch,logistic,aabc_age,52,0.000774263682681127,train,0.5728346456692913,0.019701928188173578,0.5614659978824065,0.020835827380585623,0.5736808313946018,0.019631767584404446
+flat_mae,patch,logistic,aabc_age,52,0.000774263682681127,test,0.4230769230769231,0.06208901559847898,0.4263083735909823,0.06099594284855781,0.4269688644688645,0.06277916775184823
+flat_mae,patch,logistic,aabc_age,53,9.999999999999999e-05,train,0.468503937007874,0.020784903139644335,0.4530944018923541,0.021079195068412368,0.46744920778919385,0.020673724135065075
+flat_mae,patch,logistic,aabc_age,53,9.999999999999999e-05,test,0.5192307692307693,0.058336150960409956,0.4778935185185185,0.056919293573939544,0.5112179487179487,0.057354562492450176
+flat_mae,patch,logistic,aabc_age,54,0.046415888336127774,train,0.8484251968503937,0.015622559205479515,0.8486552321005761,0.01563466009548066,0.8496502851038285,0.01558910593982665
+flat_mae,patch,logistic,aabc_age,54,0.046415888336127774,test,0.40384615384615385,0.060034248115101455,0.37649262413327744,0.05826356801998579,0.41048534798534797,0.06119411094178349
+flat_mae,patch,logistic,aabc_age,55,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_age,55,166.81005372000556,test,0.4807692307692308,0.058148249628954804,0.4733514492753623,0.054474379757819845,0.48901098901098905,0.05936499621774483
+flat_mae,patch,logistic,aabc_age,56,0.000774263682681127,train,0.5570866141732284,0.021456591417805166,0.5516428130371738,0.0221731970124133,0.5582868549202227,0.02149860627951304
+flat_mae,patch,logistic,aabc_age,56,0.000774263682681127,test,0.40384615384615385,0.06304242724654639,0.39622153209109734,0.06076532629668424,0.4015567765567766,0.0628410538051134
+flat_mae,patch,logistic,aabc_age,57,0.046415888336127774,train,0.84251968503937,0.01585001427273507,0.8429679148216352,0.015845536674912588,0.8429491943635175,0.01579359627148672
+flat_mae,patch,logistic,aabc_age,57,0.046415888336127774,test,0.5192307692307693,0.0574054921810133,0.49670506912442397,0.05769100681961992,0.5155677655677655,0.05699896632817456
+flat_mae,patch,logistic,aabc_age,58,0.3593813663804626,train,0.9881889763779528,0.0045682373228176755,0.9884091937456655,0.0044769662048898306,0.9882883748985881,0.004523856120692117
+flat_mae,patch,logistic,aabc_age,58,0.3593813663804626,test,0.4230769230769231,0.05967917182984243,0.40654511453950437,0.06115877536204072,0.4207875457875458,0.05927754848986136
+flat_mae,patch,logistic,aabc_age,59,0.000774263682681127,train,0.5531496062992126,0.02068855495274167,0.548695508777664,0.021214498182337373,0.5545221514071068,0.020672050320110234
+flat_mae,patch,logistic,aabc_age,59,0.000774263682681127,test,0.5192307692307693,0.06252805287588688,0.49991466120498373,0.06623523954461033,0.5157967032967034,0.06215046603781298
+flat_mae,patch,logistic,aabc_age,60,0.005994842503189409,train,0.6437007874015748,0.020110982295830755,0.6373354488322319,0.020686630752902453,0.6452707335597798,0.019998915197279275
+flat_mae,patch,logistic,aabc_age,60,0.005994842503189409,test,0.4423076923076923,0.05898123009024635,0.4291287565481114,0.06176947870265391,0.44459706959706957,0.059268388111311805
+flat_mae,patch,logistic,aabc_age,61,0.3593813663804626,train,0.9881889763779528,0.005015833013285185,0.9882702763688528,0.004987194708346153,0.988238388228367,0.004995454120842215
+flat_mae,patch,logistic,aabc_age,61,0.3593813663804626,test,0.3269230769230769,0.05999924802290117,0.32605377714073364,0.059493436151958734,0.3276098901098901,0.06022930518905783
+flat_mae,patch,logistic,aabc_age,62,0.046415888336127774,train,0.84251968503937,0.01590315869005001,0.8427382825934824,0.015979609557455223,0.8437694792180118,0.01587821722224166
+flat_mae,patch,logistic,aabc_age,62,0.046415888336127774,test,0.46153846153846156,0.06850689940635314,0.46724310311266837,0.06788405332706389,0.4626831501831502,0.06884036815356925
+flat_mae,patch,logistic,aabc_age,63,0.005994842503189409,train,0.6811023622047244,0.019722460101388586,0.6784399453588775,0.020234469117763814,0.6833067536906227,0.019652142929701525
+flat_mae,patch,logistic,aabc_age,63,0.005994842503189409,test,0.4230769230769231,0.057567698762660324,0.41338566827697265,0.05654634466025146,0.42078754578754574,0.05736800421015458
+flat_mae,patch,logistic,aabc_age,64,0.3593813663804626,train,0.9822834645669292,0.006042064300641128,0.9825556325174406,0.0059483277986337795,0.9826751235641717,0.005920658717795833
+flat_mae,patch,logistic,aabc_age,64,0.3593813663804626,test,0.40384615384615385,0.06583110000456206,0.4105528127267257,0.0649950965697301,0.40476190476190477,0.0660530661722229
+flat_mae,patch,logistic,aabc_age,65,0.005994842503189409,train,0.65748031496063,0.020643393974815044,0.6545504262215979,0.02089566619682225,0.6596159489363987,0.02059833384736603
+flat_mae,patch,logistic,aabc_age,65,0.005994842503189409,test,0.5769230769230769,0.06426212960815372,0.5621632996632997,0.06409258973084567,0.5707417582417582,0.06371763986513436
+flat_mae,patch,logistic,aabc_age,66,0.005994842503189409,train,0.6791338582677166,0.019349365966251357,0.6762686227659211,0.019775659507591196,0.6815081925395434,0.019316414111550124
+flat_mae,patch,logistic,aabc_age,66,0.005994842503189409,test,0.4807692307692308,0.06016132158941171,0.4596607566662013,0.06068762171743191,0.4757326007326007,0.05987106851884831
+flat_mae,patch,logistic,aabc_age,67,0.046415888336127774,train,0.84251968503937,0.016581048349315227,0.8431730211813073,0.016610184471139867,0.8425816664716748,0.016681710516475722
+flat_mae,patch,logistic,aabc_age,67,0.046415888336127774,test,0.5192307692307693,0.06417316624783438,0.5067438055165966,0.06537415870734964,0.5144230769230769,0.06381588776883253
+flat_mae,patch,logistic,aabc_age,68,0.046415888336127774,train,0.8503937007874016,0.01580411813548293,0.85022044131321,0.015899428555084014,0.850428614719528,0.015760100115970378
+flat_mae,patch,logistic,aabc_age,68,0.046415888336127774,test,0.46153846153846156,0.06188234446280311,0.4398809523809524,0.06553892205323958,0.4578754578754579,0.06168321469782148
+flat_mae,patch,logistic,aabc_age,69,0.3593813663804626,train,0.9901574803149606,0.0041434758762160795,0.9902062227487987,0.004133871688238401,0.9903045039308463,0.004102602413405874
+flat_mae,patch,logistic,aabc_age,69,0.3593813663804626,test,0.40384615384615385,0.06426305038644224,0.3963817563922753,0.06504523412636465,0.40773809523809523,0.06493784273981329
+flat_mae,patch,logistic,aabc_age,70,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_age,70,2.782559402207126,test,0.38461538461538464,0.062168065300403766,0.3762747668997669,0.06254008406954978,0.38690476190476186,0.06265227732451385
+flat_mae,patch,logistic,aabc_age,71,0.000774263682681127,train,0.5551181102362205,0.02044959506620795,0.5499779922717755,0.020836274907345898,0.5553004810228064,0.020438680146624068
+flat_mae,patch,logistic,aabc_age,71,0.000774263682681127,test,0.5192307692307693,0.06314414804202098,0.5127731463938361,0.06407723343517392,0.5187728937728937,0.06344692866309007
+flat_mae,patch,logistic,aabc_age,72,0.046415888336127774,train,0.8326771653543307,0.016223470787781788,0.8329414961230179,0.01629707972194343,0.8331537249539211,0.016275344660622817
+flat_mae,patch,logistic,aabc_age,72,0.046415888336127774,test,0.5769230769230769,0.06315197223866517,0.5670588235294118,0.0659766853724605,0.5737179487179487,0.06326882135573896
+flat_mae,patch,logistic,aabc_age,73,0.046415888336127774,train,0.844488188976378,0.015980312050793845,0.8450532450167012,0.01593449503047919,0.845500432498576,0.015872013515514965
+flat_mae,patch,logistic,aabc_age,73,0.046415888336127774,test,0.5,0.06387463451880757,0.5131649831649832,0.060681143120914265,0.4981684981684981,0.06381269474662798
+flat_mae,patch,logistic,aabc_age,74,0.005994842503189409,train,0.6633858267716536,0.020166957663771773,0.6610446188469901,0.020419262022998368,0.6647940645084572,0.02008329406786545
+flat_mae,patch,logistic,aabc_age,74,0.005994842503189409,test,0.46153846153846156,0.06643343200207232,0.4487179487179487,0.06994467821680307,0.459478021978022,0.06653138472875025
+flat_mae,patch,logistic,aabc_age,75,0.005994842503189409,train,0.6712598425196851,0.020076231123149386,0.6686696567798635,0.020533151747504353,0.6738288255026477,0.020053442925762586
+flat_mae,patch,logistic,aabc_age,75,0.005994842503189409,test,0.5192307692307693,0.07084296952670001,0.5222355488922206,0.07127761237252711,0.5192307692307693,0.07084768210974748
+flat_mae,patch,logistic,aabc_age,76,0.046415888336127774,train,0.8543307086614174,0.015593200979780704,0.8543200760840167,0.0156666603636682,0.8551635630978022,0.015575475529999617
+flat_mae,patch,logistic,aabc_age,76,0.046415888336127774,test,0.5,0.06617507485862315,0.4999823165340407,0.06630484056353396,0.5041208791208791,0.06666621260109828
+flat_mae,patch,logistic,aabc_age,77,0.3593813663804626,train,0.984251968503937,0.005426875104693928,0.9842061301638507,0.00545015171760585,0.9842061301638507,0.005442270672159946
+flat_mae,patch,logistic,aabc_age,77,0.3593813663804626,test,0.5192307692307693,0.060222209333778344,0.5068734015345269,0.06189476526915229,0.5231227106227107,0.06061078712711297
+flat_mae,patch,logistic,aabc_age,78,0.046415888336127774,train,0.8366141732283464,0.016117756343987197,0.8369691236207231,0.016134155216635103,0.837353564229395,0.016059157618446042
+flat_mae,patch,logistic,aabc_age,78,0.046415888336127774,test,0.5384615384615384,0.06728517205668887,0.5408772262220538,0.0679401843207676,0.5412087912087912,0.06763574196912679
+flat_mae,patch,logistic,aabc_age,79,0.005994842503189409,train,0.6535433070866141,0.02003785554971665,0.6503583480571433,0.0205657637424062,0.6560188266342405,0.01995186624942705
+flat_mae,patch,logistic,aabc_age,79,0.005994842503189409,test,0.5,0.06592128172764379,0.49559419075548106,0.06713670130154296,0.4967948717948718,0.06592070603861995
+flat_mae,patch,logistic,aabc_age,80,0.3593813663804626,train,0.9803149606299213,0.006362497069762622,0.9805187052833295,0.006291234886607907,0.9803914399805136,0.006353343997347285
+flat_mae,patch,logistic,aabc_age,80,0.3593813663804626,test,0.5384615384615384,0.05982340876852798,0.5142857142857142,0.05976425329194862,0.5485347985347985,0.060826067221274344
+flat_mae,patch,logistic,aabc_age,81,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_age,81,21.54434690031882,test,0.4807692307692308,0.06616140386166806,0.49083639998182726,0.0655279966055778,0.48489010989010983,0.06648539090209249
+flat_mae,patch,logistic,aabc_age,82,0.3593813663804626,train,0.9881889763779528,0.004984761949752797,0.9881514076651944,0.0049891112081997075,0.988020820347188,0.005057113610509552
+flat_mae,patch,logistic,aabc_age,82,0.3593813663804626,test,0.5576923076923077,0.06565176194249356,0.5509476031215161,0.06716327472566319,0.5604395604395604,0.06582553511038909
+flat_mae,patch,logistic,aabc_age,83,0.3593813663804626,train,0.9862204724409449,0.004917094300054926,0.9862445801892442,0.00490188961419264,0.9864398270772878,0.004846208046352392
+flat_mae,patch,logistic,aabc_age,83,0.3593813663804626,test,0.40384615384615385,0.062201369398346744,0.39993961352657004,0.0607564887485186,0.40453296703296704,0.062312736109117806
+flat_mae,patch,logistic,aabc_age,84,0.3593813663804626,train,0.9862204724409449,0.005196886184558207,0.9864395045862365,0.005088050594614636,0.9864398270772878,0.005099977340462223
+flat_mae,patch,logistic,aabc_age,84,0.3593813663804626,test,0.38461538461538464,0.06326046233370854,0.37483766233766236,0.06495607335268341,0.3869047619047619,0.06373309536986203
+flat_mae,patch,logistic,aabc_age,85,9.999999999999999e-05,train,0.4704724409448819,0.020838330522669944,0.43745539649888754,0.021093376759289125,0.4667897913074499,0.02063862728475097
+flat_mae,patch,logistic,aabc_age,85,9.999999999999999e-05,test,0.46153846153846156,0.053040943769660925,0.3851662404092072,0.04907110049187005,0.4519230769230769,0.0511340841744937
+flat_mae,patch,logistic,aabc_age,86,0.3593813663804626,train,0.9901574803149606,0.004232926155194254,0.9902265318889731,0.004198225305084891,0.9905220718120253,0.0040883434528134815
+flat_mae,patch,logistic,aabc_age,86,0.3593813663804626,test,0.4423076923076923,0.061233571640546894,0.42765567765567764,0.06052194979561253,0.43727106227106227,0.06066059562879579
+flat_mae,patch,logistic,aabc_age,87,0.000774263682681127,train,0.5531496062992126,0.020486880129109816,0.5423112069186254,0.021584580744139086,0.5534695544017998,0.02042158499736588
+flat_mae,patch,logistic,aabc_age,87,0.000774263682681127,test,0.5192307692307693,0.06085460166970694,0.5059791758161323,0.06262279436498262,0.5247252747252747,0.061538786983484466
+flat_mae,patch,logistic,aabc_age,88,9.999999999999999e-05,train,0.49606299212598426,0.019055880686894986,0.4659381949782827,0.01960224329282935,0.49406968693240555,0.018882776380105135
+flat_mae,patch,logistic,aabc_age,88,9.999999999999999e-05,test,0.4230769230769231,0.06529472327511593,0.409168956043956,0.06741595981671455,0.4194139194139195,0.0648659643165565
+flat_mae,patch,logistic,aabc_age,89,0.005994842503189409,train,0.6456692913385826,0.020597108303879483,0.6411799893613015,0.02112791802960769,0.6469517001701225,0.02054778477592729
+flat_mae,patch,logistic,aabc_age,89,0.005994842503189409,test,0.36538461538461536,0.06330889567974905,0.36523809523809525,0.06335446073116911,0.36469780219780223,0.06316007137707644
+flat_mae,patch,logistic,aabc_age,90,0.3593813663804626,train,0.9763779527559056,0.0067190907712548304,0.9767636745357544,0.006590533987604508,0.9767943176783553,0.0066146771550144885
+flat_mae,patch,logistic,aabc_age,90,0.3593813663804626,test,0.5576923076923077,0.06483902461038442,0.5511001642036124,0.06610021730548786,0.5558608058608059,0.06470996539416102
+flat_mae,patch,logistic,aabc_age,91,0.005994842503189409,train,0.6515748031496063,0.02078358762136675,0.6482322954913878,0.021185465070569478,0.6531500472775605,0.020702216188663792
+flat_mae,patch,logistic,aabc_age,91,0.005994842503189409,test,0.5,0.0690114474682049,0.48875562218890556,0.07052548400517213,0.502518315018315,0.06933149024853995
+flat_mae,patch,logistic,aabc_age,92,0.046415888336127774,train,0.8366141732283464,0.017213568691991053,0.8363365216422284,0.017284848392465666,0.8367008605858581,0.017227081518123688
+flat_mae,patch,logistic,aabc_age,92,0.046415888336127774,test,0.4423076923076923,0.06068426387175513,0.429047619047619,0.05864433060922454,0.440018315018315,0.06037021971931907
+flat_mae,patch,logistic,aabc_age,93,0.046415888336127774,train,0.8503937007874016,0.015749477787917337,0.8508925145060335,0.01580819985686153,0.8510137104925497,0.015784680987908653
+flat_mae,patch,logistic,aabc_age,93,0.046415888336127774,test,0.4230769230769231,0.055069670773060325,0.4018481518481518,0.05030816480280495,0.42651098901098905,0.056241979872844475
+flat_mae,patch,logistic,aabc_age,94,0.046415888336127774,train,0.8464566929133859,0.01551094365855111,0.847043854490663,0.015546579860273689,0.8468314924173698,0.015425055337285597
+flat_mae,patch,logistic,aabc_age,94,0.046415888336127774,test,0.4423076923076923,0.06667827684898806,0.43925925925925924,0.0660750310674081,0.4461996336996337,0.06714379839131301
+flat_mae,patch,logistic,aabc_age,95,0.005994842503189409,train,0.6614173228346457,0.02029038677304085,0.659764679474932,0.02067499530347954,0.6633630312492208,0.020267012650694926
+flat_mae,patch,logistic,aabc_age,95,0.005994842503189409,test,0.4807692307692308,0.05936926864272119,0.4450167887667888,0.05625942854371244,0.4741300366300366,0.058357027881618034
+flat_mae,patch,logistic,aabc_age,96,0.3593813663804626,train,0.9862204724409449,0.005267385498584652,0.9864242341793276,0.005190113121863687,0.9862222591961088,0.0053097431931405405
+flat_mae,patch,logistic,aabc_age,96,0.3593813663804626,test,0.5,0.0666339799059497,0.5007411067193676,0.06746337764280509,0.49702380952380953,0.06696772910592041
+flat_mae,patch,logistic,aabc_age,97,0.000774263682681127,train,0.5649606299212598,0.021254226438153015,0.5606024827719674,0.021795878497464567,0.5655310861947606,0.021299319819808742
+flat_mae,patch,logistic,aabc_age,97,0.000774263682681127,test,0.4423076923076923,0.06360951279756194,0.43199905463806143,0.06659253905666594,0.44024725274725274,0.06332580940705823
+flat_mae,patch,logistic,aabc_age,98,0.005994842503189409,train,0.6614173228346457,0.0205487142906635,0.6595842398225854,0.020924703036748118,0.6634806257899573,0.02059742805569028
+flat_mae,patch,logistic,aabc_age,98,0.005994842503189409,test,0.4230769230769231,0.06037274503611101,0.3986236802413273,0.06471730328954058,0.42376373626373626,0.06077272210040215
+flat_mae,patch,logistic,aabc_age,99,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_age,99,21.54434690031882,test,0.4230769230769231,0.0604779288803511,0.40507700695106996,0.0595019975378678,0.42673992673992667,0.0612278220368281
+flat_mae,patch,logistic,aabc_age,100,0.046415888336127774,train,0.8543307086614174,0.015237253520493446,0.8548547905001331,0.015295170074683243,0.8550459685570657,0.015208881882936889
+flat_mae,patch,logistic,aabc_age,100,0.046415888336127774,test,0.4230769230769231,0.062172824120926,0.4162878787878789,0.06087541953541348,0.42994505494505497,0.06295408741684336

data_scaling/n400_1/eval_v2/aabc_age__patch__logistic/log.txt

@@ -0,0 +1,245 @@
+fMRI foundation model logistic probe eval
+version: 0.1.dev66+g7ddd3aa04
+sha: 58906bf7243fb545e1349221e6921a1797e2e666, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-02-26 17:14:46
+config:
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (aabc_age patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_1/eval_v2/aabc_age__patch__logistic
+model: flat_mae
+representation: patch
+dataset: aabc_age
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/aabc_age__patch__logistic
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=True, reg_tokens=0, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: aabc_age (flat)
+train (n=455):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'segment', 'bold', 'mean', 'std'],
+    num_rows: 471
+}),
+    labels=[0 1 2 3],
+    counts=[110 127 109 109]
+)
+
+validation (n=53):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'segment', 'bold', 'mean', 'std'],
+    num_rows: 58
+}),
+    labels=[0 1 2 3],
+    counts=[14 13 12 14]
+)
+
+test (n=52):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'segment', 'bold', 'mean', 'std'],
+    num_rows: 55
+}),
+    labels=[0 1 2 3],
+    counts=[13 13 12 14]
+)
+
+extracting features for all splits
+extract (train)  [  0/228]  eta: 0:25:57    time: 6.8296  data: 5.6508  max mem: 3205
+extract (train)  [ 20/228]  eta: 0:02:04    time: 0.2894  data: 0.0840  max mem: 3393
+extract (train)  [ 40/228]  eta: 0:01:19    time: 0.2309  data: 0.0582  max mem: 3393
+extract (train)  [ 60/228]  eta: 0:01:01    time: 0.2477  data: 0.0690  max mem: 3393
+extract (train)  [ 80/228]  eta: 0:00:48    time: 0.2062  data: 0.0548  max mem: 3393
+extract (train)  [100/228]  eta: 0:00:38    time: 0.2223  data: 0.0717  max mem: 3393
+extract (train)  [120/228]  eta: 0:00:31    time: 0.2430  data: 0.0775  max mem: 3393
+extract (train)  [140/228]  eta: 0:00:24    time: 0.2038  data: 0.0610  max mem: 3393
+extract (train)  [160/228]  eta: 0:00:18    time: 0.2514  data: 0.0840  max mem: 3393
+extract (train)  [180/228]  eta: 0:00:13    time: 0.2168  data: 0.0721  max mem: 3393
+extract (train)  [200/228]  eta: 0:00:07    time: 0.2066  data: 0.0699  max mem: 3393
+extract (train)  [220/228]  eta: 0:00:02    time: 0.1907  data: 0.0627  max mem: 3393
+extract (train)  [227/228]  eta: 0:00:00    time: 0.1901  data: 0.0632  max mem: 3393
+extract (train) Total time: 0:00:58 (0.2581 s / it)
+extract (validation)  [ 0/27]  eta: 0:02:08    time: 4.7776  data: 4.6140  max mem: 3393
+extract (validation)  [20/27]  eta: 0:00:02    time: 0.1772  data: 0.0503  max mem: 3393
+extract (validation)  [26/27]  eta: 0:00:00    time: 0.1666  data: 0.0494  max mem: 3393
+extract (validation) Total time: 0:00:09 (0.3605 s / it)
+extract (test)  [ 0/26]  eta: 0:02:03    time: 4.7632  data: 4.5802  max mem: 3393
+extract (test)  [20/26]  eta: 0:00:02    time: 0.1666  data: 0.0418  max mem: 3393
+extract (test)  [25/26]  eta: 0:00:00    time: 0.1589  data: 0.0380  max mem: 3393
+extract (test) Total time: 0:00:09 (0.3558 s / it)
+feature extraction time: 0:01:17
+train features: (455, 768)
+validation features: (53, 768)
+test features: (52, 768)
+evaluating fixed splits
+eval results (fixed splits):
+
+| model    | repr   | clf      | dataset   | trial   |        C | split   |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:----------|:--------|---------:|:--------|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | aabc_age  |         | 0.046416 | train   | 0.83661 |  0.017304 | 0.83742 | 0.01727  | 0.83711 |   0.017311 |
+| flat_mae | patch  | logistic | aabc_age  |         | 0.046416 | test    | 0.38462 |  0.057949 | 0.3786  | 0.057977 | 0.37408 |   0.057542 |
+
+
+evaluating random splits (n=100)
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 1, "C": 0.046415888336127774, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.0623871762729618, "f1": 0.5219915848527349, "f1_std": 0.06017802432524102, "bacc": 0.5157967032967032, "bacc_std": 0.062236807096345596}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 2, "C": 9.999999999999999e-05, "split": "test", "acc": 0.46153846153846156, "acc_std": 0.06388899165300599, "f1": 0.426585173193946, "f1_std": 0.06119357747452577, "bacc": 0.4535256410256411, "bacc_std": 0.06262066287810407}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 3, "C": 21.54434690031882, "split": "test", "acc": 0.46153846153846156, "acc_std": 0.06524192575131152, "f1": 0.45466570466570466, "f1_std": 0.06482470264517816, "bacc": 0.45970695970695974, "bacc_std": 0.06498110562727226}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 4, "C": 0.3593813663804626, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.06443055065718564, "f1": 0.5116483516483517, "f1_std": 0.0650331388161421, "bacc": 0.5201465201465201, "bacc_std": 0.06462726599464377}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 5, "C": 2.782559402207126, "split": "test", "acc": 0.4807692307692308, "acc_std": 0.0617952694422784, "f1": 0.47609978588239454, "f1_std": 0.06279175548211148, "bacc": 0.48031135531135527, "bacc_std": 0.06206483155694821}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 6, "C": 0.3593813663804626, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.06371162843338048, "f1": 0.5122053872053872, "f1_std": 0.06416665991261807, "bacc": 0.5249542124542124, "bacc_std": 0.0642360784865812}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 7, "C": 0.000774263682681127, "split": "test", "acc": 0.4807692307692308, "acc_std": 0.0615353845384577, "f1": 0.4376142142062609, "f1_std": 0.06121649682855208, "bacc": 0.4741300366300366, "bacc_std": 0.06066880513261205}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 8, "C": 21.54434690031882, "split": "test", "acc": 0.46153846153846156, "acc_std": 0.06576022205508227, "f1": 0.46042572463768117, "f1_std": 0.06487791290788249, "bacc": 0.46863553113553114, "bacc_std": 0.06635234674296442}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 9, "C": 0.005994842503189409, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.06190242130412678, "f1": 0.5288539553752536, "f1_std": 0.0679387437598958, "bacc": 0.538003663003663, "bacc_std": 0.06219816353545317}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 10, "C": 0.3593813663804626, "split": "test", "acc": 0.5, "acc_std": 0.07216594125294201, "f1": 0.5118248992386923, "f1_std": 0.07041445830854252, "bacc": 0.5027472527472527, "bacc_std": 0.0723557845570276}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 11, "C": 0.046415888336127774, "split": "test", "acc": 0.5576923076923077, "acc_std": 0.06693654815786321, "f1": 0.5756060606060607, "f1_std": 0.06473787371741364, "bacc": 0.5606684981684982, "bacc_std": 0.06697772160222111}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 12, "C": 0.005994842503189409, "split": "test", "acc": 0.46153846153846156, "acc_std": 0.06553687024070302, "f1": 0.45795074812824, "f1_std": 0.06634004442730149, "bacc": 0.4597069597069597, "bacc_std": 0.0654509001386346}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 13, "C": 0.046415888336127774, "split": "test", "acc": 0.38461538461538464, "acc_std": 0.06320731129531956, "f1": 0.38568376068376065, "f1_std": 0.06044551634336766, "bacc": 0.3882783882783883, "bacc_std": 0.06412114375397251}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 14, "C": 0.3593813663804626, "split": "test", "acc": 0.5, "acc_std": 0.06341728060465807, "f1": 0.5011733094491715, "f1_std": 0.06418396716573467, "bacc": 0.5057234432234432, "bacc_std": 0.06352568745341663}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 15, "C": 0.3593813663804626, "split": "test", "acc": 0.5769230769230769, "acc_std": 0.06423075771533752, "f1": 0.5815322580645161, "f1_std": 0.06388026233559048, "bacc": 0.5771520146520146, "bacc_std": 0.06442034689101897}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 16, "C": 0.000774263682681127, "split": "test", "acc": 0.4807692307692308, "acc_std": 0.0633639809579562, "f1": 0.47022546419098143, "f1_std": 0.06343769505593227, "bacc": 0.47870879120879123, "bacc_std": 0.06334870908274497}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 17, "C": 2.782559402207126, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.06329348377801144, "f1": 0.4166269841269841, "f1_std": 0.06275823129090746, "bacc": 0.4237637362637363, "bacc_std": 0.06342433434426233}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 18, "C": 0.3593813663804626, "split": "test", "acc": 0.40384615384615385, "acc_std": 0.06913209848847436, "f1": 0.4096962629796213, "f1_std": 0.06846346696398492, "bacc": 0.40796703296703296, "bacc_std": 0.06957443001465052}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 19, "C": 0.046415888336127774, "split": "test", "acc": 0.4423076923076923, "acc_std": 0.06807665145534687, "f1": 0.4446703296703297, "f1_std": 0.06880843796884462, "bacc": 0.44505494505494503, "bacc_std": 0.06839798620861753}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 20, "C": 2.782559402207126, "split": "test", "acc": 0.5, "acc_std": 0.06821011525049057, "f1": 0.4909688013136289, "f1_std": 0.06848508840388282, "bacc": 0.49793956043956045, "bacc_std": 0.06819440187499051}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 21, "C": 0.3593813663804626, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.05233137190258917, "f1": 0.48503637566137564, "f1_std": 0.04975114814772035, "bacc": 0.5258699633699634, "bacc_std": 0.053755919961926545}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 22, "C": 0.3593813663804626, "split": "test", "acc": 0.4423076923076923, "acc_std": 0.06771820964230939, "f1": 0.45055858120374254, "f1_std": 0.06617821920266428, "bacc": 0.4466575091575092, "bacc_std": 0.06814361301590238}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 23, "C": 0.046415888336127774, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.05982469459044682, "f1": 0.41280786099865047, "f1_std": 0.05780069976389986, "bacc": 0.41941391941391937, "bacc_std": 0.05946350142948324}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 24, "C": 0.046415888336127774, "split": "test", "acc": 0.36538461538461536, "acc_std": 0.060143750382389456, "f1": 0.35896993505689156, "f1_std": 0.05873125603552922, "bacc": 0.36744505494505497, "bacc_std": 0.060525387632811414}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 25, "C": 0.046415888336127774, "split": "test", "acc": 0.3076923076923077, "acc_std": 0.06092573566459571, "f1": 0.32169055082848186, "f1_std": 0.058994593669884424, "bacc": 0.3067765567765568, "bacc_std": 0.060883066869048846}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 26, "C": 0.000774263682681127, "split": "test", "acc": 0.40384615384615385, "acc_std": 0.06300440249480833, "f1": 0.39202551834130783, "f1_std": 0.06404430005433341, "bacc": 0.4001831501831502, "bacc_std": 0.06261948844984895}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 27, "C": 0.046415888336127774, "split": "test", "acc": 0.38461538461538464, "acc_std": 0.06725628806307757, "f1": 0.38421481899742765, "f1_std": 0.06760645536317131, "bacc": 0.38278388278388276, "bacc_std": 0.06702383451020436}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 28, "C": 0.005994842503189409, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.06609267073027351, "f1": 0.5105042016806722, "f1_std": 0.0700139892398969, "bacc": 0.5247252747252747, "bacc_std": 0.06648703176882743}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 29, "C": 0.000774263682681127, "split": "test", "acc": 0.46153846153846156, "acc_std": 0.06464829747846347, "f1": 0.45146520146520147, "f1_std": 0.06577861543241968, "bacc": 0.459478021978022, "bacc_std": 0.06465919222965302}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 30, "C": 0.046415888336127774, "split": "test", "acc": 0.5769230769230769, "acc_std": 0.069163332540063, "f1": 0.5751719576719576, "f1_std": 0.06955112487931323, "bacc": 0.5828754578754578, "bacc_std": 0.06896529519250544}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 31, "C": 0.005994842503189409, "split": "test", "acc": 0.5576923076923077, "acc_std": 0.06704819206489909, "f1": 0.5408740176232436, "f1_std": 0.07130820324176729, "bacc": 0.5558608058608059, "bacc_std": 0.06677979720913636}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 32, "C": 0.3593813663804626, "split": "test", "acc": 0.4423076923076923, "acc_std": 0.06118968470904961, "f1": 0.4282176157176157, "f1_std": 0.06127410700816197, "bacc": 0.44299450549450553, "bacc_std": 0.061556466042595244}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 33, "C": 0.046415888336127774, "split": "test", "acc": 0.5576923076923077, "acc_std": 0.0652738428178778, "f1": 0.5582919254658385, "f1_std": 0.06614300974257094, "bacc": 0.5590659340659341, "bacc_std": 0.06556453681880652}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 34, "C": 0.046415888336127774, "split": "test", "acc": 0.4423076923076923, "acc_std": 0.06530171214805516, "f1": 0.4492713737875028, "f1_std": 0.06484840239157982, "bacc": 0.4432234432234432, "bacc_std": 0.06558736264401135}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 35, "C": 0.005994842503189409, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.06423058498361414, "f1": 0.4195623985522118, "f1_std": 0.06469782093245457, "bacc": 0.4223901098901099, "bacc_std": 0.06419945702259432}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 36, "C": 0.3593813663804626, "split": "test", "acc": 0.4807692307692308, "acc_std": 0.06299219215259419, "f1": 0.4533880237300705, "f1_std": 0.0646290857410857, "bacc": 0.4860347985347986, "bacc_std": 0.06399464873685771}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 37, "C": 166.81005372000556, "split": "test", "acc": 0.38461538461538464, "acc_std": 0.06413009152537719, "f1": 0.38902116402116405, "f1_std": 0.06428738273190888, "bacc": 0.3841575091575091, "bacc_std": 0.06433070209897496}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 38, "C": 0.046415888336127774, "split": "test", "acc": 0.5769230769230769, "acc_std": 0.06492617564606488, "f1": 0.5681697612732095, "f1_std": 0.0663409114958182, "bacc": 0.575091575091575, "bacc_std": 0.06485475122448725}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 39, "C": 2.782559402207126, "split": "test", "acc": 0.4807692307692308, "acc_std": 0.06719661163618755, "f1": 0.4769137866963954, "f1_std": 0.06720643835893196, "bacc": 0.4775641025641026, "bacc_std": 0.06707026983024775}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 40, "C": 0.046415888336127774, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.06683995722363087, "f1": 0.5308862433862434, "f1_std": 0.06828493723383323, "bacc": 0.5396062271062272, "bacc_std": 0.06709156417340942}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 41, "C": 0.3593813663804626, "split": "test", "acc": 0.5, "acc_std": 0.06252781629523685, "f1": 0.5004417434140073, "f1_std": 0.060324350117111, "bacc": 0.5022893772893773, "bacc_std": 0.06288170989868064}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 42, "C": 0.000774263682681127, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.04863485491317966, "f1": 0.38369963369963367, "f1_std": 0.05961748395225562, "bacc": 0.4251373626373626, "bacc_std": 0.04911938812421867}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 43, "C": 0.046415888336127774, "split": "test", "acc": 0.5769230769230769, "acc_std": 0.06724338686386251, "f1": 0.5769047619047619, "f1_std": 0.06881765818112363, "bacc": 0.5771520146520146, "bacc_std": 0.06738253467740424}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 44, "C": 0.046415888336127774, "split": "test", "acc": 0.5, "acc_std": 0.07110550332815524, "f1": 0.5025000000000001, "f1_std": 0.07129035176151845, "bacc": 0.4983974358974359, "bacc_std": 0.07138900894804541}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 45, "C": 0.046415888336127774, "split": "test", "acc": 0.4807692307692308, "acc_std": 0.06487214325435997, "f1": 0.48148693510387663, "f1_std": 0.06353068449861626, "bacc": 0.4862637362637363, "bacc_std": 0.06574287996009155}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 46, "C": 0.000774263682681127, "split": "test", "acc": 0.5769230769230769, "acc_std": 0.06438516680216036, "f1": 0.5743686061381075, "f1_std": 0.07027834743891563, "bacc": 0.5753205128205128, "bacc_std": 0.0645452214856435}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 47, "C": 0.000774263682681127, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.06323548316697633, "f1": 0.4089279895731508, "f1_std": 0.06411366667151348, "bacc": 0.41941391941391937, "bacc_std": 0.06257990943290731}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 48, "C": 0.046415888336127774, "split": "test", "acc": 0.40384615384615385, "acc_std": 0.06832149678598182, "f1": 0.4122980481676134, "f1_std": 0.06558963688299861, "bacc": 0.40613553113553114, "bacc_std": 0.0686484628371634}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 49, "C": 0.005994842503189409, "split": "test", "acc": 0.40384615384615385, "acc_std": 0.0649302197128014, "f1": 0.4029971988795518, "f1_std": 0.06477845406577831, "bacc": 0.40041208791208793, "bacc_std": 0.0647196560149341}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 50, "C": 0.3593813663804626, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.06745526679030689, "f1": 0.5433214882943144, "f1_std": 0.06734358620754474, "bacc": 0.5384615384615385, "bacc_std": 0.06762030133758463}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 51, "C": 0.046415888336127774, "split": "test", "acc": 0.5769230769230769, "acc_std": 0.06488036327319337, "f1": 0.5727609427609428, "f1_std": 0.06440225643183178, "bacc": 0.5737179487179487, "bacc_std": 0.06493301202675035}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 52, "C": 0.000774263682681127, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.06208901559847898, "f1": 0.4263083735909823, "f1_std": 0.06099594284855781, "bacc": 0.4269688644688645, "bacc_std": 0.06277916775184823}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 53, "C": 9.999999999999999e-05, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.058336150960409956, "f1": 0.4778935185185185, "f1_std": 0.056919293573939544, "bacc": 0.5112179487179487, "bacc_std": 0.057354562492450176}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 54, "C": 0.046415888336127774, "split": "test", "acc": 0.40384615384615385, "acc_std": 0.060034248115101455, "f1": 0.37649262413327744, "f1_std": 0.05826356801998579, "bacc": 0.41048534798534797, "bacc_std": 0.06119411094178349}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 55, "C": 166.81005372000556, "split": "test", "acc": 0.4807692307692308, "acc_std": 0.058148249628954804, "f1": 0.4733514492753623, "f1_std": 0.054474379757819845, "bacc": 0.48901098901098905, "bacc_std": 0.05936499621774483}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 56, "C": 0.000774263682681127, "split": "test", "acc": 0.40384615384615385, "acc_std": 0.06304242724654639, "f1": 0.39622153209109734, "f1_std": 0.06076532629668424, "bacc": 0.4015567765567766, "bacc_std": 0.0628410538051134}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 57, "C": 0.046415888336127774, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.0574054921810133, "f1": 0.49670506912442397, "f1_std": 0.05769100681961992, "bacc": 0.5155677655677655, "bacc_std": 0.05699896632817456}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 58, "C": 0.3593813663804626, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.05967917182984243, "f1": 0.40654511453950437, "f1_std": 0.06115877536204072, "bacc": 0.4207875457875458, "bacc_std": 0.05927754848986136}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 59, "C": 0.000774263682681127, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.06252805287588688, "f1": 0.49991466120498373, "f1_std": 0.06623523954461033, "bacc": 0.5157967032967034, "bacc_std": 0.06215046603781298}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 60, "C": 0.005994842503189409, "split": "test", "acc": 0.4423076923076923, "acc_std": 0.05898123009024635, "f1": 0.4291287565481114, "f1_std": 0.06176947870265391, "bacc": 0.44459706959706957, "bacc_std": 0.059268388111311805}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 61, "C": 0.3593813663804626, "split": "test", "acc": 0.3269230769230769, "acc_std": 0.05999924802290117, "f1": 0.32605377714073364, "f1_std": 0.059493436151958734, "bacc": 0.3276098901098901, "bacc_std": 0.06022930518905783}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 62, "C": 0.046415888336127774, "split": "test", "acc": 0.46153846153846156, "acc_std": 0.06850689940635314, "f1": 0.46724310311266837, "f1_std": 0.06788405332706389, "bacc": 0.4626831501831502, "bacc_std": 0.06884036815356925}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 63, "C": 0.005994842503189409, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.057567698762660324, "f1": 0.41338566827697265, "f1_std": 0.05654634466025146, "bacc": 0.42078754578754574, "bacc_std": 0.05736800421015458}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 64, "C": 0.3593813663804626, "split": "test", "acc": 0.40384615384615385, "acc_std": 0.06583110000456206, "f1": 0.4105528127267257, "f1_std": 0.0649950965697301, "bacc": 0.40476190476190477, "bacc_std": 0.0660530661722229}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 65, "C": 0.005994842503189409, "split": "test", "acc": 0.5769230769230769, "acc_std": 0.06426212960815372, "f1": 0.5621632996632997, "f1_std": 0.06409258973084567, "bacc": 0.5707417582417582, "bacc_std": 0.06371763986513436}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 66, "C": 0.005994842503189409, "split": "test", "acc": 0.4807692307692308, "acc_std": 0.06016132158941171, "f1": 0.4596607566662013, "f1_std": 0.06068762171743191, "bacc": 0.4757326007326007, "bacc_std": 0.05987106851884831}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 67, "C": 0.046415888336127774, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.06417316624783438, "f1": 0.5067438055165966, "f1_std": 0.06537415870734964, "bacc": 0.5144230769230769, "bacc_std": 0.06381588776883253}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 68, "C": 0.046415888336127774, "split": "test", "acc": 0.46153846153846156, "acc_std": 0.06188234446280311, "f1": 0.4398809523809524, "f1_std": 0.06553892205323958, "bacc": 0.4578754578754579, "bacc_std": 0.06168321469782148}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 69, "C": 0.3593813663804626, "split": "test", "acc": 0.40384615384615385, "acc_std": 0.06426305038644224, "f1": 0.3963817563922753, "f1_std": 0.06504523412636465, "bacc": 0.40773809523809523, "bacc_std": 0.06493784273981329}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 70, "C": 2.782559402207126, "split": "test", "acc": 0.38461538461538464, "acc_std": 0.062168065300403766, "f1": 0.3762747668997669, "f1_std": 0.06254008406954978, "bacc": 0.38690476190476186, "bacc_std": 0.06265227732451385}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 71, "C": 0.000774263682681127, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.06314414804202098, "f1": 0.5127731463938361, "f1_std": 0.06407723343517392, "bacc": 0.5187728937728937, "bacc_std": 0.06344692866309007}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 72, "C": 0.046415888336127774, "split": "test", "acc": 0.5769230769230769, "acc_std": 0.06315197223866517, "f1": 0.5670588235294118, "f1_std": 0.0659766853724605, "bacc": 0.5737179487179487, "bacc_std": 0.06326882135573896}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 73, "C": 0.046415888336127774, "split": "test", "acc": 0.5, "acc_std": 0.06387463451880757, "f1": 0.5131649831649832, "f1_std": 0.060681143120914265, "bacc": 0.4981684981684981, "bacc_std": 0.06381269474662798}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 74, "C": 0.005994842503189409, "split": "test", "acc": 0.46153846153846156, "acc_std": 0.06643343200207232, "f1": 0.4487179487179487, "f1_std": 0.06994467821680307, "bacc": 0.459478021978022, "bacc_std": 0.06653138472875025}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 75, "C": 0.005994842503189409, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.07084296952670001, "f1": 0.5222355488922206, "f1_std": 0.07127761237252711, "bacc": 0.5192307692307693, "bacc_std": 0.07084768210974748}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 76, "C": 0.046415888336127774, "split": "test", "acc": 0.5, "acc_std": 0.06617507485862315, "f1": 0.4999823165340407, "f1_std": 0.06630484056353396, "bacc": 0.5041208791208791, "bacc_std": 0.06666621260109828}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 77, "C": 0.3593813663804626, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.060222209333778344, "f1": 0.5068734015345269, "f1_std": 0.06189476526915229, "bacc": 0.5231227106227107, "bacc_std": 0.06061078712711297}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 78, "C": 0.046415888336127774, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.06728517205668887, "f1": 0.5408772262220538, "f1_std": 0.0679401843207676, "bacc": 0.5412087912087912, "bacc_std": 0.06763574196912679}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 79, "C": 0.005994842503189409, "split": "test", "acc": 0.5, "acc_std": 0.06592128172764379, "f1": 0.49559419075548106, "f1_std": 0.06713670130154296, "bacc": 0.4967948717948718, "bacc_std": 0.06592070603861995}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 80, "C": 0.3593813663804626, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.05982340876852798, "f1": 0.5142857142857142, "f1_std": 0.05976425329194862, "bacc": 0.5485347985347985, "bacc_std": 0.060826067221274344}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 81, "C": 21.54434690031882, "split": "test", "acc": 0.4807692307692308, "acc_std": 0.06616140386166806, "f1": 0.49083639998182726, "f1_std": 0.0655279966055778, "bacc": 0.48489010989010983, "bacc_std": 0.06648539090209249}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 82, "C": 0.3593813663804626, "split": "test", "acc": 0.5576923076923077, "acc_std": 0.06565176194249356, "f1": 0.5509476031215161, "f1_std": 0.06716327472566319, "bacc": 0.5604395604395604, "bacc_std": 0.06582553511038909}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 83, "C": 0.3593813663804626, "split": "test", "acc": 0.40384615384615385, "acc_std": 0.062201369398346744, "f1": 0.39993961352657004, "f1_std": 0.0607564887485186, "bacc": 0.40453296703296704, "bacc_std": 0.062312736109117806}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 84, "C": 0.3593813663804626, "split": "test", "acc": 0.38461538461538464, "acc_std": 0.06326046233370854, "f1": 0.37483766233766236, "f1_std": 0.06495607335268341, "bacc": 0.3869047619047619, "bacc_std": 0.06373309536986203}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 85, "C": 9.999999999999999e-05, "split": "test", "acc": 0.46153846153846156, "acc_std": 0.053040943769660925, "f1": 0.3851662404092072, "f1_std": 0.04907110049187005, "bacc": 0.4519230769230769, "bacc_std": 0.0511340841744937}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 86, "C": 0.3593813663804626, "split": "test", "acc": 0.4423076923076923, "acc_std": 0.061233571640546894, "f1": 0.42765567765567764, "f1_std": 0.06052194979561253, "bacc": 0.43727106227106227, "bacc_std": 0.06066059562879579}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 87, "C": 0.000774263682681127, "split": "test", "acc": 0.5192307692307693, "acc_std": 0.06085460166970694, "f1": 0.5059791758161323, "f1_std": 0.06262279436498262, "bacc": 0.5247252747252747, "bacc_std": 0.061538786983484466}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 88, "C": 9.999999999999999e-05, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.06529472327511593, "f1": 0.409168956043956, "f1_std": 0.06741595981671455, "bacc": 0.4194139194139195, "bacc_std": 0.0648659643165565}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 89, "C": 0.005994842503189409, "split": "test", "acc": 0.36538461538461536, "acc_std": 0.06330889567974905, "f1": 0.36523809523809525, "f1_std": 0.06335446073116911, "bacc": 0.36469780219780223, "bacc_std": 0.06316007137707644}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 90, "C": 0.3593813663804626, "split": "test", "acc": 0.5576923076923077, "acc_std": 0.06483902461038442, "f1": 0.5511001642036124, "f1_std": 0.06610021730548786, "bacc": 0.5558608058608059, "bacc_std": 0.06470996539416102}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 91, "C": 0.005994842503189409, "split": "test", "acc": 0.5, "acc_std": 0.0690114474682049, "f1": 0.48875562218890556, "f1_std": 0.07052548400517213, "bacc": 0.502518315018315, "bacc_std": 0.06933149024853995}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 92, "C": 0.046415888336127774, "split": "test", "acc": 0.4423076923076923, "acc_std": 0.06068426387175513, "f1": 0.429047619047619, "f1_std": 0.05864433060922454, "bacc": 0.440018315018315, "bacc_std": 0.06037021971931907}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 93, "C": 0.046415888336127774, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.055069670773060325, "f1": 0.4018481518481518, "f1_std": 0.05030816480280495, "bacc": 0.42651098901098905, "bacc_std": 0.056241979872844475}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 94, "C": 0.046415888336127774, "split": "test", "acc": 0.4423076923076923, "acc_std": 0.06667827684898806, "f1": 0.43925925925925924, "f1_std": 0.0660750310674081, "bacc": 0.4461996336996337, "bacc_std": 0.06714379839131301}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 95, "C": 0.005994842503189409, "split": "test", "acc": 0.4807692307692308, "acc_std": 0.05936926864272119, "f1": 0.4450167887667888, "f1_std": 0.05625942854371244, "bacc": 0.4741300366300366, "bacc_std": 0.058357027881618034}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 96, "C": 0.3593813663804626, "split": "test", "acc": 0.5, "acc_std": 0.0666339799059497, "f1": 0.5007411067193676, "f1_std": 0.06746337764280509, "bacc": 0.49702380952380953, "bacc_std": 0.06696772910592041}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 97, "C": 0.000774263682681127, "split": "test", "acc": 0.4423076923076923, "acc_std": 0.06360951279756194, "f1": 0.43199905463806143, "f1_std": 0.06659253905666594, "bacc": 0.44024725274725274, "bacc_std": 0.06332580940705823}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 98, "C": 0.005994842503189409, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.06037274503611101, "f1": 0.3986236802413273, "f1_std": 0.06471730328954058, "bacc": 0.42376373626373626, "bacc_std": 0.06077272210040215}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 99, "C": 21.54434690031882, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.0604779288803511, "f1": 0.40507700695106996, "f1_std": 0.0595019975378678, "bacc": 0.42673992673992667, "bacc_std": 0.0612278220368281}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_age", "trial": 100, "C": 0.046415888336127774, "split": "test", "acc": 0.4230769230769231, "acc_std": 0.062172824120926, "f1": 0.4162878787878789, "f1_std": 0.06087541953541348, "bacc": 0.42994505494505497, "bacc_std": 0.06295408741684336}
+eval results (random splits):
+
+| model    | repr   | clf      | dataset   | split   |   n_trials |      C |   C_std |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:----------|:--------|-----------:|-------:|--------:|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | aabc_age  | train   |        100 | 4.4389 |  23.694 | 0.81199 |  0.16843  | 0.80964 | 0.17237  | 0.81261 |   0.16822  |
+| flat_mae | patch  | logistic | aabc_age  | test    |        100 | 4.4389 |  23.694 | 0.47212 |  0.061569 | 0.46373 | 0.062725 | 0.47234 |   0.061544 |
+
+
+done! total time: 0:05:47

data_scaling/n400_1/eval_v2/aabc_sex__patch__logistic/config.yaml

@@ -0,0 +1,30 @@
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (aabc_sex patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_1/eval_v2/aabc_sex__patch__logistic
+model: flat_mae
+representation: patch
+dataset: aabc_sex
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/aabc_sex__patch__logistic
+remote_dir: null

data_scaling/n400_1/eval_v2/aabc_sex__patch__logistic/eval_table.csv

@@ -0,0 +1,203 @@
+model,repr,clf,dataset,trial,C,split,acc,acc_std,f1,f1_std,bacc,bacc_std
+flat_mae,patch,logistic,aabc_sex,,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,,2.782559402207126,test,0.8909090909090909,0.04386321337525678,0.8891129032258065,0.04392565910285936,0.9015151515151516,0.040231321487743096
+flat_mae,patch,logistic,aabc_sex,1,0.3593813663804626,train,0.9829867674858223,0.006411646870366585,0.9825679104559584,0.0065669777000106466,0.9828614554940063,0.006536396026268371
+flat_mae,patch,logistic,aabc_sex,1,0.3593813663804626,test,0.8181818181818182,0.05265588898336156,0.8166666666666667,0.05265479976916016,0.8254076086956521,0.05117159420062498
+flat_mae,patch,logistic,aabc_sex,2,0.046415888336127774,train,0.9319470699432892,0.01086034226940898,0.929871851524525,0.011259943188207838,0.9277968287464463,0.011680840048076627
+flat_mae,patch,logistic,aabc_sex,2,0.046415888336127774,test,0.8363636363636363,0.051022263816481866,0.8328267477203647,0.052100542340955486,0.8349184782608696,0.052040162627840746
+flat_mae,patch,logistic,aabc_sex,3,0.046415888336127774,train,0.9319470699432892,0.011067920394260862,0.9301434985474073,0.011376229768835774,0.9296213253612357,0.011579021635585057
+flat_mae,patch,logistic,aabc_sex,3,0.046415888336127774,test,0.7818181818181819,0.0554105458829032,0.7727272727272727,0.05863437044538351,0.7697010869565217,0.05848903869876677
+flat_mae,patch,logistic,aabc_sex,4,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,4,166.81005372000556,test,0.8181818181818182,0.050275682964238864,0.8166666666666667,0.05041723963269043,0.8254076086956521,0.049211660530014
+flat_mae,patch,logistic,aabc_sex,5,0.3593813663804626,train,0.9848771266540642,0.005187503776616098,0.9845140515222482,0.005302540181286461,0.9851036079603739,0.005121918594376673
+flat_mae,patch,logistic,aabc_sex,5,0.3593813663804626,test,0.8,0.05273855678629085,0.790003471017008,0.056230055783621925,0.7853260869565217,0.05558520219851598
+flat_mae,patch,logistic,aabc_sex,6,0.046415888336127774,train,0.9224952741020794,0.012217292881259117,0.9201850291269996,0.012659539177079727,0.918410563029397,0.013071585646073917
+flat_mae,patch,logistic,aabc_sex,6,0.046415888336127774,test,0.9090909090909091,0.03693972754466904,0.9045470322804582,0.03992807000558217,0.8974184782608696,0.04171778202044614
+flat_mae,patch,logistic,aabc_sex,7,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,7,2.782559402207126,test,0.8545454545454545,0.04707285017014047,0.8521505376344086,0.04774533501656769,0.8566576086956521,0.047426358897281806
+flat_mae,patch,logistic,aabc_sex,8,0.046415888336127774,train,0.9338374291115312,0.010641879984276134,0.9320413294426397,0.01093859371464472,0.9312553122893402,0.011089091819029057
+flat_mae,patch,logistic,aabc_sex,8,0.046415888336127774,test,0.8,0.05253974174669985,0.795677136102668,0.053761764767330654,0.7975543478260869,0.05411603893049472
+flat_mae,patch,logistic,aabc_sex,9,0.046415888336127774,train,0.9281663516068053,0.011227603005859503,0.9261694188164776,0.011568458626444254,0.9251370204285002,0.011798028807798027
+flat_mae,patch,logistic,aabc_sex,9,0.046415888336127774,test,0.8727272727272727,0.046299049487771055,0.8683760683760684,0.048387140550290475,0.8661684782608696,0.04900389469866049
+flat_mae,patch,logistic,aabc_sex,10,0.046415888336127774,train,0.9281663516068053,0.011193308769022714,0.9260738452486026,0.011581707043272778,0.9245288548902371,0.011932826505040623
+flat_mae,patch,logistic,aabc_sex,10,0.046415888336127774,test,0.8545454545454545,0.048698770497147895,0.8505434782608696,0.050083757649930345,0.8505434782608696,0.050190351568757355
+flat_mae,patch,logistic,aabc_sex,11,0.005994842503189409,train,0.8790170132325141,0.01374664196089188,0.8753277360435999,0.014250556812847759,0.8735308772238342,0.014485665767677215
+flat_mae,patch,logistic,aabc_sex,11,0.005994842503189409,test,0.9090909090909091,0.03667531878885939,0.9027925061859314,0.04190960075868798,0.8913043478260869,0.04385092463885362
+flat_mae,patch,logistic,aabc_sex,12,0.3593813663804626,train,0.9867674858223062,0.005020354081444497,0.9864417081324122,0.005141133475496999,0.9867375948884786,0.005112195158987262
+flat_mae,patch,logistic,aabc_sex,12,0.3593813663804626,test,0.8,0.05178686134787916,0.7861435136090491,0.057965069555226255,0.7792119565217391,0.05650578427288931
+flat_mae,patch,logistic,aabc_sex,13,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,13,166.81005372000556,test,0.8909090909090909,0.041913551330374846,0.8879076086956521,0.04308790390649044,0.8879076086956521,0.04317491706667154
+flat_mae,patch,logistic,aabc_sex,14,0.3593813663804626,train,0.9829867674858223,0.005482542184301136,0.9825679104559584,0.005615199175599613,0.9828614554940063,0.00560513262947946
+flat_mae,patch,logistic,aabc_sex,14,0.3593813663804626,test,0.8909090909090909,0.03883382695154418,0.8879076086956521,0.03999668307543158,0.8879076086956521,0.04032041382000847
+flat_mae,patch,logistic,aabc_sex,15,0.046415888336127774,train,0.9357277882797732,0.010606162295353015,0.9339410589410589,0.010951724833190065,0.9328892992174448,0.011283099850146543
+flat_mae,patch,logistic,aabc_sex,15,0.046415888336127774,test,0.7818181818181819,0.05706815991455355,0.7758152173913043,0.058976953550242794,0.7758152173913043,0.059001860339285686
+flat_mae,patch,logistic,aabc_sex,16,0.3593813663804626,train,0.9829867674858223,0.005437009488705511,0.9825885657235016,0.0055536891560321495,0.9834696210322693,0.005368527684275893
+flat_mae,patch,logistic,aabc_sex,16,0.3593813663804626,test,0.8363636363636363,0.051427027656220974,0.8307692307692308,0.05383410687411033,0.8288043478260869,0.054278706998561854
+flat_mae,patch,logistic,aabc_sex,17,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,17,2.782559402207126,test,0.8,0.051782839622203425,0.790003471017008,0.0555870379546963,0.7853260869565217,0.05496704207452725
+flat_mae,patch,logistic,aabc_sex,18,0.046415888336127774,train,0.9262759924385633,0.011199720654216351,0.9243690085598548,0.01149185670707677,0.9241111990386588,0.01162629757108021
+flat_mae,patch,logistic,aabc_sex,18,0.046415888336127774,test,0.8181818181818182,0.05072647767379556,0.8176392572944298,0.05062105628133805,0.8315217391304348,0.047879830440684626
+flat_mae,patch,logistic,aabc_sex,19,0.3593813663804626,train,0.9810964083175804,0.005691982892131781,0.9806193030276386,0.005839464865998627,0.9806193030276386,0.005975351154701763
+flat_mae,patch,logistic,aabc_sex,19,0.3593813663804626,test,0.7818181818181819,0.054926378976207925,0.7727272727272727,0.05772363026484643,0.7697010869565217,0.0571900716828565
+flat_mae,patch,logistic,aabc_sex,20,0.046415888336127774,train,0.9281663516068053,0.011587251621567082,0.9260738452486026,0.011993725891308314,0.9245288548902371,0.012375563916449365
+flat_mae,patch,logistic,aabc_sex,20,0.046415888336127774,test,0.8363636363636363,0.04610408633580652,0.8250265111346766,0.05230800624333397,0.8165760869565217,0.051851235104635476
+flat_mae,patch,logistic,aabc_sex,21,0.046415888336127774,train,0.9281663516068053,0.011182620940990545,0.9259758432758874,0.011577964166298232,0.923920689351974,0.01188135540486852
+flat_mae,patch,logistic,aabc_sex,21,0.046415888336127774,test,0.7818181818181819,0.052523229020851894,0.7727272727272727,0.05581268638224436,0.7697010869565217,0.055373226476392326
+flat_mae,patch,logistic,aabc_sex,22,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,22,21.54434690031882,test,0.8363636363636363,0.0519799579939985,0.8328267477203647,0.053059633258707335,0.8349184782608696,0.05295779550098547
+flat_mae,patch,logistic,aabc_sex,23,0.046415888336127774,train,0.9206049149338374,0.011682282011412184,0.9182921447484554,0.012081929223847643,0.9167765761012925,0.012419142391638986
+flat_mae,patch,logistic,aabc_sex,23,0.046415888336127774,test,0.8727272727272727,0.046282938683959336,0.8699763593380614,0.04729635012452601,0.8722826086956521,0.04700199204630469
+flat_mae,patch,logistic,aabc_sex,24,0.3593813663804626,train,0.9773156899810964,0.006661942191693025,0.9767431636331663,0.006832896592937214,0.9767431636331663,0.00691825329498112
+flat_mae,patch,logistic,aabc_sex,24,0.3593813663804626,test,0.8727272727272727,0.04799629462282125,0.8699763593380614,0.04910219932790519,0.8722826086956521,0.04885495493974446
+flat_mae,patch,logistic,aabc_sex,25,0.3593813663804626,train,0.9829867674858223,0.0054636534054791465,0.9825885657235016,0.005579542161040209,0.9834696210322693,0.005346358469112914
+flat_mae,patch,logistic,aabc_sex,25,0.3593813663804626,test,0.8181818181818182,0.05094568814429603,0.8151881720430108,0.05180721624321444,0.8192934782608696,0.05138047076561111
+flat_mae,patch,logistic,aabc_sex,26,0.046415888336127774,train,0.9357277882797732,0.01051201793541408,0.9339410589410589,0.010807699747920707,0.9328892992174448,0.010931716798019658
+flat_mae,patch,logistic,aabc_sex,26,0.046415888336127774,test,0.7818181818181819,0.05115633468606205,0.7727272727272727,0.05421765843013767,0.7697010869565217,0.053793701941407264
+flat_mae,patch,logistic,aabc_sex,27,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,27,2.782559402207126,test,0.8727272727272727,0.04253984422993856,0.8699763593380614,0.04326416544506314,0.8722826086956521,0.042897966358643964
+flat_mae,patch,logistic,aabc_sex,28,0.005994842503189409,train,0.8846880907372401,0.013423413699526822,0.8812508969938286,0.013929901414925486,0.8796491690846742,0.014290777024908783
+flat_mae,patch,logistic,aabc_sex,28,0.005994842503189409,test,0.8,0.05296143614941217,0.790003471017008,0.057136675986417966,0.7853260869565217,0.05670747727149538
+flat_mae,patch,logistic,aabc_sex,29,0.3593813663804626,train,0.9848771266540642,0.005168698287105385,0.9845140515222482,0.005284565854105462,0.9851036079603739,0.005127239953298392
+flat_mae,patch,logistic,aabc_sex,29,0.3593813663804626,test,0.8363636363636363,0.04905835957946512,0.8307692307692308,0.051178039635536604,0.8288043478260869,0.05126013007047176
+flat_mae,patch,logistic,aabc_sex,30,0.046415888336127774,train,0.9224952741020794,0.011869800493876193,0.9201850291269996,0.012283679653695662,0.918410563029397,0.01258180560817222
+flat_mae,patch,logistic,aabc_sex,30,0.046415888336127774,test,0.7636363636363637,0.059432144793123046,0.7555555555555555,0.061952149867716945,0.7540760869565217,0.061502168699851326
+flat_mae,patch,logistic,aabc_sex,31,0.046415888336127774,train,0.9262759924385633,0.010994583740998246,0.9241777748376498,0.011362782928780985,0.9228948679621325,0.011686003205109577
+flat_mae,patch,logistic,aabc_sex,31,0.046415888336127774,test,0.7818181818181819,0.05530347231039635,0.7782258064516129,0.055718803885056456,0.7819293478260869,0.055179640142934576
+flat_mae,patch,logistic,aabc_sex,32,0.046415888336127774,train,0.9243856332703214,0.011819434381201154,0.9223816650526748,0.012158151141020557,0.9218690465722911,0.012359061931176817
+flat_mae,patch,logistic,aabc_sex,32,0.046415888336127774,test,0.8727272727272727,0.04434598754758142,0.8683760683760684,0.046439782848170005,0.8661684782608696,0.04699345640291025
+flat_mae,patch,logistic,aabc_sex,33,0.046415888336127774,train,0.9206049149338374,0.01109500350179081,0.9185007483053085,0.011391645868646123,0.9179929071778188,0.011494820734193207
+flat_mae,patch,logistic,aabc_sex,33,0.046415888336127774,test,0.8545454545454545,0.045888820007196636,0.8505434782608696,0.047194520021074556,0.8505434782608696,0.04709612353182535
+flat_mae,patch,logistic,aabc_sex,34,0.3593813663804626,train,0.9792060491493384,0.006539549937667335,0.9787439225298349,0.006664929886575387,0.9802016471760602,0.006298504167404621
+flat_mae,patch,logistic,aabc_sex,34,0.3593813663804626,test,0.8363636363636363,0.049449344646318734,0.8307692307692308,0.05149415963006614,0.8288043478260869,0.0517093403774025
+flat_mae,patch,logistic,aabc_sex,35,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,35,2.782559402207126,test,0.8727272727272727,0.043602577829624116,0.8699763593380614,0.044378928880580606,0.8722826086956521,0.04402821761910765
+flat_mae,patch,logistic,aabc_sex,36,0.3593813663804626,train,0.9792060491493384,0.005942495509706932,0.9787193581065019,0.0060715560203381275,0.9795934816377971,0.005928774749430285
+flat_mae,patch,logistic,aabc_sex,36,0.3593813663804626,test,0.8363636363636363,0.04887976139856006,0.8328267477203647,0.04978049940641453,0.8349184782608696,0.04963764381180495
+flat_mae,patch,logistic,aabc_sex,37,0.3593813663804626,train,0.9829867674858223,0.005662676827736662,0.9825466942830434,0.005807475308733624,0.9822532899557432,0.005859545146694079
+flat_mae,patch,logistic,aabc_sex,37,0.3593813663804626,test,0.8181818181818182,0.052573267253919684,0.8151881720430108,0.053486615006010346,0.8192934782608696,0.053490187548087584
+flat_mae,patch,logistic,aabc_sex,38,0.046415888336127774,train,0.9319470699432892,0.01107592458261744,0.9302294908994988,0.011365434637962543,0.9302294908994988,0.011496074121349422
+flat_mae,patch,logistic,aabc_sex,38,0.046415888336127774,test,0.8727272727272727,0.04203854980159861,0.8639095086603039,0.04749952680671482,0.8539402173913043,0.04816852800661158
+flat_mae,patch,logistic,aabc_sex,39,0.046415888336127774,train,0.9262759924385633,0.011081137103900445,0.9240784423403167,0.01147087894886883,0.9222867024238695,0.011791069819642274
+flat_mae,patch,logistic,aabc_sex,39,0.046415888336127774,test,0.8545454545454545,0.04533743459004501,0.8505434782608696,0.046700870489602,0.8505434782608696,0.046837935783746884
+flat_mae,patch,logistic,aabc_sex,40,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,40,2.782559402207126,test,0.8181818181818182,0.049738489677229825,0.8131793478260869,0.05141701602857101,0.8131793478260869,0.05157697266027768
+flat_mae,patch,logistic,aabc_sex,41,0.005994842503189409,train,0.8752362948960303,0.014227651867596825,0.8712572642260834,0.014776466713096511,0.8690465722910987,0.015020895367525319
+flat_mae,patch,logistic,aabc_sex,41,0.005994842503189409,test,0.9090909090909091,0.038480287966483194,0.905982905982906,0.040362495966507414,0.9035326086956521,0.04107032746648314
+flat_mae,patch,logistic,aabc_sex,42,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,42,166.81005372000556,test,0.7818181818181819,0.055751941744341615,0.7782258064516129,0.05660795085808481,0.7819293478260869,0.05630541567402165
+flat_mae,patch,logistic,aabc_sex,43,0.046415888336127774,train,0.9224952741020794,0.011731038978893372,0.920289455598555,0.012104568368655253,0.9190187285676603,0.012398425713796818
+flat_mae,patch,logistic,aabc_sex,43,0.046415888336127774,test,0.8545454545454545,0.0439057743900889,0.8521505376344086,0.04445632866046411,0.8566576086956521,0.04384902873841091
+flat_mae,patch,logistic,aabc_sex,44,0.046415888336127774,train,0.9357277882797732,0.010737023747121861,0.9337678597731625,0.011125548686214466,0.9316729681409186,0.011524648482303512
+flat_mae,patch,logistic,aabc_sex,44,0.046415888336127774,test,0.8,0.05146166349737786,0.790003471017008,0.0554262734827424,0.7853260869565217,0.05497276355871275
+flat_mae,patch,logistic,aabc_sex,45,0.046415888336127774,train,0.9187145557655955,0.012329651719900681,0.916611983796763,0.012652940681382932,0.9163589202497142,0.012767122775365512
+flat_mae,patch,logistic,aabc_sex,45,0.046415888336127774,test,0.9272727272727272,0.034506216950927315,0.9266666666666667,0.03439735271827876,0.9375,0.029653780192203154
+flat_mae,patch,logistic,aabc_sex,46,0.046415888336127774,train,0.9224952741020794,0.011827916136415323,0.920289455598555,0.012206163443696166,0.9190187285676603,0.012444793296179587
+flat_mae,patch,logistic,aabc_sex,46,0.046415888336127774,test,0.9090909090909091,0.0393400097055957,0.9071259709557582,0.0400938976349243,0.9096467391304348,0.03956758514624431
+flat_mae,patch,logistic,aabc_sex,47,0.3593813663804626,train,0.9829867674858223,0.005712427699137532,0.9825885657235016,0.005834601586498123,0.9834696210322693,0.0056027681443567914
+flat_mae,patch,logistic,aabc_sex,47,0.3593813663804626,test,0.8909090909090909,0.04032423957771133,0.8863636363636364,0.042786489845962436,0.8817934782608696,0.04366744834135569
+flat_mae,patch,logistic,aabc_sex,48,0.005994842503189409,train,0.8752362948960303,0.014022730180077377,0.8714317277949624,0.01451698768635156,0.869654737829362,0.014751154643700679
+flat_mae,patch,logistic,aabc_sex,48,0.005994842503189409,test,0.8909090909090909,0.04058312151703114,0.8879076086956521,0.041885923840686734,0.8879076086956521,0.04216694092273785
+flat_mae,patch,logistic,aabc_sex,49,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,49,166.81005372000556,test,0.8363636363636363,0.047862060752793846,0.8343927735028438,0.04817250193499985,0.8410326086956521,0.04748820110383062
+flat_mae,patch,logistic,aabc_sex,50,0.046415888336127774,train,0.9224952741020794,0.01128131065195158,0.9201850291269996,0.01166506840396556,0.918410563029397,0.011942675922216046
+flat_mae,patch,logistic,aabc_sex,50,0.046415888336127774,test,0.8545454545454545,0.04820247927507941,0.8484848484848485,0.0507837081955852,0.8444293478260869,0.05090977446513677
+flat_mae,patch,logistic,aabc_sex,51,0.046415888336127774,train,0.9130434782608695,0.011595257890140247,0.910738914810576,0.0119401649416894,0.9102406283888742,0.012186540766105692
+flat_mae,patch,logistic,aabc_sex,51,0.046415888336127774,test,0.8181818181818182,0.04529986773191422,0.8074229691876751,0.05073726921765426,0.8009510869565217,0.050273032794018735
+flat_mae,patch,logistic,aabc_sex,52,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,52,21.54434690031882,test,0.8363636363636363,0.04892133675690641,0.8307692307692308,0.051052479399323294,0.8288043478260869,0.05116841613459943
+flat_mae,patch,logistic,aabc_sex,53,0.3593813663804626,train,0.9810964083175804,0.005998668143578677,0.9806193030276386,0.0061502017493534735,0.9806193030276386,0.00622021430283144
+flat_mae,patch,logistic,aabc_sex,53,0.3593813663804626,test,0.8545454545454545,0.049437483687870154,0.8521505376344086,0.05005743999904591,0.8566576086956521,0.049228255619581746
+flat_mae,patch,logistic,aabc_sex,54,0.005994842503189409,train,0.8809073724007561,0.014199795752770005,0.8770244091437427,0.014763356969062115,0.8745566986136757,0.015012534847057559
+flat_mae,patch,logistic,aabc_sex,54,0.005994842503189409,test,0.8545454545454545,0.047307106380010025,0.8521505376344086,0.04786339862693188,0.8566576086956521,0.04702470759122328
+flat_mae,patch,logistic,aabc_sex,55,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,55,21.54434690031882,test,0.8,0.053532546676942294,0.7931623931623932,0.05615577862990481,0.7914402173913043,0.05628428399847978
+flat_mae,patch,logistic,aabc_sex,56,0.046415888336127774,train,0.9300567107750473,0.01153923980765489,0.9281579768393621,0.01187831115229994,0.9273791728948679,0.012100881028858139
+flat_mae,patch,logistic,aabc_sex,56,0.046415888336127774,test,0.8363636363636363,0.048303106345706434,0.8307692307692308,0.050494865110545385,0.8288043478260869,0.05078730095183825
+flat_mae,patch,logistic,aabc_sex,57,0.046415888336127774,train,0.9262759924385633,0.010633434151226888,0.9241777748376498,0.010966500221882669,0.9228948679621325,0.011173700982284691
+flat_mae,patch,logistic,aabc_sex,57,0.046415888336127774,test,0.8909090909090909,0.04187341806559225,0.8879076086956521,0.04324370316732738,0.8879076086956521,0.04343313034334395
+flat_mae,patch,logistic,aabc_sex,58,0.046415888336127774,train,0.9243856332703214,0.011444890109943479,0.9220798350272499,0.011829487554457478,0.9200445499575016,0.0120498928722987
+flat_mae,patch,logistic,aabc_sex,58,0.046415888336127774,test,0.8,0.05269962911720449,0.7975911676145868,0.052978822356543255,0.8036684782608696,0.05225068674519407
+flat_mae,patch,logistic,aabc_sex,59,0.046415888336127774,train,0.9168241965973535,0.012166120743473485,0.9145119586296058,0.012555386906840141,0.9135086022450833,0.012798129046671157
+flat_mae,patch,logistic,aabc_sex,59,0.046415888336127774,test,0.8727272727272727,0.04536135902463663,0.8699763593380614,0.04632550080479619,0.8722826086956521,0.04603412930281962
+flat_mae,patch,logistic,aabc_sex,60,0.005994842503189409,train,0.8809073724007561,0.013806404837909878,0.8773574837805116,0.01422266443461675,0.875773029690202,0.014268415174205667
+flat_mae,patch,logistic,aabc_sex,60,0.005994842503189409,test,0.8363636363636363,0.04786946437798012,0.8250265111346766,0.05401258850812746,0.8165760869565217,0.05336562533883418
+flat_mae,patch,logistic,aabc_sex,61,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,61,2.782559402207126,test,0.7636363636363637,0.05629291751173917,0.7585275244849713,0.057295239610100114,0.7601902173913043,0.05716227324842578
+flat_mae,patch,logistic,aabc_sex,62,0.3593813663804626,train,0.9792060491493384,0.006119888839228942,0.9786941127795048,0.00627515442143354,0.978985316099534,0.0063844501604304665
+flat_mae,patch,logistic,aabc_sex,62,0.3593813663804626,test,0.8909090909090909,0.042532383384471135,0.8879076086956521,0.043869281572619365,0.8879076086956521,0.04397386528961242
+flat_mae,patch,logistic,aabc_sex,63,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,63,2.782559402207126,test,0.9454545454545454,0.030001123945887918,0.9435897435897436,0.031403955628115855,0.9408967391304348,0.03285763773574193
+flat_mae,patch,logistic,aabc_sex,64,0.3593813663804626,train,0.9810964083175804,0.0062271128177261775,0.9806193030276386,0.0063854341297253925,0.9806193030276386,0.006470150763331976
+flat_mae,patch,logistic,aabc_sex,64,0.3593813663804626,test,0.8,0.05472460277389255,0.7975911676145868,0.05498243239251332,0.8036684782608696,0.05441472233104929
+flat_mae,patch,logistic,aabc_sex,65,0.046415888336127774,train,0.9168241965973535,0.012194560330175588,0.9144012944983819,0.012605951161707313,0.9129004367068203,0.012861589890578437
+flat_mae,patch,logistic,aabc_sex,65,0.046415888336127774,test,0.8727272727272727,0.045901497121733706,0.8683760683760684,0.04795800107377598,0.8661684782608696,0.048408551303511474
+flat_mae,patch,logistic,aabc_sex,66,0.005994842503189409,train,0.8827977315689981,0.013499480879100606,0.8790598542729874,0.013964036552584425,0.8767988510800433,0.01408346683722175
+flat_mae,patch,logistic,aabc_sex,66,0.005994842503189409,test,0.7818181818181819,0.05590187388748458,0.7758152173913043,0.057523772721999294,0.7758152173913043,0.05755525874797466
+flat_mae,patch,logistic,aabc_sex,67,0.046415888336127774,train,0.9130434782608695,0.01185416425812814,0.9105104442483083,0.012262764179998872,0.9090242973123479,0.012507610801665132
+flat_mae,patch,logistic,aabc_sex,67,0.046415888336127774,test,0.9090909090909091,0.03910915430450041,0.9071259709557582,0.03983721099297673,0.9096467391304348,0.0393662235498639
+flat_mae,patch,logistic,aabc_sex,68,0.046415888336127774,train,0.9243856332703214,0.01154680897149735,0.9225702576112412,0.011834050305909541,0.9230853776488174,0.011986265103417331
+flat_mae,patch,logistic,aabc_sex,68,0.046415888336127774,test,0.9090909090909091,0.03885738280362057,0.9045470322804582,0.042070079932440566,0.8974184782608696,0.043712359140952936
+flat_mae,patch,logistic,aabc_sex,69,0.046415888336127774,train,0.9149338374291115,0.01199138487422192,0.9125128171203651,0.012370957852902594,0.9112664497787157,0.012610897536273077
+flat_mae,patch,logistic,aabc_sex,69,0.046415888336127774,test,0.9818181818181818,0.018124236091054922,0.9811965811965813,0.019033310260698953,0.9782608695652174,0.02167028228278305
+flat_mae,patch,logistic,aabc_sex,70,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,70,2.782559402207126,test,0.8545454545454545,0.048198967788405656,0.8533333333333333,0.04815418032403406,0.8627717391304348,0.04631582983263305
+flat_mae,patch,logistic,aabc_sex,71,0.046415888336127774,train,0.9130434782608695,0.011907668750406795,0.9105104442483083,0.01228180814891195,0.9090242973123479,0.012471793063889338
+flat_mae,patch,logistic,aabc_sex,71,0.046415888336127774,test,0.8545454545454545,0.042705387044781905,0.8428571428571429,0.04959198270756321,0.8322010869565217,0.048915602288142374
+flat_mae,patch,logistic,aabc_sex,72,0.3593813663804626,train,0.9848771266540642,0.0051731653410044855,0.9845140515222482,0.005293019915856736,0.9851036079603739,0.005184221081145655
+flat_mae,patch,logistic,aabc_sex,72,0.3593813663804626,test,0.8727272727272727,0.042466375614728866,0.8663658451926415,0.04573044341950363,0.8600543478260869,0.04643619318417449
+flat_mae,patch,logistic,aabc_sex,73,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,73,21.54434690031882,test,0.9090909090909091,0.038264412830412,0.9045470322804582,0.041573816592526514,0.8974184782608696,0.04329736525477135
+flat_mae,patch,logistic,aabc_sex,74,0.005994842503189409,train,0.8790170132325141,0.013284897127159228,0.8753277360435999,0.013770957615428923,0.8735308772238342,0.01405747786262297
+flat_mae,patch,logistic,aabc_sex,74,0.005994842503189409,test,0.8727272727272727,0.04654447442149876,0.8699763593380614,0.047472654440359224,0.8722826086956521,0.04711054185814096
+flat_mae,patch,logistic,aabc_sex,75,0.3593813663804626,train,0.9810964083175804,0.005925574980472664,0.9806425644028103,0.006061830611257943,0.9812274685659017,0.0059428712821264825
+flat_mae,patch,logistic,aabc_sex,75,0.3593813663804626,test,0.8545454545454545,0.04482577475788485,0.8505434782608696,0.04628517780841475,0.8505434782608696,0.04650196342541968
+flat_mae,patch,logistic,aabc_sex,76,0.3593813663804626,train,0.9867674858223062,0.004809066293344714,0.9864577733405013,0.004912546492483836,0.9873457604267417,0.004667884311138566
+flat_mae,patch,logistic,aabc_sex,76,0.3593813663804626,test,0.8181818181818182,0.05271973613849365,0.8106060606060606,0.05573070000133594,0.8070652173913043,0.05548169430489548
+flat_mae,patch,logistic,aabc_sex,77,0.046415888336127774,train,0.9243856332703214,0.011434804952975448,0.9220798350272499,0.011836343839198599,0.9200445499575016,0.012127629908291459
+flat_mae,patch,logistic,aabc_sex,77,0.046415888336127774,test,0.8363636363636363,0.04852091991524362,0.8250265111346766,0.05446445629425495,0.8165760869565217,0.053927853743368455
+flat_mae,patch,logistic,aabc_sex,78,0.046415888336127774,train,0.9224952741020794,0.011497051946718318,0.9203912716328067,0.011855368162087213,0.9196268941059234,0.01214576630283103
+flat_mae,patch,logistic,aabc_sex,78,0.046415888336127774,test,0.9272727272727272,0.03654997710596111,0.9252717391304348,0.0376408995727102,0.9252717391304348,0.037813069227190715
+flat_mae,patch,logistic,aabc_sex,79,0.046415888336127774,train,0.9187145557655955,0.011964943200071017,0.9165079190295289,0.01232157112279228,0.9157507547114512,0.012554500918901236
+flat_mae,patch,logistic,aabc_sex,79,0.046415888336127774,test,0.8545454545454545,0.04746194299707766,0.84593837535014,0.05219844710407987,0.8383152173913043,0.05233849217449428
+flat_mae,patch,logistic,aabc_sex,80,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,80,2.782559402207126,test,0.7090909090909091,0.062484113683478934,0.696969696969697,0.06558017612631742,0.6949728260869565,0.06471991278065005
+flat_mae,patch,logistic,aabc_sex,81,0.046415888336127774,train,0.9168241965973535,0.01124217051953436,0.9144012944983819,0.011606435760512894,0.9129004367068203,0.011827257885670173
+flat_mae,patch,logistic,aabc_sex,81,0.046415888336127774,test,0.8909090909090909,0.041228650874028876,0.8863636363636364,0.04358952704766233,0.8817934782608696,0.044539619255488855
+flat_mae,patch,logistic,aabc_sex,82,0.046415888336127774,train,0.9243856332703214,0.011549707776792909,0.922283598754187,0.011917302347633335,0.9212608810340279,0.012219877492282878
+flat_mae,patch,logistic,aabc_sex,82,0.046415888336127774,test,0.8545454545454545,0.04423042593636641,0.84593837535014,0.04841159693018112,0.8383152173913043,0.048776578278426956
+flat_mae,patch,logistic,aabc_sex,83,0.046415888336127774,train,0.9262759924385633,0.010915301819398107,0.924637543514869,0.011126032755325742,0.9259356956534482,0.011074466032620429
+flat_mae,patch,logistic,aabc_sex,83,0.046415888336127774,test,0.8909090909090909,0.03819523573769183,0.8821428571428571,0.04486985290714586,0.8695652173913043,0.045668216642892404
+flat_mae,patch,logistic,aabc_sex,84,0.046415888336127774,train,0.9300567107750473,0.011492761897415023,0.9280660940767447,0.011862949940555215,0.9267710073566048,0.012102672412429915
+flat_mae,patch,logistic,aabc_sex,84,0.046415888336127774,test,0.8,0.0560218493385017,0.7931623931623932,0.05827916762984857,0.7914402173913043,0.058255216809721115
+flat_mae,patch,logistic,aabc_sex,85,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,85,2.782559402207126,test,0.8909090909090909,0.04269276749865502,0.8879076086956521,0.044144738763871746,0.8879076086956521,0.044726083601717544
+flat_mae,patch,logistic,aabc_sex,86,0.046415888336127774,train,0.9130434782608695,0.012600970303903957,0.9106261385673151,0.012971798846859171,0.9096324628506112,0.01318939373536892
+flat_mae,patch,logistic,aabc_sex,86,0.046415888336127774,test,0.8727272727272727,0.04418742949776365,0.8699763593380614,0.04506553745968865,0.8722826086956521,0.044936406075238904
+flat_mae,patch,logistic,aabc_sex,87,0.005994842503189409,train,0.8733459357277883,0.014605526608566781,0.8690327944572749,0.015250837150799971,0.866196254286468,0.015541183347975251
+flat_mae,patch,logistic,aabc_sex,87,0.005994842503189409,test,0.8363636363636363,0.04991475708852653,0.8328267477203647,0.05084546982440388,0.8349184782608696,0.05046261887237688
+flat_mae,patch,logistic,aabc_sex,88,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,88,2.782559402207126,test,0.8181818181818182,0.05145016364252861,0.8106060606060606,0.05452648044088764,0.8070652173913043,0.05459618185922766
+flat_mae,patch,logistic,aabc_sex,89,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,89,2.782559402207126,test,0.8545454545454545,0.04382540868355924,0.84593837535014,0.04801965265036465,0.8383152173913043,0.04842113524321146
+flat_mae,patch,logistic,aabc_sex,90,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,90,166.81005372000556,test,0.8363636363636363,0.04808730352526097,0.8307692307692308,0.050320559157121116,0.8288043478260869,0.0504716924623748
+flat_mae,patch,logistic,aabc_sex,91,0.046415888336127774,train,0.9243856332703214,0.012274880994454413,0.922182994998529,0.012696916679473919,0.9206527154957649,0.013035118368354505
+flat_mae,patch,logistic,aabc_sex,91,0.046415888336127774,test,0.8363636363636363,0.04921847398912505,0.8328267477203647,0.05017049025434237,0.8349184782608696,0.05016582556230712
+flat_mae,patch,logistic,aabc_sex,92,0.046415888336127774,train,0.9149338374291115,0.012637726149938685,0.9123982027003654,0.01308055591951302,0.9106582842404525,0.01337527183095333
+flat_mae,patch,logistic,aabc_sex,92,0.046415888336127774,test,0.9090909090909091,0.037837477594889986,0.905982905982906,0.03944760922374126,0.9035326086956521,0.04011543204446846
+flat_mae,patch,logistic,aabc_sex,93,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,93,166.81005372000556,test,0.8181818181818182,0.04752462954324986,0.8035714285714286,0.054673392528910764,0.7948369565217391,0.05323249361949366
+flat_mae,patch,logistic,aabc_sex,94,0.046415888336127774,train,0.9111531190926276,0.012316593925511089,0.9087412138229735,0.012685549730581079,0.9079984759225066,0.01286583971074051
+flat_mae,patch,logistic,aabc_sex,94,0.046415888336127774,test,0.9090909090909091,0.038471197769906255,0.9071259709557582,0.039199159975394214,0.9096467391304348,0.03881062805028247
+flat_mae,patch,logistic,aabc_sex,95,0.046415888336127774,train,0.9206049149338374,0.011955376858895257,0.9183977786918963,0.0123151766397194,0.9173847416395557,0.012513533139835687
+flat_mae,patch,logistic,aabc_sex,95,0.046415888336127774,test,0.8909090909090909,0.04221785294680458,0.8879076086956521,0.04346054227662537,0.8879076086956521,0.043627257776545324
+flat_mae,patch,logistic,aabc_sex,96,0.046415888336127774,train,0.9168241965973535,0.0117937107280632,0.9147249333216096,0.012080897298253821,0.9147249333216096,0.012149202759156088
+flat_mae,patch,logistic,aabc_sex,96,0.046415888336127774,test,0.9272727272727272,0.034779000028396546,0.9242424242424243,0.03690396877201548,0.9191576086956521,0.0386411165712355
+flat_mae,patch,logistic,aabc_sex,97,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,aabc_sex,97,21.54434690031882,test,0.8363636363636363,0.0499274051511188,0.8307692307692308,0.05211739005349969,0.8288043478260869,0.05203445604333177
+flat_mae,patch,logistic,aabc_sex,98,0.3593813663804626,train,0.9810964083175804,0.005912660317600871,0.9806425644028103,0.006046251952659396,0.9812274685659017,0.005921402391254824
+flat_mae,patch,logistic,aabc_sex,98,0.3593813663804626,test,0.8,0.05175989065861454,0.7861435136090491,0.057448597315347114,0.7792119565217391,0.05626817148987758
+flat_mae,patch,logistic,aabc_sex,99,0.005994842503189409,train,0.8752362948960303,0.014886755482489229,0.8717679379444085,0.01530706156831197,0.8708710689058883,0.01542111997727275
+flat_mae,patch,logistic,aabc_sex,99,0.005994842503189409,test,0.8727272727272727,0.04610954975395439,0.8663658451926415,0.05005989706526048,0.8600543478260869,0.050864338815354994
+flat_mae,patch,logistic,aabc_sex,100,0.046415888336127774,train,0.9243856332703214,0.011957094212854265,0.922283598754187,0.012335065093687186,0.9212608810340279,0.0126029765418789
+flat_mae,patch,logistic,aabc_sex,100,0.046415888336127774,test,0.8727272727272727,0.042902988780281856,0.8699763593380614,0.04380719510719203,0.8722826086956521,0.04361875269678864

data_scaling/n400_1/eval_v2/aabc_sex__patch__logistic/log.txt

@@ -0,0 +1,245 @@
+fMRI foundation model logistic probe eval
+version: 0.1.dev66+g7ddd3aa04
+sha: 58906bf7243fb545e1349221e6921a1797e2e666, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-02-26 17:14:46
+config:
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (aabc_sex patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_1/eval_v2/aabc_sex__patch__logistic
+model: flat_mae
+representation: patch
+dataset: aabc_sex
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/aabc_sex__patch__logistic
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=True, reg_tokens=0, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: aabc_sex (flat)
+train (n=471):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'segment', 'bold', 'mean', 'std'],
+    num_rows: 471
+}),
+    labels=[0 1],
+    counts=[269 202]
+)
+
+validation (n=58):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'segment', 'bold', 'mean', 'std'],
+    num_rows: 58
+}),
+    labels=[0 1],
+    counts=[36 22]
+)
+
+test (n=55):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'segment', 'bold', 'mean', 'std'],
+    num_rows: 55
+}),
+    labels=[0 1],
+    counts=[33 22]
+)
+
+extracting features for all splits
+extract (train)  [  0/236]  eta: 0:24:33    time: 6.2439  data: 5.1290  max mem: 3205
+extract (train)  [ 20/236]  eta: 0:02:00    time: 0.2718  data: 0.0882  max mem: 3393
+extract (train)  [ 40/236]  eta: 0:01:14    time: 0.1938  data: 0.0549  max mem: 3393
+extract (train)  [ 60/236]  eta: 0:00:56    time: 0.2028  data: 0.0637  max mem: 3393
+extract (train)  [ 80/236]  eta: 0:00:47    time: 0.2412  data: 0.0853  max mem: 3393
+extract (train)  [100/236]  eta: 0:00:38    time: 0.2030  data: 0.0641  max mem: 3393
+extract (train)  [120/236]  eta: 0:00:31    time: 0.2039  data: 0.0645  max mem: 3393
+extract (train)  [140/236]  eta: 0:00:25    time: 0.2108  data: 0.0708  max mem: 3393
+extract (train)  [160/236]  eta: 0:00:19    time: 0.2120  data: 0.0720  max mem: 3393
+extract (train)  [180/236]  eta: 0:00:13    time: 0.2062  data: 0.0690  max mem: 3393
+extract (train)  [200/236]  eta: 0:00:08    time: 0.2141  data: 0.0733  max mem: 3393
+extract (train)  [220/236]  eta: 0:00:03    time: 0.1917  data: 0.0613  max mem: 3393
+extract (train)  [235/236]  eta: 0:00:00    time: 0.1737  data: 0.0523  max mem: 3393
+extract (train) Total time: 0:00:56 (0.2388 s / it)
+extract (validation)  [ 0/29]  eta: 0:02:06    time: 4.3553  data: 4.1916  max mem: 3393
+extract (validation)  [20/29]  eta: 0:00:03    time: 0.2010  data: 0.0659  max mem: 3393
+extract (validation)  [28/29]  eta: 0:00:00    time: 0.1704  data: 0.0500  max mem: 3393
+extract (validation) Total time: 0:00:10 (0.3482 s / it)
+extract (test)  [ 0/28]  eta: 0:02:04    time: 4.4628  data: 4.3233  max mem: 3393
+extract (test)  [20/28]  eta: 0:00:03    time: 0.1915  data: 0.0588  max mem: 3393
+extract (test)  [27/28]  eta: 0:00:00    time: 0.1625  data: 0.0467  max mem: 3393
+extract (test) Total time: 0:00:09 (0.3503 s / it)
+feature extraction time: 0:01:16
+train features: (471, 768)
+validation features: (58, 768)
+test features: (55, 768)
+evaluating fixed splits
+eval results (fixed splits):
+
+| model    | repr   | clf      | dataset   | trial   |      C | split   |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:----------|:--------|-------:|:--------|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | aabc_sex  |         | 2.7826 | train   | 1       |  0        | 1       | 0        | 1       |   0        |
+| flat_mae | patch  | logistic | aabc_sex  |         | 2.7826 | test    | 0.89091 |  0.043863 | 0.88911 | 0.043926 | 0.90152 |   0.040231 |
+
+
+evaluating random splits (n=100)
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 1, "C": 0.3593813663804626, "split": "test", "acc": 0.8181818181818182, "acc_std": 0.05265588898336156, "f1": 0.8166666666666667, "f1_std": 0.05265479976916016, "bacc": 0.8254076086956521, "bacc_std": 0.05117159420062498}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 2, "C": 0.046415888336127774, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.051022263816481866, "f1": 0.8328267477203647, "f1_std": 0.052100542340955486, "bacc": 0.8349184782608696, "bacc_std": 0.052040162627840746}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 3, "C": 0.046415888336127774, "split": "test", "acc": 0.7818181818181819, "acc_std": 0.0554105458829032, "f1": 0.7727272727272727, "f1_std": 0.05863437044538351, "bacc": 0.7697010869565217, "bacc_std": 0.05848903869876677}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 4, "C": 166.81005372000556, "split": "test", "acc": 0.8181818181818182, "acc_std": 0.050275682964238864, "f1": 0.8166666666666667, "f1_std": 0.05041723963269043, "bacc": 0.8254076086956521, "bacc_std": 0.049211660530014}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 5, "C": 0.3593813663804626, "split": "test", "acc": 0.8, "acc_std": 0.05273855678629085, "f1": 0.790003471017008, "f1_std": 0.056230055783621925, "bacc": 0.7853260869565217, "bacc_std": 0.05558520219851598}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 6, "C": 0.046415888336127774, "split": "test", "acc": 0.9090909090909091, "acc_std": 0.03693972754466904, "f1": 0.9045470322804582, "f1_std": 0.03992807000558217, "bacc": 0.8974184782608696, "bacc_std": 0.04171778202044614}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 7, "C": 2.782559402207126, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.04707285017014047, "f1": 0.8521505376344086, "f1_std": 0.04774533501656769, "bacc": 0.8566576086956521, "bacc_std": 0.047426358897281806}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 8, "C": 0.046415888336127774, "split": "test", "acc": 0.8, "acc_std": 0.05253974174669985, "f1": 0.795677136102668, "f1_std": 0.053761764767330654, "bacc": 0.7975543478260869, "bacc_std": 0.05411603893049472}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 9, "C": 0.046415888336127774, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.046299049487771055, "f1": 0.8683760683760684, "f1_std": 0.048387140550290475, "bacc": 0.8661684782608696, "bacc_std": 0.04900389469866049}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 10, "C": 0.046415888336127774, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.048698770497147895, "f1": 0.8505434782608696, "f1_std": 0.050083757649930345, "bacc": 0.8505434782608696, "bacc_std": 0.050190351568757355}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 11, "C": 0.005994842503189409, "split": "test", "acc": 0.9090909090909091, "acc_std": 0.03667531878885939, "f1": 0.9027925061859314, "f1_std": 0.04190960075868798, "bacc": 0.8913043478260869, "bacc_std": 0.04385092463885362}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 12, "C": 0.3593813663804626, "split": "test", "acc": 0.8, "acc_std": 0.05178686134787916, "f1": 0.7861435136090491, "f1_std": 0.057965069555226255, "bacc": 0.7792119565217391, "bacc_std": 0.05650578427288931}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 13, "C": 166.81005372000556, "split": "test", "acc": 0.8909090909090909, "acc_std": 0.041913551330374846, "f1": 0.8879076086956521, "f1_std": 0.04308790390649044, "bacc": 0.8879076086956521, "bacc_std": 0.04317491706667154}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 14, "C": 0.3593813663804626, "split": "test", "acc": 0.8909090909090909, "acc_std": 0.03883382695154418, "f1": 0.8879076086956521, "f1_std": 0.03999668307543158, "bacc": 0.8879076086956521, "bacc_std": 0.04032041382000847}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 15, "C": 0.046415888336127774, "split": "test", "acc": 0.7818181818181819, "acc_std": 0.05706815991455355, "f1": 0.7758152173913043, "f1_std": 0.058976953550242794, "bacc": 0.7758152173913043, "bacc_std": 0.059001860339285686}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 16, "C": 0.3593813663804626, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.051427027656220974, "f1": 0.8307692307692308, "f1_std": 0.05383410687411033, "bacc": 0.8288043478260869, "bacc_std": 0.054278706998561854}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 17, "C": 2.782559402207126, "split": "test", "acc": 0.8, "acc_std": 0.051782839622203425, "f1": 0.790003471017008, "f1_std": 0.0555870379546963, "bacc": 0.7853260869565217, "bacc_std": 0.05496704207452725}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 18, "C": 0.046415888336127774, "split": "test", "acc": 0.8181818181818182, "acc_std": 0.05072647767379556, "f1": 0.8176392572944298, "f1_std": 0.05062105628133805, "bacc": 0.8315217391304348, "bacc_std": 0.047879830440684626}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 19, "C": 0.3593813663804626, "split": "test", "acc": 0.7818181818181819, "acc_std": 0.054926378976207925, "f1": 0.7727272727272727, "f1_std": 0.05772363026484643, "bacc": 0.7697010869565217, "bacc_std": 0.0571900716828565}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 20, "C": 0.046415888336127774, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.04610408633580652, "f1": 0.8250265111346766, "f1_std": 0.05230800624333397, "bacc": 0.8165760869565217, "bacc_std": 0.051851235104635476}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 21, "C": 0.046415888336127774, "split": "test", "acc": 0.7818181818181819, "acc_std": 0.052523229020851894, "f1": 0.7727272727272727, "f1_std": 0.05581268638224436, "bacc": 0.7697010869565217, "bacc_std": 0.055373226476392326}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 22, "C": 21.54434690031882, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.0519799579939985, "f1": 0.8328267477203647, "f1_std": 0.053059633258707335, "bacc": 0.8349184782608696, "bacc_std": 0.05295779550098547}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 23, "C": 0.046415888336127774, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.046282938683959336, "f1": 0.8699763593380614, "f1_std": 0.04729635012452601, "bacc": 0.8722826086956521, "bacc_std": 0.04700199204630469}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 24, "C": 0.3593813663804626, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.04799629462282125, "f1": 0.8699763593380614, "f1_std": 0.04910219932790519, "bacc": 0.8722826086956521, "bacc_std": 0.04885495493974446}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 25, "C": 0.3593813663804626, "split": "test", "acc": 0.8181818181818182, "acc_std": 0.05094568814429603, "f1": 0.8151881720430108, "f1_std": 0.05180721624321444, "bacc": 0.8192934782608696, "bacc_std": 0.05138047076561111}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 26, "C": 0.046415888336127774, "split": "test", "acc": 0.7818181818181819, "acc_std": 0.05115633468606205, "f1": 0.7727272727272727, "f1_std": 0.05421765843013767, "bacc": 0.7697010869565217, "bacc_std": 0.053793701941407264}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 27, "C": 2.782559402207126, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.04253984422993856, "f1": 0.8699763593380614, "f1_std": 0.04326416544506314, "bacc": 0.8722826086956521, "bacc_std": 0.042897966358643964}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 28, "C": 0.005994842503189409, "split": "test", "acc": 0.8, "acc_std": 0.05296143614941217, "f1": 0.790003471017008, "f1_std": 0.057136675986417966, "bacc": 0.7853260869565217, "bacc_std": 0.05670747727149538}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 29, "C": 0.3593813663804626, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.04905835957946512, "f1": 0.8307692307692308, "f1_std": 0.051178039635536604, "bacc": 0.8288043478260869, "bacc_std": 0.05126013007047176}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 30, "C": 0.046415888336127774, "split": "test", "acc": 0.7636363636363637, "acc_std": 0.059432144793123046, "f1": 0.7555555555555555, "f1_std": 0.061952149867716945, "bacc": 0.7540760869565217, "bacc_std": 0.061502168699851326}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 31, "C": 0.046415888336127774, "split": "test", "acc": 0.7818181818181819, "acc_std": 0.05530347231039635, "f1": 0.7782258064516129, "f1_std": 0.055718803885056456, "bacc": 0.7819293478260869, "bacc_std": 0.055179640142934576}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 32, "C": 0.046415888336127774, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.04434598754758142, "f1": 0.8683760683760684, "f1_std": 0.046439782848170005, "bacc": 0.8661684782608696, "bacc_std": 0.04699345640291025}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 33, "C": 0.046415888336127774, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.045888820007196636, "f1": 0.8505434782608696, "f1_std": 0.047194520021074556, "bacc": 0.8505434782608696, "bacc_std": 0.04709612353182535}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 34, "C": 0.3593813663804626, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.049449344646318734, "f1": 0.8307692307692308, "f1_std": 0.05149415963006614, "bacc": 0.8288043478260869, "bacc_std": 0.0517093403774025}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 35, "C": 2.782559402207126, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.043602577829624116, "f1": 0.8699763593380614, "f1_std": 0.044378928880580606, "bacc": 0.8722826086956521, "bacc_std": 0.04402821761910765}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 36, "C": 0.3593813663804626, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.04887976139856006, "f1": 0.8328267477203647, "f1_std": 0.04978049940641453, "bacc": 0.8349184782608696, "bacc_std": 0.04963764381180495}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 37, "C": 0.3593813663804626, "split": "test", "acc": 0.8181818181818182, "acc_std": 0.052573267253919684, "f1": 0.8151881720430108, "f1_std": 0.053486615006010346, "bacc": 0.8192934782608696, "bacc_std": 0.053490187548087584}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 38, "C": 0.046415888336127774, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.04203854980159861, "f1": 0.8639095086603039, "f1_std": 0.04749952680671482, "bacc": 0.8539402173913043, "bacc_std": 0.04816852800661158}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 39, "C": 0.046415888336127774, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.04533743459004501, "f1": 0.8505434782608696, "f1_std": 0.046700870489602, "bacc": 0.8505434782608696, "bacc_std": 0.046837935783746884}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 40, "C": 2.782559402207126, "split": "test", "acc": 0.8181818181818182, "acc_std": 0.049738489677229825, "f1": 0.8131793478260869, "f1_std": 0.05141701602857101, "bacc": 0.8131793478260869, "bacc_std": 0.05157697266027768}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 41, "C": 0.005994842503189409, "split": "test", "acc": 0.9090909090909091, "acc_std": 0.038480287966483194, "f1": 0.905982905982906, "f1_std": 0.040362495966507414, "bacc": 0.9035326086956521, "bacc_std": 0.04107032746648314}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 42, "C": 166.81005372000556, "split": "test", "acc": 0.7818181818181819, "acc_std": 0.055751941744341615, "f1": 0.7782258064516129, "f1_std": 0.05660795085808481, "bacc": 0.7819293478260869, "bacc_std": 0.05630541567402165}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 43, "C": 0.046415888336127774, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.0439057743900889, "f1": 0.8521505376344086, "f1_std": 0.04445632866046411, "bacc": 0.8566576086956521, "bacc_std": 0.04384902873841091}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 44, "C": 0.046415888336127774, "split": "test", "acc": 0.8, "acc_std": 0.05146166349737786, "f1": 0.790003471017008, "f1_std": 0.0554262734827424, "bacc": 0.7853260869565217, "bacc_std": 0.05497276355871275}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 45, "C": 0.046415888336127774, "split": "test", "acc": 0.9272727272727272, "acc_std": 0.034506216950927315, "f1": 0.9266666666666667, "f1_std": 0.03439735271827876, "bacc": 0.9375, "bacc_std": 0.029653780192203154}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 46, "C": 0.046415888336127774, "split": "test", "acc": 0.9090909090909091, "acc_std": 0.0393400097055957, "f1": 0.9071259709557582, "f1_std": 0.0400938976349243, "bacc": 0.9096467391304348, "bacc_std": 0.03956758514624431}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 47, "C": 0.3593813663804626, "split": "test", "acc": 0.8909090909090909, "acc_std": 0.04032423957771133, "f1": 0.8863636363636364, "f1_std": 0.042786489845962436, "bacc": 0.8817934782608696, "bacc_std": 0.04366744834135569}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 48, "C": 0.005994842503189409, "split": "test", "acc": 0.8909090909090909, "acc_std": 0.04058312151703114, "f1": 0.8879076086956521, "f1_std": 0.041885923840686734, "bacc": 0.8879076086956521, "bacc_std": 0.04216694092273785}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 49, "C": 166.81005372000556, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.047862060752793846, "f1": 0.8343927735028438, "f1_std": 0.04817250193499985, "bacc": 0.8410326086956521, "bacc_std": 0.04748820110383062}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 50, "C": 0.046415888336127774, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.04820247927507941, "f1": 0.8484848484848485, "f1_std": 0.0507837081955852, "bacc": 0.8444293478260869, "bacc_std": 0.05090977446513677}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 51, "C": 0.046415888336127774, "split": "test", "acc": 0.8181818181818182, "acc_std": 0.04529986773191422, "f1": 0.8074229691876751, "f1_std": 0.05073726921765426, "bacc": 0.8009510869565217, "bacc_std": 0.050273032794018735}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 52, "C": 21.54434690031882, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.04892133675690641, "f1": 0.8307692307692308, "f1_std": 0.051052479399323294, "bacc": 0.8288043478260869, "bacc_std": 0.05116841613459943}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 53, "C": 0.3593813663804626, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.049437483687870154, "f1": 0.8521505376344086, "f1_std": 0.05005743999904591, "bacc": 0.8566576086956521, "bacc_std": 0.049228255619581746}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 54, "C": 0.005994842503189409, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.047307106380010025, "f1": 0.8521505376344086, "f1_std": 0.04786339862693188, "bacc": 0.8566576086956521, "bacc_std": 0.04702470759122328}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 55, "C": 21.54434690031882, "split": "test", "acc": 0.8, "acc_std": 0.053532546676942294, "f1": 0.7931623931623932, "f1_std": 0.05615577862990481, "bacc": 0.7914402173913043, "bacc_std": 0.05628428399847978}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 56, "C": 0.046415888336127774, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.048303106345706434, "f1": 0.8307692307692308, "f1_std": 0.050494865110545385, "bacc": 0.8288043478260869, "bacc_std": 0.05078730095183825}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 57, "C": 0.046415888336127774, "split": "test", "acc": 0.8909090909090909, "acc_std": 0.04187341806559225, "f1": 0.8879076086956521, "f1_std": 0.04324370316732738, "bacc": 0.8879076086956521, "bacc_std": 0.04343313034334395}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 58, "C": 0.046415888336127774, "split": "test", "acc": 0.8, "acc_std": 0.05269962911720449, "f1": 0.7975911676145868, "f1_std": 0.052978822356543255, "bacc": 0.8036684782608696, "bacc_std": 0.05225068674519407}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 59, "C": 0.046415888336127774, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.04536135902463663, "f1": 0.8699763593380614, "f1_std": 0.04632550080479619, "bacc": 0.8722826086956521, "bacc_std": 0.04603412930281962}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 60, "C": 0.005994842503189409, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.04786946437798012, "f1": 0.8250265111346766, "f1_std": 0.05401258850812746, "bacc": 0.8165760869565217, "bacc_std": 0.05336562533883418}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 61, "C": 2.782559402207126, "split": "test", "acc": 0.7636363636363637, "acc_std": 0.05629291751173917, "f1": 0.7585275244849713, "f1_std": 0.057295239610100114, "bacc": 0.7601902173913043, "bacc_std": 0.05716227324842578}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 62, "C": 0.3593813663804626, "split": "test", "acc": 0.8909090909090909, "acc_std": 0.042532383384471135, "f1": 0.8879076086956521, "f1_std": 0.043869281572619365, "bacc": 0.8879076086956521, "bacc_std": 0.04397386528961242}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 63, "C": 2.782559402207126, "split": "test", "acc": 0.9454545454545454, "acc_std": 0.030001123945887918, "f1": 0.9435897435897436, "f1_std": 0.031403955628115855, "bacc": 0.9408967391304348, "bacc_std": 0.03285763773574193}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 64, "C": 0.3593813663804626, "split": "test", "acc": 0.8, "acc_std": 0.05472460277389255, "f1": 0.7975911676145868, "f1_std": 0.05498243239251332, "bacc": 0.8036684782608696, "bacc_std": 0.05441472233104929}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 65, "C": 0.046415888336127774, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.045901497121733706, "f1": 0.8683760683760684, "f1_std": 0.04795800107377598, "bacc": 0.8661684782608696, "bacc_std": 0.048408551303511474}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 66, "C": 0.005994842503189409, "split": "test", "acc": 0.7818181818181819, "acc_std": 0.05590187388748458, "f1": 0.7758152173913043, "f1_std": 0.057523772721999294, "bacc": 0.7758152173913043, "bacc_std": 0.05755525874797466}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 67, "C": 0.046415888336127774, "split": "test", "acc": 0.9090909090909091, "acc_std": 0.03910915430450041, "f1": 0.9071259709557582, "f1_std": 0.03983721099297673, "bacc": 0.9096467391304348, "bacc_std": 0.0393662235498639}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 68, "C": 0.046415888336127774, "split": "test", "acc": 0.9090909090909091, "acc_std": 0.03885738280362057, "f1": 0.9045470322804582, "f1_std": 0.042070079932440566, "bacc": 0.8974184782608696, "bacc_std": 0.043712359140952936}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 69, "C": 0.046415888336127774, "split": "test", "acc": 0.9818181818181818, "acc_std": 0.018124236091054922, "f1": 0.9811965811965813, "f1_std": 0.019033310260698953, "bacc": 0.9782608695652174, "bacc_std": 0.02167028228278305}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 70, "C": 2.782559402207126, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.048198967788405656, "f1": 0.8533333333333333, "f1_std": 0.04815418032403406, "bacc": 0.8627717391304348, "bacc_std": 0.04631582983263305}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 71, "C": 0.046415888336127774, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.042705387044781905, "f1": 0.8428571428571429, "f1_std": 0.04959198270756321, "bacc": 0.8322010869565217, "bacc_std": 0.048915602288142374}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 72, "C": 0.3593813663804626, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.042466375614728866, "f1": 0.8663658451926415, "f1_std": 0.04573044341950363, "bacc": 0.8600543478260869, "bacc_std": 0.04643619318417449}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 73, "C": 21.54434690031882, "split": "test", "acc": 0.9090909090909091, "acc_std": 0.038264412830412, "f1": 0.9045470322804582, "f1_std": 0.041573816592526514, "bacc": 0.8974184782608696, "bacc_std": 0.04329736525477135}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 74, "C": 0.005994842503189409, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.04654447442149876, "f1": 0.8699763593380614, "f1_std": 0.047472654440359224, "bacc": 0.8722826086956521, "bacc_std": 0.04711054185814096}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 75, "C": 0.3593813663804626, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.04482577475788485, "f1": 0.8505434782608696, "f1_std": 0.04628517780841475, "bacc": 0.8505434782608696, "bacc_std": 0.04650196342541968}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 76, "C": 0.3593813663804626, "split": "test", "acc": 0.8181818181818182, "acc_std": 0.05271973613849365, "f1": 0.8106060606060606, "f1_std": 0.05573070000133594, "bacc": 0.8070652173913043, "bacc_std": 0.05548169430489548}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 77, "C": 0.046415888336127774, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.04852091991524362, "f1": 0.8250265111346766, "f1_std": 0.05446445629425495, "bacc": 0.8165760869565217, "bacc_std": 0.053927853743368455}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 78, "C": 0.046415888336127774, "split": "test", "acc": 0.9272727272727272, "acc_std": 0.03654997710596111, "f1": 0.9252717391304348, "f1_std": 0.0376408995727102, "bacc": 0.9252717391304348, "bacc_std": 0.037813069227190715}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 79, "C": 0.046415888336127774, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.04746194299707766, "f1": 0.84593837535014, "f1_std": 0.05219844710407987, "bacc": 0.8383152173913043, "bacc_std": 0.05233849217449428}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 80, "C": 2.782559402207126, "split": "test", "acc": 0.7090909090909091, "acc_std": 0.062484113683478934, "f1": 0.696969696969697, "f1_std": 0.06558017612631742, "bacc": 0.6949728260869565, "bacc_std": 0.06471991278065005}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 81, "C": 0.046415888336127774, "split": "test", "acc": 0.8909090909090909, "acc_std": 0.041228650874028876, "f1": 0.8863636363636364, "f1_std": 0.04358952704766233, "bacc": 0.8817934782608696, "bacc_std": 0.044539619255488855}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 82, "C": 0.046415888336127774, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.04423042593636641, "f1": 0.84593837535014, "f1_std": 0.04841159693018112, "bacc": 0.8383152173913043, "bacc_std": 0.048776578278426956}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 83, "C": 0.046415888336127774, "split": "test", "acc": 0.8909090909090909, "acc_std": 0.03819523573769183, "f1": 0.8821428571428571, "f1_std": 0.04486985290714586, "bacc": 0.8695652173913043, "bacc_std": 0.045668216642892404}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 84, "C": 0.046415888336127774, "split": "test", "acc": 0.8, "acc_std": 0.0560218493385017, "f1": 0.7931623931623932, "f1_std": 0.05827916762984857, "bacc": 0.7914402173913043, "bacc_std": 0.058255216809721115}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 85, "C": 2.782559402207126, "split": "test", "acc": 0.8909090909090909, "acc_std": 0.04269276749865502, "f1": 0.8879076086956521, "f1_std": 0.044144738763871746, "bacc": 0.8879076086956521, "bacc_std": 0.044726083601717544}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 86, "C": 0.046415888336127774, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.04418742949776365, "f1": 0.8699763593380614, "f1_std": 0.04506553745968865, "bacc": 0.8722826086956521, "bacc_std": 0.044936406075238904}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 87, "C": 0.005994842503189409, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.04991475708852653, "f1": 0.8328267477203647, "f1_std": 0.05084546982440388, "bacc": 0.8349184782608696, "bacc_std": 0.05046261887237688}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 88, "C": 2.782559402207126, "split": "test", "acc": 0.8181818181818182, "acc_std": 0.05145016364252861, "f1": 0.8106060606060606, "f1_std": 0.05452648044088764, "bacc": 0.8070652173913043, "bacc_std": 0.05459618185922766}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 89, "C": 2.782559402207126, "split": "test", "acc": 0.8545454545454545, "acc_std": 0.04382540868355924, "f1": 0.84593837535014, "f1_std": 0.04801965265036465, "bacc": 0.8383152173913043, "bacc_std": 0.04842113524321146}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 90, "C": 166.81005372000556, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.04808730352526097, "f1": 0.8307692307692308, "f1_std": 0.050320559157121116, "bacc": 0.8288043478260869, "bacc_std": 0.0504716924623748}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 91, "C": 0.046415888336127774, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.04921847398912505, "f1": 0.8328267477203647, "f1_std": 0.05017049025434237, "bacc": 0.8349184782608696, "bacc_std": 0.05016582556230712}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 92, "C": 0.046415888336127774, "split": "test", "acc": 0.9090909090909091, "acc_std": 0.037837477594889986, "f1": 0.905982905982906, "f1_std": 0.03944760922374126, "bacc": 0.9035326086956521, "bacc_std": 0.04011543204446846}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 93, "C": 166.81005372000556, "split": "test", "acc": 0.8181818181818182, "acc_std": 0.04752462954324986, "f1": 0.8035714285714286, "f1_std": 0.054673392528910764, "bacc": 0.7948369565217391, "bacc_std": 0.05323249361949366}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 94, "C": 0.046415888336127774, "split": "test", "acc": 0.9090909090909091, "acc_std": 0.038471197769906255, "f1": 0.9071259709557582, "f1_std": 0.039199159975394214, "bacc": 0.9096467391304348, "bacc_std": 0.03881062805028247}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 95, "C": 0.046415888336127774, "split": "test", "acc": 0.8909090909090909, "acc_std": 0.04221785294680458, "f1": 0.8879076086956521, "f1_std": 0.04346054227662537, "bacc": 0.8879076086956521, "bacc_std": 0.043627257776545324}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 96, "C": 0.046415888336127774, "split": "test", "acc": 0.9272727272727272, "acc_std": 0.034779000028396546, "f1": 0.9242424242424243, "f1_std": 0.03690396877201548, "bacc": 0.9191576086956521, "bacc_std": 0.0386411165712355}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 97, "C": 21.54434690031882, "split": "test", "acc": 0.8363636363636363, "acc_std": 0.0499274051511188, "f1": 0.8307692307692308, "f1_std": 0.05211739005349969, "bacc": 0.8288043478260869, "bacc_std": 0.05203445604333177}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 98, "C": 0.3593813663804626, "split": "test", "acc": 0.8, "acc_std": 0.05175989065861454, "f1": 0.7861435136090491, "f1_std": 0.057448597315347114, "bacc": 0.7792119565217391, "bacc_std": 0.05626817148987758}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 99, "C": 0.005994842503189409, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.04610954975395439, "f1": 0.8663658451926415, "f1_std": 0.05005989706526048, "bacc": 0.8600543478260869, "bacc_std": 0.050864338815354994}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "aabc_sex", "trial": 100, "C": 0.046415888336127774, "split": "test", "acc": 0.8727272727272727, "acc_std": 0.042902988780281856, "f1": 0.8699763593380614, "f1_std": 0.04380719510719203, "bacc": 0.8722826086956521, "bacc_std": 0.04361875269678864}
+eval results (random splits):
+
+| model    | repr   | clf      | dataset   | split   |   n_trials |      C |   C_std |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:----------|:--------|-----------:|-------:|--------:|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | aabc_sex  | train   |        100 | 11.514 |   39.71 | 0.94839 |  0.040725 | 0.94692 | 0.04194  | 0.94631 |   0.042678 |
+| flat_mae | patch  | logistic | aabc_sex  | test    |        100 | 11.514 |   39.71 | 0.84836 |  0.046317 | 0.84311 | 0.048044 | 0.84205 |   0.048153 |
+
+
+done! total time: 0:05:14

data_scaling/n400_1/eval_v2/abide_dx__patch__logistic/config.yaml

@@ -0,0 +1,30 @@
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (abide_dx patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_1/eval_v2/abide_dx__patch__logistic
+model: flat_mae
+representation: patch
+dataset: abide_dx
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/abide_dx__patch__logistic
+remote_dir: null

data_scaling/n400_1/eval_v2/abide_dx__patch__logistic/eval_table.csv

@@ -0,0 +1,203 @@
+model,repr,clf,dataset,trial,C,split,acc,acc_std,f1,f1_std,bacc,bacc_std
+flat_mae,patch,logistic,abide_dx,,0.3593813663804626,train,0.9316239316239316,0.009578303089056401,0.9307242539123857,0.009733780399628968,0.9299445137566484,0.009901145254784898
+flat_mae,patch,logistic,abide_dx,,0.3593813663804626,test,0.5241935483870968,0.041284081559585684,0.5127539127539127,0.04184667249699166,0.5151872217858078,0.04119939857251283
+flat_mae,patch,logistic,abide_dx,1,0.005994842503189409,train,0.7165242165242165,0.017482637341670913,0.7095210667820004,0.018113957275255634,0.7080472499077151,0.01788823652685562
+flat_mae,patch,logistic,abide_dx,1,0.005994842503189409,test,0.6774193548387096,0.041113137346047476,0.6688034188034189,0.04295419279844657,0.6680672268907563,0.04203937864030787
+flat_mae,patch,logistic,abide_dx,2,0.3593813663804626,train,0.9216524216524217,0.010031799424812969,0.920486676730254,0.010219301282402097,0.9188999630860096,0.010373358578822559
+flat_mae,patch,logistic,abide_dx,2,0.3593813663804626,test,0.6209677419354839,0.04211467917333632,0.6179613241560145,0.042344171626269744,0.618172268907563,0.04239843385697718
+flat_mae,patch,logistic,abide_dx,3,0.046415888336127774,train,0.8105413105413105,0.01409940941434186,0.8074463324142113,0.01432886945855722,0.806016980435585,0.014272212961785177
+flat_mae,patch,logistic,abide_dx,3,0.046415888336127774,test,0.6048387096774194,0.04441990535419158,0.5989703649924097,0.04504116366441694,0.5987394957983193,0.04472068721015768
+flat_mae,patch,logistic,abide_dx,4,0.046415888336127774,train,0.8176638176638177,0.014569762408488346,0.8144774332080769,0.014924683905774245,0.8127722406792174,0.014932497700045554
+flat_mae,patch,logistic,abide_dx,4,0.046415888336127774,test,0.7016129032258065,0.03831011527471725,0.6944388944388944,0.039845593860878076,0.6932773109243697,0.03926182668976251
+flat_mae,patch,logistic,abide_dx,5,0.046415888336127774,train,0.8133903133903134,0.014677475114763954,0.8100566967190259,0.01498334369122216,0.8083056478405315,0.014953154093032708
+flat_mae,patch,logistic,abide_dx,5,0.046415888336127774,test,0.5967741935483871,0.04129690912688167,0.5915678524374176,0.04169908351233194,0.5913865546218487,0.04146902108947524
+flat_mae,patch,logistic,abide_dx,6,0.046415888336127774,train,0.8205128205128205,0.014663587948363357,0.8170922355393434,0.015071834588745036,0.8150609080841639,0.015078872003588357
+flat_mae,patch,logistic,abide_dx,6,0.046415888336127774,test,0.6370967741935484,0.04635304281882685,0.6342182890855457,0.046843179294802285,0.634453781512605,0.04677036162353719
+flat_mae,patch,logistic,abide_dx,7,0.005994842503189409,train,0.7165242165242165,0.01711661469307375,0.7105322594347735,0.017686557922821953,0.7092284976005906,0.017546572760843048
+flat_mae,patch,logistic,abide_dx,7,0.005994842503189409,test,0.5887096774193549,0.04462083000509283,0.5765651155005022,0.046826300397298085,0.5777310924369747,0.04541367312022607
+flat_mae,patch,logistic,abide_dx,8,0.3593813663804626,train,0.9373219373219374,0.009215859718798448,0.9365785950705092,0.009340085595253544,0.9360649686231082,0.009454571985151825
+flat_mae,patch,logistic,abide_dx,8,0.3593813663804626,test,0.5645161290322581,0.04229725774192688,0.5588932806324111,0.04293049709371854,0.5588235294117647,0.04267437159659311
+flat_mae,patch,logistic,abide_dx,9,0.3593813663804626,train,0.9216524216524217,0.00987335121733457,0.920486676730254,0.01006350335407584,0.9188999630860096,0.01024171581578018
+flat_mae,patch,logistic,abide_dx,9,0.3593813663804626,test,0.6048387096774194,0.04283920763410549,0.5972691721349506,0.04353892823111837,0.5971638655462186,0.04304172193901062
+flat_mae,patch,logistic,abide_dx,10,0.046415888336127774,train,0.8304843304843305,0.01399679348998892,0.8278401187053696,0.014270933065766584,0.8264673311184939,0.014318802047782402
+flat_mae,patch,logistic,abide_dx,10,0.046415888336127774,test,0.6612903225806451,0.040930955671154394,0.6502820306204673,0.043394048096207456,0.6502100840336134,0.042008753482431
+flat_mae,patch,logistic,abide_dx,11,0.046415888336127774,train,0.8076923076923077,0.014650219783339935,0.8039491513151771,0.015071280719662841,0.8019564414913252,0.015072563097944986
+flat_mae,patch,logistic,abide_dx,11,0.046415888336127774,test,0.6209677419354839,0.042008679916720464,0.6167554415729598,0.04245925815837763,0.6165966386554622,0.0423885638229868
+flat_mae,patch,logistic,abide_dx,12,0.005994842503189409,train,0.7193732193732194,0.01634174002689427,0.7129513739702229,0.016875859362168937,0.7115171650055371,0.016717499953353667
+flat_mae,patch,logistic,abide_dx,12,0.005994842503189409,test,0.5887096774193549,0.044167039376957366,0.5841388834089565,0.04458292141122282,0.5840336134453781,0.04437302500699791
+flat_mae,patch,logistic,abide_dx,13,0.3593813663804626,train,0.9230769230769231,0.010035358872074964,0.9220128373930218,0.010211408943347216,0.9207825765965301,0.010391390340097752
+flat_mae,patch,logistic,abide_dx,13,0.3593813663804626,test,0.5725806451612904,0.04249018678406638,0.5703170970905524,0.042903984753441864,0.5709033613445378,0.04313978178191481
+flat_mae,patch,logistic,abide_dx,14,0.3593813663804626,train,0.9273504273504274,0.009652200911040482,0.9263204469917579,0.009831232730783478,0.9249538575119971,0.010009716910258716
+flat_mae,patch,logistic,abide_dx,14,0.3593813663804626,test,0.6048387096774194,0.04310701056721079,0.5907590759075907,0.0453912057515939,0.592436974789916,0.04383432933102287
+flat_mae,patch,logistic,abide_dx,15,0.046415888336127774,train,0.8219373219373219,0.014186975497978607,0.8191597885560604,0.014484624294715072,0.8178294573643411,0.01455401087225457
+flat_mae,patch,logistic,abide_dx,15,0.046415888336127774,test,0.6532258064516129,0.039270305166193877,0.6480760345851759,0.039930931836775055,0.6475840336134454,0.03964505782892502
+flat_mae,patch,logistic,abide_dx,16,0.046415888336127774,train,0.8105413105413105,0.014546443589820905,0.8077223273658869,0.014795612022782697,0.8066076042820229,0.014817358049660674
+flat_mae,patch,logistic,abide_dx,16,0.046415888336127774,test,0.6129032258064516,0.04254916663914516,0.6112852664576802,0.042812067293305056,0.6123949579831933,0.04292760821989454
+flat_mae,patch,logistic,abide_dx,17,0.046415888336127774,train,0.8176638176638177,0.014606378047169018,0.8153846153846154,0.014829584234071961,0.8148394241417497,0.014887649285549748
+flat_mae,patch,logistic,abide_dx,17,0.046415888336127774,test,0.6129032258064516,0.043490570939756265,0.6045708211533352,0.04493554455811725,0.6045168067226891,0.04407857663726617
+flat_mae,patch,logistic,abide_dx,18,0.3593813663804626,train,0.9330484330484331,0.009340343630694485,0.9320992354629924,0.00951938682820236,0.9307124400147656,0.00975718669328957
+flat_mae,patch,logistic,abide_dx,18,0.3593813663804626,test,0.6290322580645161,0.043240642013598486,0.6242424242424243,0.04386617606615634,0.6239495798319328,0.043601468943845846
+flat_mae,patch,logistic,abide_dx,19,0.005994842503189409,train,0.7421652421652422,0.01589959930750264,0.7362649679624891,0.016402493074144674,0.7345514950166112,0.016262966153999667
+flat_mae,patch,logistic,abide_dx,19,0.005994842503189409,test,0.5725806451612904,0.04351731268865347,0.5478500171998624,0.04790838098260214,0.5551470588235294,0.04451250840553695
+flat_mae,patch,logistic,abide_dx,20,0.046415888336127774,train,0.8176638176638177,0.014051374476050232,0.8151415404871627,0.014344345271512923,0.8142488002953119,0.01447380682295407
+flat_mae,patch,logistic,abide_dx,20,0.046415888336127774,test,0.6532258064516129,0.03989847613008047,0.6408702094699266,0.04215016598399647,0.641281512605042,0.040701763083595376
+flat_mae,patch,logistic,abide_dx,21,0.000774263682681127,train,0.6723646723646723,0.017309744874948758,0.6579730038892702,0.018452832671581765,0.6582502768549281,0.01773169799557126
+flat_mae,patch,logistic,abide_dx,21,0.000774263682681127,test,0.5887096774193549,0.04177785200071881,0.5740553647201454,0.04370057909705704,0.576155462184874,0.04227325722468146
+flat_mae,patch,logistic,abide_dx,22,0.046415888336127774,train,0.8290598290598291,0.014248603536410246,0.8260725936325721,0.014600688490252695,0.8242894056847545,0.014642178289329923
+flat_mae,patch,logistic,abide_dx,22,0.046415888336127774,test,0.5967741935483871,0.043081544899482405,0.575109649122807,0.047269246976230334,0.5803571428571428,0.04413932884337094
+flat_mae,patch,logistic,abide_dx,23,0.3593813663804626,train,0.9173789173789174,0.010538961392205707,0.9161787593567035,0.01073990595336288,0.9147286821705426,0.01094052398004128
+flat_mae,patch,logistic,abide_dx,23,0.3593813663804626,test,0.5967741935483871,0.044992194952677395,0.58994708994709,0.046026024497452335,0.5898109243697479,0.04552468870036837
+flat_mae,patch,logistic,abide_dx,24,0.046415888336127774,train,0.8148148148148148,0.014788005547585569,0.8112856398421797,0.015198838560204549,0.8093023255813954,0.015205868959525269
+flat_mae,patch,logistic,abide_dx,24,0.046415888336127774,test,0.6048387096774194,0.04403463304477666,0.6017043592264831,0.04417293170351323,0.601890756302521,0.04416192525932249
+flat_mae,patch,logistic,abide_dx,25,0.005994842503189409,train,0.7264957264957265,0.016466748235548594,0.7205970149253731,0.01701122638613265,0.7191583610188261,0.016884888432620912
+flat_mae,patch,logistic,abide_dx,25,0.005994842503189409,test,0.6290322580645161,0.04344978083643979,0.6274817136886102,0.04353433023167341,0.6286764705882353,0.043653220708625845
+flat_mae,patch,logistic,abide_dx,26,0.005994842503189409,train,0.7321937321937322,0.015719189846824414,0.724676018258134,0.01635662134398017,0.7228497600590624,0.016119348636000657
+flat_mae,patch,logistic,abide_dx,26,0.005994842503189409,test,0.5725806451612904,0.04468924962737756,0.5623043623043623,0.04620564046687654,0.5630252100840336,0.0451854926690979
+flat_mae,patch,logistic,abide_dx,27,0.046415888336127774,train,0.8105413105413105,0.015038982038523297,0.8078552833314468,0.01527872189648326,0.8069029162052418,0.015296056643130312
+flat_mae,patch,logistic,abide_dx,27,0.046415888336127774,test,0.5725806451612904,0.045398212064891565,0.5643931861867832,0.04639725160756262,0.5646008403361344,0.045871421786329955
+flat_mae,patch,logistic,abide_dx,28,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,abide_dx,28,1291.5496650148827,test,0.6370967741935484,0.04268278317059871,0.6330637206549615,0.04332768949551267,0.6328781512605042,0.04321975378321963
+flat_mae,patch,logistic,abide_dx,29,0.046415888336127774,train,0.8048433048433048,0.01456864732981116,0.8007181886561003,0.015033265286728471,0.7984865263935031,0.015007451949615315
+flat_mae,patch,logistic,abide_dx,29,0.046415888336127774,test,0.6774193548387096,0.039879824024235046,0.6688034188034189,0.04188124334564985,0.6680672268907563,0.04090223241051902
+flat_mae,patch,logistic,abide_dx,30,0.005994842503189409,train,0.7122507122507122,0.01694076219616888,0.7052749719416387,0.017461167455087984,0.703875968992248,0.01725587200310047
+flat_mae,patch,logistic,abide_dx,30,0.005994842503189409,test,0.6370967741935484,0.04234016894566617,0.6342182890855457,0.04261392435327109,0.634453781512605,0.04252657852203198
+flat_mae,patch,logistic,abide_dx,31,0.046415888336127774,train,0.8162393162393162,0.014769910854079187,0.8129565944790409,0.015115680384265761,0.8111849390919158,0.015132672063206037
+flat_mae,patch,logistic,abide_dx,31,0.046415888336127774,test,0.6451612903225806,0.042695275809892956,0.6375232527238905,0.04452496015712942,0.6370798319327731,0.04367422892699733
+flat_mae,patch,logistic,abide_dx,32,0.046415888336127774,train,0.8205128205128205,0.01422023485822427,0.8177780341485925,0.014498082040798018,0.8165374677002584,0.014528849303628048
+flat_mae,patch,logistic,abide_dx,32,0.046415888336127774,test,0.6370967741935484,0.04212897050827861,0.6317074780542539,0.04261838304020082,0.6313025210084033,0.042234307116611074
+flat_mae,patch,logistic,abide_dx,33,0.046415888336127774,train,0.8190883190883191,0.013842097819626784,0.8161329640346229,0.014144024197177904,0.8146548541897379,0.01418329954816489
+flat_mae,patch,logistic,abide_dx,33,0.046415888336127774,test,0.5725806451612904,0.04284765986953694,0.5643931861867832,0.04379133700030574,0.5646008403361344,0.04322708815632735
+flat_mae,patch,logistic,abide_dx,34,0.3593813663804626,train,0.9330484330484331,0.009739044857283777,0.9321450385918152,0.009898948424039381,0.9310077519379845,0.010028826459785688
+flat_mae,patch,logistic,abide_dx,34,0.3593813663804626,test,0.5403225806451613,0.0432549582729253,0.537888198757764,0.04354100764742364,0.5383403361344539,0.04366090718453375
+flat_mae,patch,logistic,abide_dx,35,0.046415888336127774,train,0.8091168091168092,0.014703339086436949,0.8062083855231064,0.014962704366723404,0.8050203026947212,0.0149667357376967
+flat_mae,patch,logistic,abide_dx,35,0.046415888336127774,test,0.6693548387096774,0.04091259941439645,0.6667322189446083,0.04139352722244033,0.6670168067226891,0.04130047995845927
+flat_mae,patch,logistic,abide_dx,36,0.005994842503189409,train,0.717948717948718,0.01607809711807512,0.7113678649744168,0.01656608460047531,0.7099298634182355,0.01639043941145779
+flat_mae,patch,logistic,abide_dx,36,0.005994842503189409,test,0.6774193548387096,0.041091241335156826,0.6688034188034189,0.04277763648525594,0.6680672268907563,0.04185253933161877
+flat_mae,patch,logistic,abide_dx,37,0.046415888336127774,train,0.8361823361823362,0.014120366328593998,0.833382522604706,0.014508518384888302,0.8316352897748247,0.014641372689916821
+flat_mae,patch,logistic,abide_dx,37,0.046415888336127774,test,0.6532258064516129,0.0397184550451784,0.6448884448884449,0.04115653437550358,0.6444327731092437,0.040399868390830165
+flat_mae,patch,logistic,abide_dx,38,0.005994842503189409,train,0.7108262108262108,0.016777573135978704,0.703948229477205,0.017300964409583227,0.7025839793281654,0.01710826250197527
+flat_mae,patch,logistic,abide_dx,38,0.005994842503189409,test,0.6532258064516129,0.04019436389169068,0.6448884448884449,0.04151157851580024,0.6444327731092437,0.040804104287870346
+flat_mae,patch,logistic,abide_dx,39,0.046415888336127774,train,0.8162393162393162,0.013534250630543096,0.8129565944790409,0.013822063312745285,0.8111849390919158,0.013818937940601222
+flat_mae,patch,logistic,abide_dx,39,0.046415888336127774,test,0.5967741935483871,0.04396794310417141,0.5836690840719849,0.045798049335887776,0.5850840336134454,0.04444940641644906
+flat_mae,patch,logistic,abide_dx,40,0.046415888336127774,train,0.8162393162393162,0.014752659445214882,0.81375978677328,0.014980307597183154,0.8129568106312293,0.015013710817275242
+flat_mae,patch,logistic,abide_dx,40,0.046415888336127774,test,0.6532258064516129,0.03919338576894447,0.6331613347093223,0.042771518458555646,0.6365546218487395,0.04010720712249162
+flat_mae,patch,logistic,abide_dx,41,0.000774263682681127,train,0.6467236467236467,0.01745606411531348,0.6312056737588653,0.01867543600115551,0.6323366555924695,0.017897132485508598
+flat_mae,patch,logistic,abide_dx,41,0.000774263682681127,test,0.6290322580645161,0.03924210121545952,0.6119727891156463,0.04231364298836883,0.6144957983193278,0.04015843813684551
+flat_mae,patch,logistic,abide_dx,42,0.046415888336127774,train,0.8048433048433048,0.013984618298275662,0.8013570034389814,0.01429406650128978,0.7996677740863787,0.014277481250283159
+flat_mae,patch,logistic,abide_dx,42,0.046415888336127774,test,0.6290322580645161,0.04409677419354838,0.6255252100840336,0.044836666351665194,0.6255252100840336,0.044590379240014404
+flat_mae,patch,logistic,abide_dx,43,0.046415888336127774,train,0.8076923076923077,0.014073640373030877,0.8039491513151771,0.014460420101863084,0.8019564414913252,0.014452952581534347
+flat_mae,patch,logistic,abide_dx,43,0.046415888336127774,test,0.717741935483871,0.042134993773728924,0.710955710955711,0.04364201444051302,0.7095588235294117,0.04301362189664417
+flat_mae,patch,logistic,abide_dx,44,0.000774263682681127,train,0.6552706552706553,0.01728865706665264,0.640128117135667,0.018473523846953072,0.6409745293466224,0.01772114591718627
+flat_mae,patch,logistic,abide_dx,44,0.000774263682681127,test,0.6129032258064516,0.03997531431299166,0.5921052631578947,0.04397497149481937,0.5966386554621849,0.04103224726333098
+flat_mae,patch,logistic,abide_dx,45,0.046415888336127774,train,0.8233618233618234,0.014668026631684797,0.8205417886447037,0.015008825082479246,0.8191214470284238,0.015086857569313085
+flat_mae,patch,logistic,abide_dx,45,0.046415888336127774,test,0.5806451612903226,0.041908800443708635,0.5643243243243243,0.043704722143080695,0.5672268907563025,0.04222818336737672
+flat_mae,patch,logistic,abide_dx,46,0.046415888336127774,train,0.8262108262108262,0.014397080084802706,0.8233065953654188,0.014697865822600197,0.8217054263565892,0.014741011759662897
+flat_mae,patch,logistic,abide_dx,46,0.046415888336127774,test,0.5725806451612904,0.04469744223500621,0.5691904293674206,0.045167292227623355,0.569327731092437,0.045194090900702234
+flat_mae,patch,logistic,abide_dx,47,0.046415888336127774,train,0.8148148148148148,0.015179620859587782,0.8119932098358494,0.01544736288400603,0.8107788851974899,0.015453680789704872
+flat_mae,patch,logistic,abide_dx,47,0.046415888336127774,test,0.6129032258064516,0.04297369909299979,0.607905138339921,0.0437169917842997,0.6076680672268908,0.0434851656123859
+flat_mae,patch,logistic,abide_dx,48,0.046415888336127774,train,0.8205128205128205,0.01398314136614389,0.8176473013647796,0.014258152616334653,0.8162421557770395,0.014264240159438838
+flat_mae,patch,logistic,abide_dx,48,0.046415888336127774,test,0.6451612903225806,0.04267607524765116,0.6391534391534391,0.04381060237156751,0.6386554621848739,0.04334719713144787
+flat_mae,patch,logistic,abide_dx,49,0.000774263682681127,train,0.6709401709401709,0.017217957851632108,0.6571069709022399,0.018147002702927635,0.6572535991140642,0.01751083365900478
+flat_mae,patch,logistic,abide_dx,49,0.000774263682681127,test,0.6048387096774194,0.039736266353089376,0.578494623655914,0.04418238570013005,0.5861344537815126,0.04059818158704357
+flat_mae,patch,logistic,abide_dx,50,0.046415888336127774,train,0.8162393162393162,0.013903051001574334,0.813098655791366,0.014229187660698616,0.8114802510151347,0.014255921715475118
+flat_mae,patch,logistic,abide_dx,50,0.046415888336127774,test,0.5967741935483871,0.04465999427283362,0.5810810810810811,0.04713720256675513,0.5835084033613446,0.04528510431816933
+flat_mae,patch,logistic,abide_dx,51,0.046415888336127774,train,0.8262108262108262,0.014700204274575215,0.8231738035264484,0.015078082138637632,0.8214101144333703,0.015121153166270716
+flat_mae,patch,logistic,abide_dx,51,0.046415888336127774,test,0.5967741935483871,0.041211631163837505,0.5810810810810811,0.04356186821528533,0.5835084033613446,0.04182421273594904
+flat_mae,patch,logistic,abide_dx,52,0.005994842503189409,train,0.7222222222222222,0.016479023571547197,0.7134243994347622,0.017286387941185366,0.7117386489479514,0.01691431917440416
+flat_mae,patch,logistic,abide_dx,52,0.005994842503189409,test,0.6290322580645161,0.04385020856099628,0.6242424242424243,0.04462218152686849,0.6239495798319328,0.04452530415289145
+flat_mae,patch,logistic,abide_dx,53,0.046415888336127774,train,0.8190883190883191,0.014420001852908272,0.8152643062724434,0.014827164070846322,0.8128829826504245,0.014779086244230381
+flat_mae,patch,logistic,abide_dx,53,0.046415888336127774,test,0.6290322580645161,0.04276284029254295,0.6242424242424243,0.04342207004187337,0.6239495798319328,0.043176670429289805
+flat_mae,patch,logistic,abide_dx,54,0.005994842503189409,train,0.7250712250712251,0.016750247000801768,0.719719897806142,0.01716177287397828,0.7184569952011812,0.017046344699817206
+flat_mae,patch,logistic,abide_dx,54,0.005994842503189409,test,0.6209677419354839,0.04510232037057044,0.6137071651090342,0.0463636833031733,0.6134453781512605,0.04571655086801859
+flat_mae,patch,logistic,abide_dx,55,0.005994842503189409,train,0.7207977207977208,0.017173415793444505,0.7137698042870457,0.017810733704264333,0.712218530823182,0.01759028040594031
+flat_mae,patch,logistic,abide_dx,55,0.005994842503189409,test,0.6129032258064516,0.042043706509107076,0.6003223207091055,0.04395201151796931,0.6013655462184874,0.04262738162186557
+flat_mae,patch,logistic,abide_dx,56,0.005994842503189409,train,0.7165242165242165,0.016547743100567787,0.7084249084249084,0.01721236103531624,0.7068660022148394,0.016902526070245796
+flat_mae,patch,logistic,abide_dx,56,0.005994842503189409,test,0.6612903225806451,0.03989701558058263,0.6502820306204673,0.04235785137587265,0.6502100840336134,0.04101881706001485
+flat_mae,patch,logistic,abide_dx,57,0.046415888336127774,train,0.8219373219373219,0.013685425911501679,0.8190285079080932,0.01401972479330757,0.8175341454411222,0.014095154826991082
+flat_mae,patch,logistic,abide_dx,57,0.046415888336127774,test,0.6532258064516129,0.039863512587277675,0.6480760345851759,0.0406830702303827,0.6475840336134454,0.04038080695452615
+flat_mae,patch,logistic,abide_dx,58,0.046415888336127774,train,0.8262108262108262,0.01400073049449309,0.8228988312365071,0.014405602454043664,0.8208194905869325,0.014446985523044019
+flat_mae,patch,logistic,abide_dx,58,0.046415888336127774,test,0.6129032258064516,0.04346805912865296,0.6112852664576802,0.04353824942154756,0.6123949579831933,0.04371216285030784
+flat_mae,patch,logistic,abide_dx,59,0.046415888336127774,train,0.811965811965812,0.015300966006035414,0.8088235294117647,0.015612507265701651,0.8073089700996677,0.01562627123933995
+flat_mae,patch,logistic,abide_dx,59,0.046415888336127774,test,0.6290322580645161,0.04456227155100645,0.6266038229903116,0.0448244940169691,0.6271008403361344,0.044791266846952324
+flat_mae,patch,logistic,abide_dx,60,0.046415888336127774,train,0.8133903133903134,0.015491394766894546,0.8103418762876817,0.015812811178201536,0.8088962716869693,0.015824577725537563
+flat_mae,patch,logistic,abide_dx,60,0.046415888336127774,test,0.5887096774193549,0.04398425541310865,0.5865315462569467,0.04417646002498987,0.5871848739495797,0.04429141821564067
+flat_mae,patch,logistic,abide_dx,61,0.046415888336127774,train,0.8276353276353277,0.014651157190122673,0.8251916487451509,0.014888629797909905,0.8241786637135474,0.014922397315113632
+flat_mae,patch,logistic,abide_dx,61,0.046415888336127774,test,0.5725806451612904,0.04423782512393834,0.5623043623043623,0.04515594725345051,0.5630252100840336,0.04431821774533121
+flat_mae,patch,logistic,abide_dx,62,0.000774263682681127,train,0.6709401709401709,0.01686058949891694,0.6594401496966376,0.017910652706293988,0.6590254706533776,0.01737571648037623
+flat_mae,patch,logistic,abide_dx,62,0.000774263682681127,test,0.6129032258064516,0.04186993636584508,0.5978378378378378,0.04441792401715795,0.5997899159663866,0.04274082997220342
+flat_mae,patch,logistic,abide_dx,63,0.046415888336127774,train,0.8304843304843305,0.015123873594540265,0.8280810429807682,0.015380056755930108,0.8270579549649317,0.015418007125727665
+flat_mae,patch,logistic,abide_dx,63,0.046415888336127774,test,0.6209677419354839,0.044036972452337696,0.6203504657677024,0.044150532010916364,0.6228991596638656,0.04416149355513142
+flat_mae,patch,logistic,abide_dx,64,0.046415888336127774,train,0.8091168091168092,0.015001372761538975,0.806069352245083,0.015261617600720697,0.8047249907715024,0.015257651478396738
+flat_mae,patch,logistic,abide_dx,64,0.046415888336127774,test,0.6451612903225806,0.04257984786643588,0.6391534391534391,0.04374301919437287,0.6386554621848739,0.0432648698667387
+flat_mae,patch,logistic,abide_dx,65,0.046415888336127774,train,0.8219373219373219,0.014542471530607202,0.8190285079080932,0.014897386934271454,0.8175341454411222,0.014987215102487492
+flat_mae,patch,logistic,abide_dx,65,0.046415888336127774,test,0.6048387096774194,0.044480467249134785,0.5931704050887178,0.04665746390703013,0.5940126050420168,0.04535424956033638
+flat_mae,patch,logistic,abide_dx,66,0.046415888336127774,train,0.8290598290598291,0.014302485491162328,0.8266951942067149,0.014526288140236232,0.825765965300849,0.014546364991278857
+flat_mae,patch,logistic,abide_dx,66,0.046415888336127774,test,0.6129032258064516,0.04579966726224051,0.6092436974789917,0.046368699932556025,0.6092436974789917,0.04625389298721849
+flat_mae,patch,logistic,abide_dx,67,0.046415888336127774,train,0.8176638176638177,0.014825843747029742,0.8143348320550119,0.015204025511261092,0.8124769287559985,0.015246896670638412
+flat_mae,patch,logistic,abide_dx,67,0.046415888336127774,test,0.6532258064516129,0.04194602845096986,0.6448884448884449,0.043576932006045295,0.6444327731092437,0.04268664970847627
+flat_mae,patch,logistic,abide_dx,68,0.000774263682681127,train,0.6581196581196581,0.01681654092652144,0.6443768996960486,0.017853916458865103,0.6447397563676633,0.017221780109342926
+flat_mae,patch,logistic,abide_dx,68,0.000774263682681127,test,0.5967741935483871,0.03919394331408622,0.5810810810810811,0.04189347456698179,0.5835084033613446,0.039982221811747945
+flat_mae,patch,logistic,abide_dx,69,0.005994842503189409,train,0.7051282051282052,0.016007562072030423,0.6957889778615167,0.016815559434444613,0.6944629014396456,0.016458173363628197
+flat_mae,patch,logistic,abide_dx,69,0.005994842503189409,test,0.6854838709677419,0.0407685020527538,0.6721578198088265,0.043651566533400986,0.6722689075630253,0.041829512970162006
+flat_mae,patch,logistic,abide_dx,70,0.046415888336127774,train,0.8205128205128205,0.015012298429850175,0.8176473013647796,0.015351469695808152,0.8162421557770395,0.01541843014295104
+flat_mae,patch,logistic,abide_dx,70,0.046415888336127774,test,0.6209677419354839,0.04453546506193729,0.6153389215233318,0.045388848751803205,0.6150210084033614,0.04509356542582331
+flat_mae,patch,logistic,abide_dx,71,0.046415888336127774,train,0.8190883190883191,0.014368339097894181,0.8158564922390557,0.0146592371191468,0.8140642303433001,0.014616513646363013
+flat_mae,patch,logistic,abide_dx,71,0.046415888336127774,test,0.6532258064516129,0.03931641103787403,0.6429862738533645,0.04107719866164188,0.6428571428571428,0.0399598841213838
+flat_mae,patch,logistic,abide_dx,72,0.3593813663804626,train,0.9145299145299145,0.010308706102825698,0.9134615384615385,0.010453765351307034,0.9127353266888151,0.010528649519402139
+flat_mae,patch,logistic,abide_dx,72,0.3593813663804626,test,0.5241935483870968,0.04360950012968361,0.5234186697934988,0.043613506635374236,0.5252100840336134,0.043709194454083354
+flat_mae,patch,logistic,abide_dx,73,0.046415888336127774,train,0.8048433048433048,0.015033486699902647,0.8020764948602122,0.015237878080858051,0.8011443337024733,0.015212191593777348
+flat_mae,patch,logistic,abide_dx,73,0.046415888336127774,test,0.6290322580645161,0.03969965283177649,0.6266038229903116,0.039870109268220065,0.6271008403361344,0.0398788773951557
+flat_mae,patch,logistic,abide_dx,74,0.005994842503189409,train,0.7136752136752137,0.016454233127379835,0.7060589332027873,0.017123270082534062,0.7045773348098929,0.01688693708699315
+flat_mae,patch,logistic,abide_dx,74,0.005994842503189409,test,0.6290322580645161,0.04281820340911307,0.6210470369386127,0.04412702440007403,0.6207983193277311,0.04343186865320686
+flat_mae,patch,logistic,abide_dx,75,0.046415888336127774,train,0.8148148148148148,0.015245762282913695,0.8118583268049313,0.015617056117240872,0.810483573274271,0.015699314896695623
+flat_mae,patch,logistic,abide_dx,75,0.046415888336127774,test,0.6451612903225806,0.0409200891258643,0.6356837606837606,0.04256190347344885,0.6355042016806722,0.04166504903969842
+flat_mae,patch,logistic,abide_dx,76,0.046415888336127774,train,0.8076923076923077,0.014371618590603886,0.8041047668460186,0.014817505863391412,0.8022517534145441,0.014886670526488574
+flat_mae,patch,logistic,abide_dx,76,0.046415888336127774,test,0.6693548387096774,0.04075176751591489,0.6575739206573719,0.04300275716343429,0.657563025210084,0.041464770753086654
+flat_mae,patch,logistic,abide_dx,77,0.046415888336127774,train,0.8276353276353277,0.013373413290946419,0.8251916487451509,0.013594323436510304,0.8241786637135474,0.013626169788903388
+flat_mae,patch,logistic,abide_dx,77,0.046415888336127774,test,0.532258064516129,0.04507058575514828,0.5311603650586701,0.045220500564835,0.532563025210084,0.04539670070997022
+flat_mae,patch,logistic,abide_dx,78,0.046415888336127774,train,0.8262108262108262,0.013563234251677646,0.8238067807768268,0.013759774345971107,0.8228866740494647,0.01375667232252695
+flat_mae,patch,logistic,abide_dx,78,0.046415888336127774,test,0.6129032258064516,0.042833860369720776,0.5978378378378378,0.044704063706268816,0.5997899159663866,0.04322538013830674
+flat_mae,patch,logistic,abide_dx,79,0.046415888336127774,train,0.8319088319088319,0.014053904368009714,0.8290998217468806,0.014396525460207378,0.8274640088593577,0.014464465460585416
+flat_mae,patch,logistic,abide_dx,79,0.046415888336127774,test,0.6048387096774194,0.043808123275625055,0.5931704050887178,0.04613649124232754,0.5940126050420168,0.04466083182381372
+flat_mae,patch,logistic,abide_dx,80,0.3593813663804626,train,0.9358974358974359,0.008763054829346162,0.9351168164313222,0.008881332956650568,0.9344776670358066,0.008995773116624847
+flat_mae,patch,logistic,abide_dx,80,0.3593813663804626,test,0.5967741935483871,0.04270167001732451,0.5941345902068604,0.042936531142214185,0.5945378151260504,0.0431073123937748
+flat_mae,patch,logistic,abide_dx,81,0.046415888336127774,train,0.8162393162393162,0.015089676822395159,0.8128112216528621,0.015435759159426497,0.810889627168697,0.015421296653370555
+flat_mae,patch,logistic,abide_dx,81,0.046415888336127774,test,0.6129032258064516,0.04402541598737713,0.5951020408163266,0.047541045925511,0.5982142857142857,0.04497705880162714
+flat_mae,patch,logistic,abide_dx,82,0.005994842503189409,train,0.7136752136752137,0.017119025154253457,0.7073756467819036,0.017652347001438858,0.7060538944259874,0.01749311908083878
+flat_mae,patch,logistic,abide_dx,82,0.005994842503189409,test,0.6129032258064516,0.04179572243844966,0.6003223207091055,0.04422265172952331,0.6013655462184874,0.04276559544832569
+flat_mae,patch,logistic,abide_dx,83,0.046415888336127774,train,0.8133903133903134,0.014838885496827731,0.8100566967190259,0.01518374206501488,0.8083056478405315,0.015176256060351298
+flat_mae,patch,logistic,abide_dx,83,0.046415888336127774,test,0.6129032258064516,0.04304336801014306,0.6092436974789917,0.043478191221444244,0.6092436974789917,0.04344047080429191
+flat_mae,patch,logistic,abide_dx,84,0.3593813663804626,train,0.9188034188034188,0.010225572507099332,0.9175368635815636,0.010439908646266306,0.9157253599114064,0.010623231528864805
+flat_mae,patch,logistic,abide_dx,84,0.3593813663804626,test,0.6048387096774194,0.04528888656854025,0.6004471624909581,0.04571977995275497,0.6003151260504203,0.04551317451707635
+flat_mae,patch,logistic,abide_dx,85,0.046415888336127774,train,0.8105413105413105,0.013941817185813424,0.8078552833314468,0.014201135637855347,0.8069029162052418,0.014259065477644377
+flat_mae,patch,logistic,abide_dx,85,0.046415888336127774,test,0.6290322580645161,0.04127939308473816,0.6210470369386127,0.04241401945253636,0.6207983193277311,0.04176528047883907
+flat_mae,patch,logistic,abide_dx,86,0.046415888336127774,train,0.8290598290598291,0.013715115990033631,0.8259389050515737,0.014088501609144785,0.8239940937615357,0.014165452617888272
+flat_mae,patch,logistic,abide_dx,86,0.046415888336127774,test,0.6290322580645161,0.04500658698466618,0.6266038229903116,0.045471849995903604,0.6271008403361344,0.04546393078776404
+flat_mae,patch,logistic,abide_dx,87,0.046415888336127774,train,0.8005698005698005,0.015280797754368148,0.7967691505992704,0.01573963982841928,0.7949058693244739,0.015761117124632976
+flat_mae,patch,logistic,abide_dx,87,0.046415888336127774,test,0.6612903225806451,0.04085137142552884,0.6569169960474308,0.04127147306952389,0.6565126050420168,0.04102570972715738
+flat_mae,patch,logistic,abide_dx,88,0.046415888336127774,train,0.811965811965812,0.014432497292878455,0.8088235294117647,0.01475480324954852,0.8073089700996677,0.014779863346403386
+flat_mae,patch,logistic,abide_dx,88,0.046415888336127774,test,0.5887096774193549,0.04596567907086079,0.5873947935016637,0.04615058046713721,0.5887605042016807,0.04619853584836723
+flat_mae,patch,logistic,abide_dx,89,0.046415888336127774,train,0.8290598290598291,0.01478972096761629,0.8264552706177072,0.015092206581189978,0.8251753414544112,0.015153844737883159
+flat_mae,patch,logistic,abide_dx,89,0.046415888336127774,test,0.5967741935483871,0.04131666246557779,0.5963541666666667,0.041389998778551144,0.5992647058823529,0.04156590435483708
+flat_mae,patch,logistic,abide_dx,90,0.046415888336127774,train,0.8048433048433048,0.014621829826821172,0.8017991282574423,0.014890026892427537,0.8005537098560354,0.014892245759094211
+flat_mae,patch,logistic,abide_dx,90,0.046415888336127774,test,0.6451612903225806,0.041285962471406985,0.6313513513513513,0.04354651373630016,0.6323529411764706,0.04198956385720837
+flat_mae,patch,logistic,abide_dx,91,0.046415888336127774,train,0.8105413105413105,0.014931052598046111,0.8074463324142113,0.015206590857396491,0.806016980435585,0.015184422468709031
+flat_mae,patch,logistic,abide_dx,91,0.046415888336127774,test,0.6129032258064516,0.04043778381713843,0.6063492063492064,0.041450922687933905,0.60609243697479,0.04103487482293122
+flat_mae,patch,logistic,abide_dx,92,0.046415888336127774,train,0.8233618233618234,0.014454517523663973,0.8207959682798489,0.014722767280470804,0.8197120708748615,0.014774491465283795
+flat_mae,patch,logistic,abide_dx,92,0.046415888336127774,test,0.5483870967741935,0.04340927601949774,0.5441176470588236,0.04347175593474291,0.5441176470588236,0.04334445608826659
+flat_mae,patch,logistic,abide_dx,93,0.046415888336127774,train,0.8148148148148148,0.014187863704510126,0.8114338138058714,0.014535431737444465,0.8095976375046142,0.014556876803373767
+flat_mae,patch,logistic,abide_dx,93,0.046415888336127774,test,0.5564516129032258,0.043894855332508066,0.5529334644378892,0.04385612774890223,0.553046218487395,0.043792677830717164
+flat_mae,patch,logistic,abide_dx,94,0.000774263682681127,train,0.6538461538461539,0.01751019401447506,0.6361057129265603,0.019263152126182344,0.6379106681432263,0.018184910487316137
+flat_mae,patch,logistic,abide_dx,94,0.000774263682681127,test,0.5967741935483871,0.042042280264808704,0.5836690840719849,0.044153889673427024,0.5850840336134454,0.04273000063675474
+flat_mae,patch,logistic,abide_dx,95,0.046415888336127774,train,0.8048433048433048,0.015115483622675157,0.802210006108057,0.015328926219235674,0.8014396456256921,0.015315812436326133
+flat_mae,patch,logistic,abide_dx,95,0.046415888336127774,test,0.6048387096774194,0.043058326847732004,0.5880957223239103,0.04592739007657337,0.5908613445378151,0.04388065219861184
+flat_mae,patch,logistic,abide_dx,96,0.046415888336127774,train,0.8105413105413105,0.014946129350932863,0.8075860150236482,0.015288111700606288,0.806312292358804,0.015368376733031314
+flat_mae,patch,logistic,abide_dx,96,0.046415888336127774,test,0.6693548387096774,0.042988549040357406,0.6630211440312852,0.04456023013336834,0.6622899159663866,0.04383846685208054
+flat_mae,patch,logistic,abide_dx,97,0.3593813663804626,train,0.9344729344729344,0.009165564078313005,0.9333778966131907,0.00939403075863723,0.9311184939091917,0.00971089743687654
+flat_mae,patch,logistic,abide_dx,97,0.3593813663804626,test,0.6370967741935484,0.04334339746787591,0.6330637206549615,0.04419768212279903,0.6328781512605042,0.04400716331604586
+flat_mae,patch,logistic,abide_dx,98,0.046415888336127774,train,0.8162393162393162,0.015386270618986444,0.8126625223678359,0.015839818444453348,0.810594315245478,0.015849388975963694
+flat_mae,patch,logistic,abide_dx,98,0.046415888336127774,test,0.6693548387096774,0.04248585490260742,0.6595915634415801,0.04456710340963384,0.6591386554621849,0.04342151480565184
+flat_mae,patch,logistic,abide_dx,99,0.046415888336127774,train,0.8034188034188035,0.01468735742719744,0.799983482677458,0.015013861622428811,0.7983757844222961,0.01499287997269589
+flat_mae,patch,logistic,abide_dx,99,0.046415888336127774,test,0.6693548387096774,0.03895119214359001,0.6575739206573719,0.041270982845458284,0.657563025210084,0.039917357671463004
+flat_mae,patch,logistic,abide_dx,100,0.000774263682681127,train,0.6666666666666666,0.01644684311805553,0.6515966472105335,0.01787485829874074,0.6521963824289405,0.017021809165914788
+flat_mae,patch,logistic,abide_dx,100,0.000774263682681127,test,0.6370967741935484,0.04103809998482146,0.6190346145968457,0.04485951415033147,0.6218487394957983,0.042330725111534864

data_scaling/n400_1/eval_v2/abide_dx__patch__logistic/log.txt

@@ -0,0 +1,252 @@
+fMRI foundation model logistic probe eval
+version: 0.1.dev66+g7ddd3aa04
+sha: 58906bf7243fb545e1349221e6921a1797e2e666, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-02-26 17:14:52
+config:
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (abide_dx patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_1/eval_v2/abide_dx__patch__logistic
+model: flat_mae
+representation: patch
+dataset: abide_dx
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/abide_dx__patch__logistic
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=True, reg_tokens=0, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: abide_dx (flat)
+train (n=578):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'site', 'dataset', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 578
+}),
+    labels=['Autism' 'Control'],
+    counts=[260 318]
+)
+
+validation (n=124):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'site', 'dataset', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 124
+}),
+    labels=['Autism' 'Control'],
+    counts=[54 70]
+)
+
+test (n=124):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'site', 'dataset', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 124
+}),
+    labels=['Autism' 'Control'],
+    counts=[57 67]
+)
+
+extracting features for all splits
+extract (train)  [  0/289]  eta: 0:16:23    time: 3.4023  data: 2.6918  max mem: 2698
+extract (train)  [ 20/289]  eta: 0:01:29    time: 0.1803  data: 0.0530  max mem: 2851
+extract (train)  [ 40/289]  eta: 0:01:04    time: 0.1770  data: 0.0452  max mem: 2851
+extract (train)  [ 60/289]  eta: 0:00:53    time: 0.1797  data: 0.0487  max mem: 2851
+extract (train)  [ 80/289]  eta: 0:00:45    time: 0.1676  data: 0.0424  max mem: 2851
+extract (train)  [100/289]  eta: 0:00:39    time: 0.1686  data: 0.0437  max mem: 2851
+extract (train)  [120/289]  eta: 0:00:34    time: 0.1766  data: 0.0463  max mem: 2851
+extract (train)  [140/289]  eta: 0:00:29    time: 0.1932  data: 0.0536  max mem: 2851
+extract (train)  [160/289]  eta: 0:00:25    time: 0.1727  data: 0.0450  max mem: 2851
+extract (train)  [180/289]  eta: 0:00:21    time: 0.1597  data: 0.0433  max mem: 2851
+extract (train)  [200/289]  eta: 0:00:16    time: 0.1661  data: 0.0456  max mem: 2851
+extract (train)  [220/289]  eta: 0:00:13    time: 0.1785  data: 0.0487  max mem: 2851
+extract (train)  [240/289]  eta: 0:00:09    time: 0.1681  data: 0.0451  max mem: 2851
+extract (train)  [260/289]  eta: 0:00:05    time: 0.1658  data: 0.0455  max mem: 2851
+extract (train)  [280/289]  eta: 0:00:01    time: 0.1333  data: 0.0318  max mem: 2851
+extract (train)  [288/289]  eta: 0:00:00    time: 0.1321  data: 0.0319  max mem: 2851
+extract (train) Total time: 0:00:52 (0.1818 s / it)
+extract (validation)  [ 0/62]  eta: 0:03:20    time: 3.2295  data: 3.1127  max mem: 2851
+extract (validation)  [20/62]  eta: 0:00:13    time: 0.1841  data: 0.0550  max mem: 2851
+extract (validation)  [40/62]  eta: 0:00:05    time: 0.1425  data: 0.0385  max mem: 2851
+extract (validation)  [60/62]  eta: 0:00:00    time: 0.1363  data: 0.0345  max mem: 2851
+extract (validation)  [61/62]  eta: 0:00:00    time: 0.1363  data: 0.0342  max mem: 2851
+extract (validation) Total time: 0:00:12 (0.2081 s / it)
+extract (test)  [ 0/62]  eta: 0:03:25    time: 3.3141  data: 3.1140  max mem: 2851
+extract (test)  [20/62]  eta: 0:00:14    time: 0.1921  data: 0.0552  max mem: 2851
+extract (test)  [40/62]  eta: 0:00:05    time: 0.1538  data: 0.0410  max mem: 2851
+extract (test)  [60/62]  eta: 0:00:00    time: 0.1322  data: 0.0317  max mem: 2851
+extract (test)  [61/62]  eta: 0:00:00    time: 0.1322  data: 0.0317  max mem: 2851
+extract (test) Total time: 0:00:13 (0.2137 s / it)
+feature extraction time: 0:01:18
+train features: (578, 768)
+validation features: (124, 768)
+test features: (124, 768)
+evaluating fixed splits
+eval results (fixed splits):
+
+| model    | repr   | clf      | dataset   | trial   |       C | split   |     acc |   acc_std |      f1 |    f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:----------|:--------|--------:|:--------|--------:|----------:|--------:|----------:|--------:|-----------:|
+| flat_mae | patch  | logistic | abide_dx  |         | 0.35938 | train   | 0.93162 | 0.0095783 | 0.93072 | 0.0097338 | 0.92994 |  0.0099011 |
+| flat_mae | patch  | logistic | abide_dx  |         | 0.35938 | test    | 0.52419 | 0.041284  | 0.51275 | 0.041847  | 0.51519 |  0.041199  |
+
+
+evaluating random splits (n=100)
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 1, "C": 0.005994842503189409, "split": "test", "acc": 0.6774193548387096, "acc_std": 0.041113137346047476, "f1": 0.6688034188034189, "f1_std": 0.04295419279844657, "bacc": 0.6680672268907563, "bacc_std": 0.04203937864030787}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 2, "C": 0.3593813663804626, "split": "test", "acc": 0.6209677419354839, "acc_std": 0.04211467917333632, "f1": 0.6179613241560145, "f1_std": 0.042344171626269744, "bacc": 0.618172268907563, "bacc_std": 0.04239843385697718}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 3, "C": 0.046415888336127774, "split": "test", "acc": 0.6048387096774194, "acc_std": 0.04441990535419158, "f1": 0.5989703649924097, "f1_std": 0.04504116366441694, "bacc": 0.5987394957983193, "bacc_std": 0.04472068721015768}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 4, "C": 0.046415888336127774, "split": "test", "acc": 0.7016129032258065, "acc_std": 0.03831011527471725, "f1": 0.6944388944388944, "f1_std": 0.039845593860878076, "bacc": 0.6932773109243697, "bacc_std": 0.03926182668976251}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 5, "C": 0.046415888336127774, "split": "test", "acc": 0.5967741935483871, "acc_std": 0.04129690912688167, "f1": 0.5915678524374176, "f1_std": 0.04169908351233194, "bacc": 0.5913865546218487, "bacc_std": 0.04146902108947524}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 6, "C": 0.046415888336127774, "split": "test", "acc": 0.6370967741935484, "acc_std": 0.04635304281882685, "f1": 0.6342182890855457, "f1_std": 0.046843179294802285, "bacc": 0.634453781512605, "bacc_std": 0.04677036162353719}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 7, "C": 0.005994842503189409, "split": "test", "acc": 0.5887096774193549, "acc_std": 0.04462083000509283, "f1": 0.5765651155005022, "f1_std": 0.046826300397298085, "bacc": 0.5777310924369747, "bacc_std": 0.04541367312022607}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 8, "C": 0.3593813663804626, "split": "test", "acc": 0.5645161290322581, "acc_std": 0.04229725774192688, "f1": 0.5588932806324111, "f1_std": 0.04293049709371854, "bacc": 0.5588235294117647, "bacc_std": 0.04267437159659311}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 9, "C": 0.3593813663804626, "split": "test", "acc": 0.6048387096774194, "acc_std": 0.04283920763410549, "f1": 0.5972691721349506, "f1_std": 0.04353892823111837, "bacc": 0.5971638655462186, "bacc_std": 0.04304172193901062}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 10, "C": 0.046415888336127774, "split": "test", "acc": 0.6612903225806451, "acc_std": 0.040930955671154394, "f1": 0.6502820306204673, "f1_std": 0.043394048096207456, "bacc": 0.6502100840336134, "bacc_std": 0.042008753482431}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 11, "C": 0.046415888336127774, "split": "test", "acc": 0.6209677419354839, "acc_std": 0.042008679916720464, "f1": 0.6167554415729598, "f1_std": 0.04245925815837763, "bacc": 0.6165966386554622, "bacc_std": 0.0423885638229868}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 12, "C": 0.005994842503189409, "split": "test", "acc": 0.5887096774193549, "acc_std": 0.044167039376957366, "f1": 0.5841388834089565, "f1_std": 0.04458292141122282, "bacc": 0.5840336134453781, "bacc_std": 0.04437302500699791}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 13, "C": 0.3593813663804626, "split": "test", "acc": 0.5725806451612904, "acc_std": 0.04249018678406638, "f1": 0.5703170970905524, "f1_std": 0.042903984753441864, "bacc": 0.5709033613445378, "bacc_std": 0.04313978178191481}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 14, "C": 0.3593813663804626, "split": "test", "acc": 0.6048387096774194, "acc_std": 0.04310701056721079, "f1": 0.5907590759075907, "f1_std": 0.0453912057515939, "bacc": 0.592436974789916, "bacc_std": 0.04383432933102287}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 15, "C": 0.046415888336127774, "split": "test", "acc": 0.6532258064516129, "acc_std": 0.039270305166193877, "f1": 0.6480760345851759, "f1_std": 0.039930931836775055, "bacc": 0.6475840336134454, "bacc_std": 0.03964505782892502}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 16, "C": 0.046415888336127774, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.04254916663914516, "f1": 0.6112852664576802, "f1_std": 0.042812067293305056, "bacc": 0.6123949579831933, "bacc_std": 0.04292760821989454}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 17, "C": 0.046415888336127774, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.043490570939756265, "f1": 0.6045708211533352, "f1_std": 0.04493554455811725, "bacc": 0.6045168067226891, "bacc_std": 0.04407857663726617}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 18, "C": 0.3593813663804626, "split": "test", "acc": 0.6290322580645161, "acc_std": 0.043240642013598486, "f1": 0.6242424242424243, "f1_std": 0.04386617606615634, "bacc": 0.6239495798319328, "bacc_std": 0.043601468943845846}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 19, "C": 0.005994842503189409, "split": "test", "acc": 0.5725806451612904, "acc_std": 0.04351731268865347, "f1": 0.5478500171998624, "f1_std": 0.04790838098260214, "bacc": 0.5551470588235294, "bacc_std": 0.04451250840553695}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 20, "C": 0.046415888336127774, "split": "test", "acc": 0.6532258064516129, "acc_std": 0.03989847613008047, "f1": 0.6408702094699266, "f1_std": 0.04215016598399647, "bacc": 0.641281512605042, "bacc_std": 0.040701763083595376}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 21, "C": 0.000774263682681127, "split": "test", "acc": 0.5887096774193549, "acc_std": 0.04177785200071881, "f1": 0.5740553647201454, "f1_std": 0.04370057909705704, "bacc": 0.576155462184874, "bacc_std": 0.04227325722468146}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 22, "C": 0.046415888336127774, "split": "test", "acc": 0.5967741935483871, "acc_std": 0.043081544899482405, "f1": 0.575109649122807, "f1_std": 0.047269246976230334, "bacc": 0.5803571428571428, "bacc_std": 0.04413932884337094}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 23, "C": 0.3593813663804626, "split": "test", "acc": 0.5967741935483871, "acc_std": 0.044992194952677395, "f1": 0.58994708994709, "f1_std": 0.046026024497452335, "bacc": 0.5898109243697479, "bacc_std": 0.04552468870036837}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 24, "C": 0.046415888336127774, "split": "test", "acc": 0.6048387096774194, "acc_std": 0.04403463304477666, "f1": 0.6017043592264831, "f1_std": 0.04417293170351323, "bacc": 0.601890756302521, "bacc_std": 0.04416192525932249}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 25, "C": 0.005994842503189409, "split": "test", "acc": 0.6290322580645161, "acc_std": 0.04344978083643979, "f1": 0.6274817136886102, "f1_std": 0.04353433023167341, "bacc": 0.6286764705882353, "bacc_std": 0.043653220708625845}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 26, "C": 0.005994842503189409, "split": "test", "acc": 0.5725806451612904, "acc_std": 0.04468924962737756, "f1": 0.5623043623043623, "f1_std": 0.04620564046687654, "bacc": 0.5630252100840336, "bacc_std": 0.0451854926690979}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 27, "C": 0.046415888336127774, "split": "test", "acc": 0.5725806451612904, "acc_std": 0.045398212064891565, "f1": 0.5643931861867832, "f1_std": 0.04639725160756262, "bacc": 0.5646008403361344, "bacc_std": 0.045871421786329955}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 28, "C": 1291.5496650148827, "split": "test", "acc": 0.6370967741935484, "acc_std": 0.04268278317059871, "f1": 0.6330637206549615, "f1_std": 0.04332768949551267, "bacc": 0.6328781512605042, "bacc_std": 0.04321975378321963}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 29, "C": 0.046415888336127774, "split": "test", "acc": 0.6774193548387096, "acc_std": 0.039879824024235046, "f1": 0.6688034188034189, "f1_std": 0.04188124334564985, "bacc": 0.6680672268907563, "bacc_std": 0.04090223241051902}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 30, "C": 0.005994842503189409, "split": "test", "acc": 0.6370967741935484, "acc_std": 0.04234016894566617, "f1": 0.6342182890855457, "f1_std": 0.04261392435327109, "bacc": 0.634453781512605, "bacc_std": 0.04252657852203198}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 31, "C": 0.046415888336127774, "split": "test", "acc": 0.6451612903225806, "acc_std": 0.042695275809892956, "f1": 0.6375232527238905, "f1_std": 0.04452496015712942, "bacc": 0.6370798319327731, "bacc_std": 0.04367422892699733}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 32, "C": 0.046415888336127774, "split": "test", "acc": 0.6370967741935484, "acc_std": 0.04212897050827861, "f1": 0.6317074780542539, "f1_std": 0.04261838304020082, "bacc": 0.6313025210084033, "bacc_std": 0.042234307116611074}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 33, "C": 0.046415888336127774, "split": "test", "acc": 0.5725806451612904, "acc_std": 0.04284765986953694, "f1": 0.5643931861867832, "f1_std": 0.04379133700030574, "bacc": 0.5646008403361344, "bacc_std": 0.04322708815632735}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 34, "C": 0.3593813663804626, "split": "test", "acc": 0.5403225806451613, "acc_std": 0.0432549582729253, "f1": 0.537888198757764, "f1_std": 0.04354100764742364, "bacc": 0.5383403361344539, "bacc_std": 0.04366090718453375}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 35, "C": 0.046415888336127774, "split": "test", "acc": 0.6693548387096774, "acc_std": 0.04091259941439645, "f1": 0.6667322189446083, "f1_std": 0.04139352722244033, "bacc": 0.6670168067226891, "bacc_std": 0.04130047995845927}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 36, "C": 0.005994842503189409, "split": "test", "acc": 0.6774193548387096, "acc_std": 0.041091241335156826, "f1": 0.6688034188034189, "f1_std": 0.04277763648525594, "bacc": 0.6680672268907563, "bacc_std": 0.04185253933161877}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 37, "C": 0.046415888336127774, "split": "test", "acc": 0.6532258064516129, "acc_std": 0.0397184550451784, "f1": 0.6448884448884449, "f1_std": 0.04115653437550358, "bacc": 0.6444327731092437, "bacc_std": 0.040399868390830165}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 38, "C": 0.005994842503189409, "split": "test", "acc": 0.6532258064516129, "acc_std": 0.04019436389169068, "f1": 0.6448884448884449, "f1_std": 0.04151157851580024, "bacc": 0.6444327731092437, "bacc_std": 0.040804104287870346}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 39, "C": 0.046415888336127774, "split": "test", "acc": 0.5967741935483871, "acc_std": 0.04396794310417141, "f1": 0.5836690840719849, "f1_std": 0.045798049335887776, "bacc": 0.5850840336134454, "bacc_std": 0.04444940641644906}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 40, "C": 0.046415888336127774, "split": "test", "acc": 0.6532258064516129, "acc_std": 0.03919338576894447, "f1": 0.6331613347093223, "f1_std": 0.042771518458555646, "bacc": 0.6365546218487395, "bacc_std": 0.04010720712249162}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 41, "C": 0.000774263682681127, "split": "test", "acc": 0.6290322580645161, "acc_std": 0.03924210121545952, "f1": 0.6119727891156463, "f1_std": 0.04231364298836883, "bacc": 0.6144957983193278, "bacc_std": 0.04015843813684551}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 42, "C": 0.046415888336127774, "split": "test", "acc": 0.6290322580645161, "acc_std": 0.04409677419354838, "f1": 0.6255252100840336, "f1_std": 0.044836666351665194, "bacc": 0.6255252100840336, "bacc_std": 0.044590379240014404}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 43, "C": 0.046415888336127774, "split": "test", "acc": 0.717741935483871, "acc_std": 0.042134993773728924, "f1": 0.710955710955711, "f1_std": 0.04364201444051302, "bacc": 0.7095588235294117, "bacc_std": 0.04301362189664417}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 44, "C": 0.000774263682681127, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.03997531431299166, "f1": 0.5921052631578947, "f1_std": 0.04397497149481937, "bacc": 0.5966386554621849, "bacc_std": 0.04103224726333098}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 45, "C": 0.046415888336127774, "split": "test", "acc": 0.5806451612903226, "acc_std": 0.041908800443708635, "f1": 0.5643243243243243, "f1_std": 0.043704722143080695, "bacc": 0.5672268907563025, "bacc_std": 0.04222818336737672}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 46, "C": 0.046415888336127774, "split": "test", "acc": 0.5725806451612904, "acc_std": 0.04469744223500621, "f1": 0.5691904293674206, "f1_std": 0.045167292227623355, "bacc": 0.569327731092437, "bacc_std": 0.045194090900702234}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 47, "C": 0.046415888336127774, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.04297369909299979, "f1": 0.607905138339921, "f1_std": 0.0437169917842997, "bacc": 0.6076680672268908, "bacc_std": 0.0434851656123859}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 48, "C": 0.046415888336127774, "split": "test", "acc": 0.6451612903225806, "acc_std": 0.04267607524765116, "f1": 0.6391534391534391, "f1_std": 0.04381060237156751, "bacc": 0.6386554621848739, "bacc_std": 0.04334719713144787}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 49, "C": 0.000774263682681127, "split": "test", "acc": 0.6048387096774194, "acc_std": 0.039736266353089376, "f1": 0.578494623655914, "f1_std": 0.04418238570013005, "bacc": 0.5861344537815126, "bacc_std": 0.04059818158704357}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 50, "C": 0.046415888336127774, "split": "test", "acc": 0.5967741935483871, "acc_std": 0.04465999427283362, "f1": 0.5810810810810811, "f1_std": 0.04713720256675513, "bacc": 0.5835084033613446, "bacc_std": 0.04528510431816933}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 51, "C": 0.046415888336127774, "split": "test", "acc": 0.5967741935483871, "acc_std": 0.041211631163837505, "f1": 0.5810810810810811, "f1_std": 0.04356186821528533, "bacc": 0.5835084033613446, "bacc_std": 0.04182421273594904}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 52, "C": 0.005994842503189409, "split": "test", "acc": 0.6290322580645161, "acc_std": 0.04385020856099628, "f1": 0.6242424242424243, "f1_std": 0.04462218152686849, "bacc": 0.6239495798319328, "bacc_std": 0.04452530415289145}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 53, "C": 0.046415888336127774, "split": "test", "acc": 0.6290322580645161, "acc_std": 0.04276284029254295, "f1": 0.6242424242424243, "f1_std": 0.04342207004187337, "bacc": 0.6239495798319328, "bacc_std": 0.043176670429289805}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 54, "C": 0.005994842503189409, "split": "test", "acc": 0.6209677419354839, "acc_std": 0.04510232037057044, "f1": 0.6137071651090342, "f1_std": 0.0463636833031733, "bacc": 0.6134453781512605, "bacc_std": 0.04571655086801859}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 55, "C": 0.005994842503189409, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.042043706509107076, "f1": 0.6003223207091055, "f1_std": 0.04395201151796931, "bacc": 0.6013655462184874, "bacc_std": 0.04262738162186557}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 56, "C": 0.005994842503189409, "split": "test", "acc": 0.6612903225806451, "acc_std": 0.03989701558058263, "f1": 0.6502820306204673, "f1_std": 0.04235785137587265, "bacc": 0.6502100840336134, "bacc_std": 0.04101881706001485}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 57, "C": 0.046415888336127774, "split": "test", "acc": 0.6532258064516129, "acc_std": 0.039863512587277675, "f1": 0.6480760345851759, "f1_std": 0.0406830702303827, "bacc": 0.6475840336134454, "bacc_std": 0.04038080695452615}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 58, "C": 0.046415888336127774, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.04346805912865296, "f1": 0.6112852664576802, "f1_std": 0.04353824942154756, "bacc": 0.6123949579831933, "bacc_std": 0.04371216285030784}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 59, "C": 0.046415888336127774, "split": "test", "acc": 0.6290322580645161, "acc_std": 0.04456227155100645, "f1": 0.6266038229903116, "f1_std": 0.0448244940169691, "bacc": 0.6271008403361344, "bacc_std": 0.044791266846952324}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 60, "C": 0.046415888336127774, "split": "test", "acc": 0.5887096774193549, "acc_std": 0.04398425541310865, "f1": 0.5865315462569467, "f1_std": 0.04417646002498987, "bacc": 0.5871848739495797, "bacc_std": 0.04429141821564067}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 61, "C": 0.046415888336127774, "split": "test", "acc": 0.5725806451612904, "acc_std": 0.04423782512393834, "f1": 0.5623043623043623, "f1_std": 0.04515594725345051, "bacc": 0.5630252100840336, "bacc_std": 0.04431821774533121}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 62, "C": 0.000774263682681127, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.04186993636584508, "f1": 0.5978378378378378, "f1_std": 0.04441792401715795, "bacc": 0.5997899159663866, "bacc_std": 0.04274082997220342}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 63, "C": 0.046415888336127774, "split": "test", "acc": 0.6209677419354839, "acc_std": 0.044036972452337696, "f1": 0.6203504657677024, "f1_std": 0.044150532010916364, "bacc": 0.6228991596638656, "bacc_std": 0.04416149355513142}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 64, "C": 0.046415888336127774, "split": "test", "acc": 0.6451612903225806, "acc_std": 0.04257984786643588, "f1": 0.6391534391534391, "f1_std": 0.04374301919437287, "bacc": 0.6386554621848739, "bacc_std": 0.0432648698667387}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 65, "C": 0.046415888336127774, "split": "test", "acc": 0.6048387096774194, "acc_std": 0.044480467249134785, "f1": 0.5931704050887178, "f1_std": 0.04665746390703013, "bacc": 0.5940126050420168, "bacc_std": 0.04535424956033638}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 66, "C": 0.046415888336127774, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.04579966726224051, "f1": 0.6092436974789917, "f1_std": 0.046368699932556025, "bacc": 0.6092436974789917, "bacc_std": 0.04625389298721849}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 67, "C": 0.046415888336127774, "split": "test", "acc": 0.6532258064516129, "acc_std": 0.04194602845096986, "f1": 0.6448884448884449, "f1_std": 0.043576932006045295, "bacc": 0.6444327731092437, "bacc_std": 0.04268664970847627}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 68, "C": 0.000774263682681127, "split": "test", "acc": 0.5967741935483871, "acc_std": 0.03919394331408622, "f1": 0.5810810810810811, "f1_std": 0.04189347456698179, "bacc": 0.5835084033613446, "bacc_std": 0.039982221811747945}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 69, "C": 0.005994842503189409, "split": "test", "acc": 0.6854838709677419, "acc_std": 0.0407685020527538, "f1": 0.6721578198088265, "f1_std": 0.043651566533400986, "bacc": 0.6722689075630253, "bacc_std": 0.041829512970162006}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 70, "C": 0.046415888336127774, "split": "test", "acc": 0.6209677419354839, "acc_std": 0.04453546506193729, "f1": 0.6153389215233318, "f1_std": 0.045388848751803205, "bacc": 0.6150210084033614, "bacc_std": 0.04509356542582331}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 71, "C": 0.046415888336127774, "split": "test", "acc": 0.6532258064516129, "acc_std": 0.03931641103787403, "f1": 0.6429862738533645, "f1_std": 0.04107719866164188, "bacc": 0.6428571428571428, "bacc_std": 0.0399598841213838}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 72, "C": 0.3593813663804626, "split": "test", "acc": 0.5241935483870968, "acc_std": 0.04360950012968361, "f1": 0.5234186697934988, "f1_std": 0.043613506635374236, "bacc": 0.5252100840336134, "bacc_std": 0.043709194454083354}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 73, "C": 0.046415888336127774, "split": "test", "acc": 0.6290322580645161, "acc_std": 0.03969965283177649, "f1": 0.6266038229903116, "f1_std": 0.039870109268220065, "bacc": 0.6271008403361344, "bacc_std": 0.0398788773951557}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 74, "C": 0.005994842503189409, "split": "test", "acc": 0.6290322580645161, "acc_std": 0.04281820340911307, "f1": 0.6210470369386127, "f1_std": 0.04412702440007403, "bacc": 0.6207983193277311, "bacc_std": 0.04343186865320686}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 75, "C": 0.046415888336127774, "split": "test", "acc": 0.6451612903225806, "acc_std": 0.0409200891258643, "f1": 0.6356837606837606, "f1_std": 0.04256190347344885, "bacc": 0.6355042016806722, "bacc_std": 0.04166504903969842}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 76, "C": 0.046415888336127774, "split": "test", "acc": 0.6693548387096774, "acc_std": 0.04075176751591489, "f1": 0.6575739206573719, "f1_std": 0.04300275716343429, "bacc": 0.657563025210084, "bacc_std": 0.041464770753086654}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 77, "C": 0.046415888336127774, "split": "test", "acc": 0.532258064516129, "acc_std": 0.04507058575514828, "f1": 0.5311603650586701, "f1_std": 0.045220500564835, "bacc": 0.532563025210084, "bacc_std": 0.04539670070997022}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 78, "C": 0.046415888336127774, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.042833860369720776, "f1": 0.5978378378378378, "f1_std": 0.044704063706268816, "bacc": 0.5997899159663866, "bacc_std": 0.04322538013830674}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 79, "C": 0.046415888336127774, "split": "test", "acc": 0.6048387096774194, "acc_std": 0.043808123275625055, "f1": 0.5931704050887178, "f1_std": 0.04613649124232754, "bacc": 0.5940126050420168, "bacc_std": 0.04466083182381372}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 80, "C": 0.3593813663804626, "split": "test", "acc": 0.5967741935483871, "acc_std": 0.04270167001732451, "f1": 0.5941345902068604, "f1_std": 0.042936531142214185, "bacc": 0.5945378151260504, "bacc_std": 0.0431073123937748}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 81, "C": 0.046415888336127774, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.04402541598737713, "f1": 0.5951020408163266, "f1_std": 0.047541045925511, "bacc": 0.5982142857142857, "bacc_std": 0.04497705880162714}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 82, "C": 0.005994842503189409, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.04179572243844966, "f1": 0.6003223207091055, "f1_std": 0.04422265172952331, "bacc": 0.6013655462184874, "bacc_std": 0.04276559544832569}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 83, "C": 0.046415888336127774, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.04304336801014306, "f1": 0.6092436974789917, "f1_std": 0.043478191221444244, "bacc": 0.6092436974789917, "bacc_std": 0.04344047080429191}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 84, "C": 0.3593813663804626, "split": "test", "acc": 0.6048387096774194, "acc_std": 0.04528888656854025, "f1": 0.6004471624909581, "f1_std": 0.04571977995275497, "bacc": 0.6003151260504203, "bacc_std": 0.04551317451707635}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 85, "C": 0.046415888336127774, "split": "test", "acc": 0.6290322580645161, "acc_std": 0.04127939308473816, "f1": 0.6210470369386127, "f1_std": 0.04241401945253636, "bacc": 0.6207983193277311, "bacc_std": 0.04176528047883907}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 86, "C": 0.046415888336127774, "split": "test", "acc": 0.6290322580645161, "acc_std": 0.04500658698466618, "f1": 0.6266038229903116, "f1_std": 0.045471849995903604, "bacc": 0.6271008403361344, "bacc_std": 0.04546393078776404}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 87, "C": 0.046415888336127774, "split": "test", "acc": 0.6612903225806451, "acc_std": 0.04085137142552884, "f1": 0.6569169960474308, "f1_std": 0.04127147306952389, "bacc": 0.6565126050420168, "bacc_std": 0.04102570972715738}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 88, "C": 0.046415888336127774, "split": "test", "acc": 0.5887096774193549, "acc_std": 0.04596567907086079, "f1": 0.5873947935016637, "f1_std": 0.04615058046713721, "bacc": 0.5887605042016807, "bacc_std": 0.04619853584836723}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 89, "C": 0.046415888336127774, "split": "test", "acc": 0.5967741935483871, "acc_std": 0.04131666246557779, "f1": 0.5963541666666667, "f1_std": 0.041389998778551144, "bacc": 0.5992647058823529, "bacc_std": 0.04156590435483708}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 90, "C": 0.046415888336127774, "split": "test", "acc": 0.6451612903225806, "acc_std": 0.041285962471406985, "f1": 0.6313513513513513, "f1_std": 0.04354651373630016, "bacc": 0.6323529411764706, "bacc_std": 0.04198956385720837}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 91, "C": 0.046415888336127774, "split": "test", "acc": 0.6129032258064516, "acc_std": 0.04043778381713843, "f1": 0.6063492063492064, "f1_std": 0.041450922687933905, "bacc": 0.60609243697479, "bacc_std": 0.04103487482293122}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 92, "C": 0.046415888336127774, "split": "test", "acc": 0.5483870967741935, "acc_std": 0.04340927601949774, "f1": 0.5441176470588236, "f1_std": 0.04347175593474291, "bacc": 0.5441176470588236, "bacc_std": 0.04334445608826659}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 93, "C": 0.046415888336127774, "split": "test", "acc": 0.5564516129032258, "acc_std": 0.043894855332508066, "f1": 0.5529334644378892, "f1_std": 0.04385612774890223, "bacc": 0.553046218487395, "bacc_std": 0.043792677830717164}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 94, "C": 0.000774263682681127, "split": "test", "acc": 0.5967741935483871, "acc_std": 0.042042280264808704, "f1": 0.5836690840719849, "f1_std": 0.044153889673427024, "bacc": 0.5850840336134454, "bacc_std": 0.04273000063675474}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 95, "C": 0.046415888336127774, "split": "test", "acc": 0.6048387096774194, "acc_std": 0.043058326847732004, "f1": 0.5880957223239103, "f1_std": 0.04592739007657337, "bacc": 0.5908613445378151, "bacc_std": 0.04388065219861184}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 96, "C": 0.046415888336127774, "split": "test", "acc": 0.6693548387096774, "acc_std": 0.042988549040357406, "f1": 0.6630211440312852, "f1_std": 0.04456023013336834, "bacc": 0.6622899159663866, "bacc_std": 0.04383846685208054}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 97, "C": 0.3593813663804626, "split": "test", "acc": 0.6370967741935484, "acc_std": 0.04334339746787591, "f1": 0.6330637206549615, "f1_std": 0.04419768212279903, "bacc": 0.6328781512605042, "bacc_std": 0.04400716331604586}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 98, "C": 0.046415888336127774, "split": "test", "acc": 0.6693548387096774, "acc_std": 0.04248585490260742, "f1": 0.6595915634415801, "f1_std": 0.04456710340963384, "bacc": 0.6591386554621849, "bacc_std": 0.04342151480565184}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 99, "C": 0.046415888336127774, "split": "test", "acc": 0.6693548387096774, "acc_std": 0.03895119214359001, "f1": 0.6575739206573719, "f1_std": 0.041270982845458284, "bacc": 0.657563025210084, "bacc_std": 0.039917357671463004}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "abide_dx", "trial": 100, "C": 0.000774263682681127, "split": "test", "acc": 0.6370967741935484, "acc_std": 0.04103809998482146, "f1": 0.6190346145968457, "f1_std": 0.04485951415033147, "bacc": 0.6218487394957983, "bacc_std": 0.042330725111534864}
+eval results (random splits):
+
+| model    | repr   | clf      | dataset   | split   |   n_trials |      C |   C_std |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:----------|:--------|-----------:|-------:|--------:|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | abide_dx  | train   |        100 | 12.989 |  129.15 | 0.80406 |  0.071932 | 0.79974 | 0.075084 | 0.79846 |   0.074815 |
+| flat_mae | patch  | logistic | abide_dx  | test    |        100 | 12.989 |  129.15 | 0.61944 |  0.035969 | 0.61111 | 0.036134 | 0.61183 |   0.035412 |
+
+
+done! total time: 0:05:23

data_scaling/n400_1/eval_v2/adhd200_dx__patch__logistic/config.yaml

@@ -0,0 +1,30 @@
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (adhd200_dx patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_1/eval_v2/adhd200_dx__patch__logistic
+model: flat_mae
+representation: patch
+dataset: adhd200_dx
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/adhd200_dx__patch__logistic
+remote_dir: null

data_scaling/n400_1/eval_v2/adhd200_dx__patch__logistic/eval_table.csv

@@ -0,0 +1,203 @@
+model,repr,clf,dataset,trial,C,split,acc,acc_std,f1,f1_std,bacc,bacc_std
+flat_mae,patch,logistic,adhd200_dx,,0.046415888336127774,train,0.8657534246575342,0.017078188059986704,0.8614950940532335,0.01786542857265286,0.857391463638029,0.018112886528454195
+flat_mae,patch,logistic,adhd200_dx,,0.046415888336127774,test,0.676923076923077,0.059504947832663864,0.6719538572458543,0.06047265227632608,0.6727799227799228,0.06038152931158118
+flat_mae,patch,logistic,adhd200_dx,1,0.000774263682681127,train,0.6767123287671233,0.022587077014058984,0.6510403163080114,0.025500796987109806,0.6526073151370825,0.02357514448176547
+flat_mae,patch,logistic,adhd200_dx,1,0.000774263682681127,test,0.5538461538461539,0.0572655890767544,0.5250692869740489,0.06076585750307687,0.5299227799227799,0.05785183843593267
+flat_mae,patch,logistic,adhd200_dx,2,0.046415888336127774,train,0.863013698630137,0.018246061122376967,0.8585271317829457,0.01910003251673321,0.8542468095499787,0.019267217692187515
+flat_mae,patch,logistic,adhd200_dx,2,0.046415888336127774,test,0.5692307692307692,0.05671171570517729,0.5376016260162602,0.06225838701724851,0.5434362934362934,0.0579687388284011
+flat_mae,patch,logistic,adhd200_dx,3,0.005994842503189409,train,0.7753424657534247,0.022369737351692617,0.7648787078934138,0.024009055738663534,0.7607925749526775,0.023508700049786866
+flat_mae,patch,logistic,adhd200_dx,3,0.005994842503189409,test,0.6,0.05724673888027191,0.5775,0.06147789908968846,0.5791505791505791,0.058772705740167124
+flat_mae,patch,logistic,adhd200_dx,4,0.005994842503189409,train,0.7506849315068493,0.023079824113354856,0.7411650107149814,0.02446588668211456,0.7382304451364718,0.024122255251734163
+flat_mae,patch,logistic,adhd200_dx,4,0.005994842503189409,test,0.6153846153846154,0.058280724431296206,0.6018132810585641,0.061335942045065636,0.6013513513513513,0.059766919147401495
+flat_mae,patch,logistic,adhd200_dx,5,0.005994842503189409,train,0.7808219178082192,0.02078671573710714,0.7736434108527132,0.022026284546400397,0.7706692312389326,0.021989550001872284
+flat_mae,patch,logistic,adhd200_dx,5,0.005994842503189409,test,0.5076923076923077,0.06166973690315324,0.5019157088122606,0.06265548856678507,0.5024131274131274,0.06307831677995987
+flat_mae,patch,logistic,adhd200_dx,6,0.005994842503189409,train,0.7753424657534247,0.020005385651299848,0.7643010143010143,0.021738756019797873,0.7600751053306467,0.021251549075675173
+flat_mae,patch,logistic,adhd200_dx,6,0.005994842503189409,test,0.676923076923077,0.05564570085522007,0.6612062546537603,0.06027805632457064,0.6597490347490347,0.058029226425179735
+flat_mae,patch,logistic,adhd200_dx,7,0.000774263682681127,train,0.6904109589041096,0.02384004430609179,0.6739551465996316,0.026013751162567152,0.6726354033095194,0.02499823872913839
+flat_mae,patch,logistic,adhd200_dx,7,0.000774263682681127,test,0.6,0.053583809265342605,0.570630081300813,0.05875139503905504,0.5748069498069498,0.05497776935868238
+flat_mae,patch,logistic,adhd200_dx,8,0.000774263682681127,train,0.6575342465753424,0.022877677610674274,0.6332071163848894,0.02618067205423451,0.6348995542529157,0.024211013993169033
+flat_mae,patch,logistic,adhd200_dx,8,0.000774263682681127,test,0.6307692307692307,0.052420466665642984,0.577922077922078,0.06534726334508818,0.5931467181467182,0.05533943784487845
+flat_mae,patch,logistic,adhd200_dx,9,0.046415888336127774,train,0.863013698630137,0.017203730824674596,0.8590777118853472,0.0178951782948399,0.8556817487940405,0.018128698401832573
+flat_mae,patch,logistic,adhd200_dx,9,0.046415888336127774,test,0.6,0.06166942986506083,0.588206627680312,0.06414252962711496,0.5878378378378378,0.0631285088032032
+flat_mae,patch,logistic,adhd200_dx,10,0.3593813663804626,train,0.9808219178082191,0.007018189914722619,0.9804551539491299,0.007165759005496705,0.9794223606277096,0.0074775484731877655
+flat_mae,patch,logistic,adhd200_dx,10,0.3593813663804626,test,0.5384615384615384,0.059654952625754924,0.5192307692307693,0.06180309132185452,0.5207528957528957,0.060215730060988015
+flat_mae,patch,logistic,adhd200_dx,11,0.046415888336127774,train,0.873972602739726,0.015797057857953677,0.8703514949345195,0.016402000539652333,0.8668254259021799,0.016596147741691346
+flat_mae,patch,logistic,adhd200_dx,11,0.046415888336127774,test,0.5692307692307692,0.05821271745394455,0.5608108108108107,0.05943390770577114,0.5608108108108107,0.05902766828279475
+flat_mae,patch,logistic,adhd200_dx,12,0.3593813663804626,train,0.9863013698630136,0.0058550169507943515,0.9860175757157852,0.006003312751636345,0.9842767295597484,0.006720381091320536
+flat_mae,patch,logistic,adhd200_dx,12,0.3593813663804626,test,0.5846153846153846,0.05862002808177161,0.5830363506771205,0.058889472645809154,0.5873552123552124,0.05932128623455292
+flat_mae,patch,logistic,adhd200_dx,13,0.046415888336127774,train,0.8904109589041096,0.01657785557762633,0.8874707115550623,0.017187596993291695,0.88425841118642,0.017500511778493447
+flat_mae,patch,logistic,adhd200_dx,13,0.046415888336127774,test,0.6,0.06020512471061013,0.5921814671814671,0.06129344119325076,0.5921814671814671,0.060771826774628876
+flat_mae,patch,logistic,adhd200_dx,14,0.046415888336127774,train,0.8493150684931506,0.01846539489916572,0.8454116324377604,0.019064076766956806,0.8428283568419125,0.01917790528908459
+flat_mae,patch,logistic,adhd200_dx,14,0.046415888336127774,test,0.6461538461538462,0.05890901619743036,0.6336682185738789,0.062146457074397456,0.6327220077220077,0.060796763063037065
+flat_mae,patch,logistic,adhd200_dx,15,0.005994842503189409,train,0.7452054794520548,0.021742253092199054,0.7316622002988118,0.023653793230580165,0.7283537888502167,0.022795503203861693
+flat_mae,patch,logistic,adhd200_dx,15,0.005994842503189409,test,0.6307692307692307,0.05951964719625667,0.6235521235521235,0.06047272103557194,0.6235521235521235,0.060358604493063656
+flat_mae,patch,logistic,adhd200_dx,16,0.005994842503189409,train,0.7698630136986301,0.02056704208476135,0.7602739726027397,0.022036332541953082,0.7566556756426696,0.021706631290715647
+flat_mae,patch,logistic,adhd200_dx,16,0.005994842503189409,test,0.5538461538461539,0.06091686059971773,0.5381034060279344,0.06258601657950763,0.5386100386100386,0.06138737497131656
+flat_mae,patch,logistic,adhd200_dx,17,0.046415888336127774,train,0.8684931506849315,0.01760480580593876,0.8652058841632302,0.01813228314639139,0.862688526592172,0.01819513376921124
+flat_mae,patch,logistic,adhd200_dx,17,0.046415888336127774,test,0.5846153846153846,0.05871990680869055,0.5699583435432491,0.06145209366131349,0.5699806949806949,0.060034387646530685
+flat_mae,patch,logistic,adhd200_dx,18,0.046415888336127774,train,0.8602739726027397,0.017702704480348722,0.8558418325860186,0.018563732784502077,0.8518196250839593,0.018808476561670995
+flat_mae,patch,logistic,adhd200_dx,18,0.046415888336127774,test,0.6153846153846154,0.059894629368741695,0.606060606060606,0.06123124537723429,0.6056949806949807,0.060530000381377465
+flat_mae,patch,logistic,adhd200_dx,19,0.046415888336127774,train,0.8712328767123287,0.01752891699825019,0.8666029035544047,0.018552076117333956,0.8615283629480369,0.018863359931584694
+flat_mae,patch,logistic,adhd200_dx,19,0.046415888336127774,test,0.676923076923077,0.06036583539420486,0.6655231560891939,0.06391764659581159,0.6640926640926641,0.06276648339768712
+flat_mae,patch,logistic,adhd200_dx,20,0.005994842503189409,train,0.7424657534246575,0.02145089729027852,0.7329212853406402,0.02267758029029618,0.7302314221163827,0.022317700473933284
+flat_mae,patch,logistic,adhd200_dx,20,0.005994842503189409,test,0.6153846153846154,0.05593973934395675,0.5966741126830479,0.05956773634875029,0.597007722007722,0.05736341308013188
+flat_mae,patch,logistic,adhd200_dx,21,0.005994842503189409,train,0.7671232876712328,0.0213903512603398,0.7571409560935898,0.022703998105863116,0.7535110215546192,0.02227510118001808
+flat_mae,patch,logistic,adhd200_dx,21,0.005994842503189409,test,0.7076923076923077,0.05511287180916318,0.6934723256391164,0.05935428871932758,0.6911196911196911,0.05751842221042249
+flat_mae,patch,logistic,adhd200_dx,22,0.046415888336127774,train,0.8410958904109589,0.01852075544991967,0.8362165005879805,0.01938356587541209,0.8326769249557306,0.019487613328053925
+flat_mae,patch,logistic,adhd200_dx,22,0.046415888336127774,test,0.6153846153846154,0.05756622119338048,0.6094688776736361,0.059000555408504046,0.61003861003861,0.05916352132013179
+flat_mae,patch,logistic,adhd200_dx,23,0.005994842503189409,train,0.7917808219178082,0.01943952905610274,0.7831050228310502,0.020821796423450743,0.7789430298589486,0.020593756305295988
+flat_mae,patch,logistic,adhd200_dx,23,0.005994842503189409,test,0.6153846153846154,0.05582575596967586,0.5905769715293525,0.06092880156668017,0.5926640926640927,0.05723420917198336
+flat_mae,patch,logistic,adhd200_dx,24,0.005994842503189409,train,0.7452054794520548,0.02150419039492512,0.7336581691500129,0.023225537391161698,0.7305061977163094,0.022753184600145818
+flat_mae,patch,logistic,adhd200_dx,24,0.005994842503189409,test,0.6615384615384615,0.05450640481338385,0.6299171842650104,0.06302193643219771,0.6332046332046332,0.057101951438083956
+flat_mae,patch,logistic,adhd200_dx,25,0.005994842503189409,train,0.7616438356164383,0.020492957291593516,0.7476336933457312,0.022551349053612554,0.7436343652683641,0.021706554413297257
+flat_mae,patch,logistic,adhd200_dx,25,0.005994842503189409,test,0.6461538461538462,0.05497279370140541,0.6167649320687003,0.0619366915297259,0.6196911196911197,0.05722116357302214
+flat_mae,patch,logistic,adhd200_dx,26,0.005994842503189409,train,0.7561643835616438,0.021106356414677466,0.7451137317672167,0.02276757266686839,0.7416498748244489,0.02223659459559254
+flat_mae,patch,logistic,adhd200_dx,26,0.005994842503189409,test,0.6153846153846154,0.0551576546265822,0.5834401435529352,0.06222831061513817,0.5883204633204633,0.05713123455936296
+flat_mae,patch,logistic,adhd200_dx,27,0.005994842503189409,train,0.7589041095890411,0.021694467568080304,0.7488584474885844,0.023092882223134936,0.7455119985345301,0.022705384174726586
+flat_mae,patch,logistic,adhd200_dx,27,0.005994842503189409,test,0.7230769230769231,0.05303444454024486,0.7176640926640927,0.05405444556692365,0.7176640926640927,0.05377801209791008
+flat_mae,patch,logistic,adhd200_dx,28,0.005994842503189409,train,0.7671232876712328,0.020626755894595308,0.7576948008840918,0.02190316980927576,0.7542284911766501,0.021540367710327007
+flat_mae,patch,logistic,adhd200_dx,28,0.005994842503189409,test,0.6153846153846154,0.0541713258502937,0.5834401435529352,0.06113224527527147,0.5883204633204633,0.05613889901634252
+flat_mae,patch,logistic,adhd200_dx,29,0.005994842503189409,train,0.7698630136986301,0.0215035509206162,0.7608125819134994,0.022827928711384598,0.7573731452647005,0.022512442760103606
+flat_mae,patch,logistic,adhd200_dx,29,0.005994842503189409,test,0.5692307692307692,0.06031255082470189,0.545,0.06448818718183073,0.5477799227799228,0.06153564747621803
+flat_mae,patch,logistic,adhd200_dx,30,0.046415888336127774,train,0.873972602739726,0.018006069742447978,0.8705913182883216,0.018731876710168793,0.8675428955242108,0.019059334448234537
+flat_mae,patch,logistic,adhd200_dx,30,0.046415888336127774,test,0.6,0.054693119111150615,0.5833333333333333,0.0571336811586876,0.5834942084942085,0.0555374175286818
+flat_mae,patch,logistic,adhd200_dx,31,0.3593813663804626,train,0.9753424657534246,0.0083709171364992,0.9749085326265458,0.008526210055046533,0.9745679916956708,0.00867276297049639
+flat_mae,patch,logistic,adhd200_dx,31,0.3593813663804626,test,0.6307692307692307,0.059494446063696034,0.6235521235521235,0.060752597031315435,0.6235521235521235,0.06019783739826568
+flat_mae,patch,logistic,adhd200_dx,32,0.046415888336127774,train,0.8493150684931506,0.018935942227686766,0.8454116324377604,0.0195802287188209,0.8428283568419125,0.019716087889845704
+flat_mae,patch,logistic,adhd200_dx,32,0.046415888336127774,test,0.6461538461538462,0.05741953155112201,0.6289401836684041,0.06168051084712195,0.6283783783783784,0.05967790280427032
+flat_mae,patch,logistic,adhd200_dx,33,0.005994842503189409,train,0.7671232876712328,0.02182961709234036,0.7587499319600937,0.023022149306424974,0.7556634304207119,0.022789463321737636
+flat_mae,patch,logistic,adhd200_dx,33,0.005994842503189409,test,0.5692307692307692,0.06055749676300353,0.564176245210728,0.06119702477144015,0.5651544401544402,0.061196796584670324
+flat_mae,patch,logistic,adhd200_dx,34,0.046415888336127774,train,0.873972602739726,0.01717838456398579,0.8714631308182264,0.017623491710827794,0.8704127740123344,0.01788683430789088
+flat_mae,patch,logistic,adhd200_dx,34,0.046415888336127774,test,0.5692307692307692,0.06031380659742999,0.564176245210728,0.06124768065242325,0.5651544401544402,0.06163327286376625
+flat_mae,patch,logistic,adhd200_dx,35,0.005994842503189409,train,0.7342465753424657,0.021562612291352177,0.7178457296323746,0.023684312368223034,0.7150577028759846,0.022611843677091154
+flat_mae,patch,logistic,adhd200_dx,35,0.005994842503189409,test,0.6307692307692307,0.057498826332246115,0.6036585365853658,0.06332847127867035,0.6061776061776062,0.05911397772664288
+flat_mae,patch,logistic,adhd200_dx,36,0.005994842503189409,train,0.7616438356164383,0.02291267782318527,0.7540831261761494,0.02397005393991643,0.751526531110704,0.023827791822369087
+flat_mae,patch,logistic,adhd200_dx,36,0.005994842503189409,test,0.6153846153846154,0.05808497346140251,0.6018132810585641,0.061038667690533496,0.6013513513513513,0.05966645370475177
+flat_mae,patch,logistic,adhd200_dx,37,0.000774263682681127,train,0.6575342465753424,0.02096302296353183,0.6219371732087072,0.024623448237372194,0.6284423276546376,0.02194461059330252
+flat_mae,patch,logistic,adhd200_dx,37,0.000774263682681127,test,0.6153846153846154,0.04606058269395328,0.554672513017265,0.05930753190875318,0.5752895752895753,0.04843900617142849
+flat_mae,patch,logistic,adhd200_dx,38,0.005994842503189409,train,0.7342465753424657,0.023001444225953283,0.7240989674654198,0.02441805742465556,0.7215149294742627,0.023990316266928532
+flat_mae,patch,logistic,adhd200_dx,38,0.005994842503189409,test,0.5692307692307692,0.05746423842899906,0.5512820512820513,0.06012947737164356,0.5521235521235521,0.05820206915826109
+flat_mae,patch,logistic,adhd200_dx,39,0.005994842503189409,train,0.7726027397260274,0.02038158482091978,0.7628552865384465,0.021818746316566745,0.759082860108689,0.021491237705092245
+flat_mae,patch,logistic,adhd200_dx,39,0.005994842503189409,test,0.6307692307692307,0.0615423075721229,0.6153846153846154,0.06404714127278875,0.6148648648648649,0.06243158787689593
+flat_mae,patch,logistic,adhd200_dx,40,0.005994842503189409,train,0.7561643835616438,0.022598059457321323,0.7468536917981687,0.023885210398192535,0.7438022836905416,0.02352146265362123
+flat_mae,patch,logistic,adhd200_dx,40,0.005994842503189409,test,0.6615384615384615,0.04774166180799899,0.6130952380952381,0.06002177730440167,0.6245173745173745,0.05071997773002365
+flat_mae,patch,logistic,adhd200_dx,41,0.046415888336127774,train,0.863013698630137,0.01789663530565597,0.8590777118853472,0.018640560725927097,0.8556817487940405,0.018831121553825235
+flat_mae,patch,logistic,adhd200_dx,41,0.046415888336127774,test,0.6461538461538462,0.05978969850515453,0.644808743169399,0.05982158801541909,0.6500965250965252,0.059381385241960516
+flat_mae,patch,logistic,adhd200_dx,42,0.005994842503189409,train,0.7917808219178082,0.020232828298696496,0.7835923360169756,0.02123527764317484,0.7796604994809794,0.020959728647286076
+flat_mae,patch,logistic,adhd200_dx,42,0.005994842503189409,test,0.6307692307692307,0.06271468972810364,0.6235521235521235,0.06377761357442285,0.6235521235521235,0.06339578516973475
+flat_mae,patch,logistic,adhd200_dx,43,0.000774263682681127,train,0.6657534246575343,0.022568342444463242,0.6403534276668605,0.025380538033943204,0.6421811076509739,0.023502473998959248
+flat_mae,patch,logistic,adhd200_dx,43,0.000774263682681127,test,0.5846153846153846,0.060512243369694504,0.5501153550371699,0.06653549857759201,0.556949806949807,0.061573446806877014
+flat_mae,patch,logistic,adhd200_dx,44,0.046415888336127774,train,0.8712328767123287,0.01711640058759064,0.8683513801809545,0.017647832420832163,0.8665506503022532,0.017941781861312157
+flat_mae,patch,logistic,adhd200_dx,44,0.046415888336127774,test,0.6153846153846154,0.05764464903293434,0.606060606060606,0.05921578486629486,0.6056949806949807,0.05872212325340343
+flat_mae,patch,logistic,adhd200_dx,45,0.005994842503189409,train,0.7780821917808219,0.022036380639370943,0.7680248570016242,0.02381929600954055,0.7639372290407278,0.02338048202951626
+flat_mae,patch,logistic,adhd200_dx,45,0.005994842503189409,test,0.5076923076923077,0.05950008701699171,0.4871794871794872,0.06236568104858205,0.48938223938223935,0.06038077148270777
+flat_mae,patch,logistic,adhd200_dx,46,0.005994842503189409,train,0.7479452054794521,0.021043395993003866,0.736833855799373,0.022497110144429934,0.7336508518043597,0.022001931833103797
+flat_mae,patch,logistic,adhd200_dx,46,0.005994842503189409,test,0.6153846153846154,0.059431141598626364,0.5966741126830479,0.06273333213340375,0.597007722007722,0.06045994029541409
+flat_mae,patch,logistic,adhd200_dx,47,0.005994842503189409,train,0.7780821917808219,0.021339051197830805,0.7685696169833032,0.022857905959625176,0.7646546986627587,0.02248936072595075
+flat_mae,patch,logistic,adhd200_dx,47,0.005994842503189409,test,0.5692307692307692,0.06130707267325128,0.5565302144249512,0.06265016787306586,0.5564671814671815,0.06192038296137542
+flat_mae,patch,logistic,adhd200_dx,48,0.046415888336127774,train,0.8684931506849315,0.01770759771118233,0.8644550349693632,0.018494065637778712,0.8605361177260793,0.01873217203779221
+flat_mae,patch,logistic,adhd200_dx,48,0.046415888336127774,test,0.6307692307692307,0.060914148529912664,0.6264367816091954,0.06175308554170438,0.627895752895753,0.061938043329468624
+flat_mae,patch,logistic,adhd200_dx,49,0.046415888336127774,train,0.863013698630137,0.017342101318424116,0.8588073280930866,0.018117311580660808,0.8549642791720096,0.018291489179414932
+flat_mae,patch,logistic,adhd200_dx,49,0.046415888336127774,test,0.5538461538461539,0.06234080032061976,0.5500119360229172,0.06306192727957428,0.5516409266409266,0.06359814774523685
+flat_mae,patch,logistic,adhd200_dx,50,0.3593813663804626,train,0.9753424657534246,0.008364901787247829,0.9748709122203099,0.00854967290561261,0.9738505220736399,0.008947496248503129
+flat_mae,patch,logistic,adhd200_dx,50,0.3593813663804626,test,0.5692307692307692,0.058078836443265204,0.5565302144249512,0.06045777555896192,0.5564671814671815,0.05953511430094472
+flat_mae,patch,logistic,adhd200_dx,51,0.046415888336127774,train,0.873972602739726,0.017171881513996033,0.8710445468509984,0.01767797475103027,0.8689778347682726,0.017893644739776993
+flat_mae,patch,logistic,adhd200_dx,51,0.046415888336127774,test,0.6,0.06084209069386094,0.5921814671814671,0.06265787776687684,0.5921814671814671,0.06242020102102711
+flat_mae,patch,logistic,adhd200_dx,52,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adhd200_dx,52,166.81005372000556,test,0.5230769230769231,0.06048955311045832,0.5115151515151515,0.06166501453032125,0.5115830115830116,0.061061613368097144
+flat_mae,patch,logistic,adhd200_dx,53,0.046415888336127774,train,0.8712328767123287,0.01686582463246114,0.8671483555204486,0.017702712573942035,0.8629633021920987,0.0179582989295863
+flat_mae,patch,logistic,adhd200_dx,53,0.046415888336127774,test,0.5230769230769231,0.05926792041680342,0.49987589972697943,0.06221983316200884,0.502895752895753,0.05999061139233182
+flat_mae,patch,logistic,adhd200_dx,54,0.000774263682681127,train,0.6794520547945205,0.021622128181495368,0.6596695914122457,0.024302370226241755,0.6593393173352873,0.022912587776226175
+flat_mae,patch,logistic,adhd200_dx,54,0.000774263682681127,test,0.5384615384615384,0.05798408269649392,0.5125,0.06153215536539919,0.5164092664092664,0.05870489862503548
+flat_mae,patch,logistic,adhd200_dx,55,0.046415888336127774,train,0.8931506849315068,0.015940105572410695,0.8905749140972719,0.016409197507246027,0.8881205348965012,0.01666344198910492
+flat_mae,patch,logistic,adhd200_dx,55,0.046415888336127774,test,0.5692307692307692,0.06375048253675135,0.5666666666666667,0.06380417795717837,0.5694980694980695,0.06425679767514877
+flat_mae,patch,logistic,adhd200_dx,56,0.005994842503189409,train,0.7726027397260274,0.021767964500068952,0.7617212386248339,0.023420409153067163,0.7576479208646272,0.02289742657845494
+flat_mae,patch,logistic,adhd200_dx,56,0.005994842503189409,test,0.5846153846153846,0.05939442724840168,0.5745454545454545,0.06070472188012568,0.5743243243243243,0.060147054191353515
+flat_mae,patch,logistic,adhd200_dx,57,0.005994842503189409,train,0.7945205479452054,0.020907293604799145,0.7866744593804793,0.022166814176650108,0.7828051535690297,0.021965267417927148
+flat_mae,patch,logistic,adhd200_dx,57,0.005994842503189409,test,0.6153846153846154,0.059952320305882836,0.5905769715293525,0.06525933957683168,0.5926640926640927,0.061561379891102906
+flat_mae,patch,logistic,adhd200_dx,58,0.005994842503189409,train,0.7863013698630137,0.020597215085571424,0.7797394318252151,0.021579319056509432,0.7769585394150332,0.021568410830982366
+flat_mae,patch,logistic,adhd200_dx,58,0.005994842503189409,test,0.6615384615384615,0.0565710194064628,0.6515594541910331,0.058946245373878714,0.6505791505791505,0.05777064500738074
+flat_mae,patch,logistic,adhd200_dx,59,0.046415888336127774,train,0.8821917808219178,0.017601838898116118,0.8786916162342229,0.018382647533215588,0.8748244489222691,0.018699544783355525
+flat_mae,patch,logistic,adhd200_dx,59,0.046415888336127774,test,0.5076923076923077,0.05877225123128174,0.48,0.062216358119852586,0.48503861003861004,0.05948481241158543
+flat_mae,patch,logistic,adhd200_dx,60,0.005994842503189409,train,0.7972602739726027,0.020075186877502173,0.7888127853881278,0.021429790625384876,0.7845148684130183,0.021151749741149483
+flat_mae,patch,logistic,adhd200_dx,60,0.005994842503189409,test,0.6307692307692307,0.05591638728372842,0.61,0.06102715186358429,0.6105212355212355,0.05797635992295799
+flat_mae,patch,logistic,adhd200_dx,61,0.005994842503189409,train,0.7835616438356164,0.020567869062099665,0.773752638310226,0.02198777375099392,0.7695090675947975,0.021584886168634444
+flat_mae,patch,logistic,adhd200_dx,61,0.005994842503189409,test,0.5538461538461539,0.05920342225835084,0.5250692869740489,0.06373726101759931,0.5299227799227799,0.06050718983622143
+flat_mae,patch,logistic,adhd200_dx,62,0.046415888336127774,train,0.8767123287671232,0.017367234579230093,0.8732883317261331,0.018102303134223993,0.8699700799902302,0.01841306453083887
+flat_mae,patch,logistic,adhd200_dx,62,0.046415888336127774,test,0.5846153846153846,0.0586200603828582,0.578226387887527,0.0595344030671501,0.5786679536679536,0.05937859601291822
+flat_mae,patch,logistic,adhd200_dx,63,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adhd200_dx,63,2.782559402207126,test,0.5384615384615384,0.060720245105569795,0.5330459770114943,0.061116563758247026,0.5337837837837838,0.061119656790161986
+flat_mae,patch,logistic,adhd200_dx,64,0.005994842503189409,train,0.7616438356164383,0.020729991019745736,0.7502379248236211,0.02245687988910006,0.7465042437564877,0.021940240232594787
+flat_mae,patch,logistic,adhd200_dx,64,0.005994842503189409,test,0.6153846153846154,0.05838481605316686,0.5966741126830479,0.06258611262848957,0.597007722007722,0.060116132732405014
+flat_mae,patch,logistic,adhd200_dx,65,0.005994842503189409,train,0.7616438356164383,0.02161347341507988,0.751426625648733,0.02308981811094439,0.7479391830005495,0.022711279961858486
+flat_mae,patch,logistic,adhd200_dx,65,0.005994842503189409,test,0.6923076923076923,0.053164698880935946,0.6697154471544715,0.0600745987168088,0.6689189189189189,0.056129481992877664
+flat_mae,patch,logistic,adhd200_dx,66,0.005994842503189409,train,0.7616438356164383,0.021225600874478022,0.751426625648733,0.022618830583177298,0.7479391830005495,0.022166094850227663
+flat_mae,patch,logistic,adhd200_dx,66,0.005994842503189409,test,0.676923076923077,0.054646846842452725,0.6612062546537603,0.058706828966698724,0.6597490347490347,0.056793099000891065
+flat_mae,patch,logistic,adhd200_dx,67,0.046415888336127774,train,0.8684931506849315,0.018039887416489733,0.8652058841632302,0.018652586025564053,0.862688526592172,0.018915955493297932
+flat_mae,patch,logistic,adhd200_dx,67,0.046415888336127774,test,0.6,0.05899682871830119,0.5775,0.06346597861177079,0.5791505791505791,0.06037054928764662
+flat_mae,patch,logistic,adhd200_dx,68,0.005994842503189409,train,0.7589041095890411,0.022723267851648438,0.745079365079365,0.02509744053733582,0.7412071808023448,0.024153722491621045
+flat_mae,patch,logistic,adhd200_dx,68,0.005994842503189409,test,0.6461538461538462,0.0559226092457431,0.6233308138070043,0.061367562292631336,0.6240347490347491,0.05793770442040621
+flat_mae,patch,logistic,adhd200_dx,69,0.005994842503189409,train,0.7753424657534247,0.020495315979667136,0.7675086999751429,0.021706199137629327,0.764379923062832,0.021594916171311655
+flat_mae,patch,logistic,adhd200_dx,69,0.005994842503189409,test,0.5538461538461539,0.06140196400305084,0.5381034060279344,0.06317904633758856,0.5386100386100386,0.061911850483780756
+flat_mae,patch,logistic,adhd200_dx,70,0.046415888336127774,train,0.8767123287671232,0.016998785514001365,0.8732883317261331,0.017715720377473303,0.8699700799902302,0.018007545352588153
+flat_mae,patch,logistic,adhd200_dx,70,0.046415888336127774,test,0.6307692307692307,0.06161260149289181,0.6235521235521235,0.06304160345581517,0.6235521235521235,0.06284827736084889
+flat_mae,patch,logistic,adhd200_dx,71,0.046415888336127774,train,0.873972602739726,0.01719617932053133,0.8701027418456397,0.017967125645906724,0.866107956280149,0.018244569176593773
+flat_mae,patch,logistic,adhd200_dx,71,0.046415888336127774,test,0.5538461538461539,0.06153762307121098,0.5469838981014179,0.06248177938605528,0.5472972972972974,0.06253792977805431
+flat_mae,patch,logistic,adhd200_dx,72,0.005994842503189409,train,0.7780821917808219,0.02136403432979856,0.7685696169833032,0.022747195386145634,0.7646546986627587,0.022351699712148475
+flat_mae,patch,logistic,adhd200_dx,72,0.005994842503189409,test,0.7384615384615385,0.05286726932309323,0.7257383966244726,0.05691980971726144,0.7224903474903475,0.05542362361576988
+flat_mae,patch,logistic,adhd200_dx,73,0.000774263682681127,train,0.6767123287671233,0.02237375080510943,0.6542672745954277,0.025059124460097493,0.6547597240031752,0.023446675949477837
+flat_mae,patch,logistic,adhd200_dx,73,0.000774263682681127,test,0.5538461538461539,0.0594706113090109,0.5250692869740489,0.06318208609007908,0.5299227799227799,0.0602026684615873
+flat_mae,patch,logistic,adhd200_dx,74,0.046415888336127774,train,0.863013698630137,0.01819109670276617,0.8588073280930866,0.019064519507832518,0.8549642791720096,0.019281632874448642
+flat_mae,patch,logistic,adhd200_dx,74,0.046415888336127774,test,0.676923076923077,0.060229179662687066,0.6741465743614228,0.06069938990241908,0.6771235521235521,0.060708661915514814
+flat_mae,patch,logistic,adhd200_dx,75,10000.0,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adhd200_dx,75,10000.0,test,0.5076923076923077,0.06054993733475372,0.4980694980694981,0.061499305600829515,0.4980694980694981,0.061220136756844686
+flat_mae,patch,logistic,adhd200_dx,76,0.005994842503189409,train,0.7835616438356164,0.02102046005775313,0.773752638310226,0.022457636024097387,0.7695090675947975,0.022043890054682106
+flat_mae,patch,logistic,adhd200_dx,76,0.005994842503189409,test,0.5384615384615384,0.05453932732306175,0.4846723044397463,0.06020184380218335,0.5033783783783784,0.05457176405896156
+flat_mae,patch,logistic,adhd200_dx,77,0.046415888336127774,train,0.873972602739726,0.01807682693349379,0.8703514949345195,0.018848966802993663,0.8668254259021799,0.019156041010845178
+flat_mae,patch,logistic,adhd200_dx,77,0.046415888336127774,test,0.6461538461538462,0.05900652055032241,0.6289401836684041,0.06280969185824659,0.6283783783783784,0.06058208559538486
+flat_mae,patch,logistic,adhd200_dx,78,0.005994842503189409,train,0.7808219178082192,0.022545938424987193,0.7740917249489385,0.02356348697234715,0.7713867008609635,0.02349825986415514
+flat_mae,patch,logistic,adhd200_dx,78,0.005994842503189409,test,0.5538461538461539,0.050594047408129894,0.48342011510002736,0.06223730091284569,0.5125482625482626,0.05192318018313542
+flat_mae,patch,logistic,adhd200_dx,79,0.046415888336127774,train,0.873972602739726,0.016502854288341317,0.8705913182883216,0.017065307867979107,0.8675428955242108,0.01719226615072895
+flat_mae,patch,logistic,adhd200_dx,79,0.046415888336127774,test,0.47692307692307695,0.06153966152676176,0.4615009746588694,0.06252886367415189,0.4623552123552124,0.061813106623232986
+flat_mae,patch,logistic,adhd200_dx,80,0.3593813663804626,train,0.9835616438356164,0.006457043123000468,0.9832844843377908,0.006564177392921705,0.9832844843377908,0.006593717959021876
+flat_mae,patch,logistic,adhd200_dx,80,0.3593813663804626,test,0.5384615384615384,0.06052609589279378,0.5374762808349146,0.060904132016181785,0.5424710424710424,0.06113460717550788
+flat_mae,patch,logistic,adhd200_dx,81,0.005994842503189409,train,0.7808219178082192,0.02118241569018412,0.7726989662473533,0.022333665419904555,0.7692342919948708,0.02214673640288412
+flat_mae,patch,logistic,adhd200_dx,81,0.005994842503189409,test,0.5692307692307692,0.05393571452919808,0.5289855072463768,0.06095492702591165,0.5390926640926641,0.055289992352924845
+flat_mae,patch,logistic,adhd200_dx,82,0.046415888336127774,train,0.8712328767123287,0.01709457318423848,0.8676567020250723,0.017759778104266606,0.8643982414361605,0.01798953182968937
+flat_mae,patch,logistic,adhd200_dx,82,0.046415888336127774,test,0.6153846153846154,0.060524406539694785,0.6150201374081972,0.06055672266851132,0.6230694980694981,0.06057061405854022
+flat_mae,patch,logistic,adhd200_dx,83,0.3593813663804626,train,0.9780821917808219,0.007591475797573075,0.9776101091890566,0.007803641307762668,0.9755602369176284,0.00847845219890788
+flat_mae,patch,logistic,adhd200_dx,83,0.3593813663804626,test,0.6615384615384615,0.056686669169553984,0.6515594541910331,0.058706671054663685,0.6505791505791505,0.05817856203597878
+flat_mae,patch,logistic,adhd200_dx,84,0.005994842503189409,train,0.7726027397260274,0.02023911924547642,0.7622970756930223,0.021660106539710292,0.7583653904866581,0.02120036665696275
+flat_mae,patch,logistic,adhd200_dx,84,0.005994842503189409,test,0.5846153846153846,0.059838836807984026,0.5644080416976918,0.06381067756323514,0.5656370656370656,0.061261480436377226
+flat_mae,patch,logistic,adhd200_dx,85,0.046415888336127774,train,0.8657534246575342,0.017509557961403622,0.8612156531050912,0.018383948980915284,0.8566739940159981,0.01855224208333624
+flat_mae,patch,logistic,adhd200_dx,85,0.046415888336127774,test,0.6153846153846154,0.06100857248911634,0.606060606060606,0.062405489865537025,0.6056949806949807,0.06197521869229974
+flat_mae,patch,logistic,adhd200_dx,86,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adhd200_dx,86,2.782559402207126,test,0.5538461538461539,0.05874369964958548,0.543030303030303,0.06038817988207976,0.542953667953668,0.06003518819625919
+flat_mae,patch,logistic,adhd200_dx,87,0.005994842503189409,train,0.7616438356164383,0.022047394997964827,0.7508415130758186,0.02370557919364191,0.7472217133785186,0.023198210544904178
+flat_mae,patch,logistic,adhd200_dx,87,0.005994842503189409,test,0.5846153846153846,0.051181903236738976,0.5308740978348035,0.061058728869043725,0.5482625482625483,0.05280501421462855
+flat_mae,patch,logistic,adhd200_dx,88,0.005994842503189409,train,0.7698630136986301,0.020909589041095893,0.761333914559721,0.02208567893630436,0.7580906148867314,0.021882564990543425
+flat_mae,patch,logistic,adhd200_dx,88,0.005994842503189409,test,0.6307692307692307,0.05859980410735277,0.6036585365853658,0.06512271871877104,0.6061776061776062,0.06055959629257867
+flat_mae,patch,logistic,adhd200_dx,89,0.046415888336127774,train,0.9013698630136986,0.015405369092786093,0.8992454913507545,0.01583441568764217,0.8975544971606522,0.01609625458655278
+flat_mae,patch,logistic,adhd200_dx,89,0.046415888336127774,test,0.5538461538461539,0.06292066198111723,0.5381034060279344,0.06561990331086198,0.5386100386100386,0.06416884384509056
+flat_mae,patch,logistic,adhd200_dx,90,0.005994842503189409,train,0.7698630136986301,0.02126099137329366,0.7602739726027397,0.02247265665382327,0.7566556756426696,0.02206959894410583
+flat_mae,patch,logistic,adhd200_dx,90,0.005994842503189409,test,0.6615384615384615,0.054909448858684574,0.6425000000000001,0.060229103255400214,0.6418918918918919,0.05716372688395238
+flat_mae,patch,logistic,adhd200_dx,91,0.046415888336127774,train,0.8657534246575342,0.017635388942161222,0.8612156531050912,0.018568613660985164,0.8566739940159981,0.01883123683622368
+flat_mae,patch,logistic,adhd200_dx,91,0.046415888336127774,test,0.6307692307692307,0.05632554571943104,0.5962732919254659,0.06466910122485398,0.6018339768339769,0.05857590421093464
+flat_mae,patch,logistic,adhd200_dx,92,0.046415888336127774,train,0.8657534246575342,0.017651352095874557,0.8614950940532335,0.01844464192480178,0.857391463638029,0.018638468995892117
+flat_mae,patch,logistic,adhd200_dx,92,0.046415888336127774,test,0.6,0.05649653076221301,0.5626293995859213,0.06274212693020575,0.5704633204633205,0.05757478748821781
+flat_mae,patch,logistic,adhd200_dx,93,0.005994842503189409,train,0.7479452054794521,0.022117535873869845,0.7348891555611696,0.023896859868020308,0.731498442938267,0.023223929220007954
+flat_mae,patch,logistic,adhd200_dx,93,0.005994842503189409,test,0.6307692307692307,0.046047580229263016,0.5666666666666667,0.062174406709198125,0.5888030888030888,0.04940116775797494
+flat_mae,patch,logistic,adhd200_dx,94,0.005994842503189409,train,0.7726027397260274,0.020976816770162478,0.763396099686819,0.022224541583159917,0.7598003297307199,0.021874359696556205
+flat_mae,patch,logistic,adhd200_dx,94,0.005994842503189409,test,0.6153846153846154,0.055571062240881104,0.5834401435529352,0.06236305682430676,0.5883204633204633,0.05741727355833475
+flat_mae,patch,logistic,adhd200_dx,95,0.005994842503189409,train,0.7808219178082192,0.021342318743229068,0.7740917249489385,0.02227124658306682,0.7713867008609635,0.022186512911127085
+flat_mae,patch,logistic,adhd200_dx,95,0.005994842503189409,test,0.5538461538461539,0.056726780042273654,0.5250692869740489,0.06176104818236881,0.5299227799227799,0.05795150249710536
+flat_mae,patch,logistic,adhd200_dx,96,0.000774263682681127,train,0.673972602739726,0.02346361512540719,0.6528671328671329,0.02577544145624314,0.6530500091591867,0.024367868763157507
+flat_mae,patch,logistic,adhd200_dx,96,0.000774263682681127,test,0.6,0.0572460773571523,0.5775,0.06183627874640533,0.5791505791505791,0.058789709270174274
+flat_mae,patch,logistic,adhd200_dx,97,0.005994842503189409,train,0.7671232876712328,0.020705093321999216,0.7565692943844204,0.022212989638314983,0.7527935519325883,0.02174798170772359
+flat_mae,patch,logistic,adhd200_dx,97,0.005994842503189409,test,0.5692307692307692,0.05417485606428317,0.5190274841437632,0.06157909892866873,0.5347490347490347,0.05486393676243797
+flat_mae,patch,logistic,adhd200_dx,98,0.005994842503189409,train,0.7561643835616438,0.02184022982838441,0.7462922032786373,0.022916735774417902,0.7430848140685107,0.022504599003110436
+flat_mae,patch,logistic,adhd200_dx,98,0.005994842503189409,test,0.6461538461538462,0.05957084390695656,0.6289401836684041,0.0631125443973189,0.6283783783783784,0.06077644959066037
+flat_mae,patch,logistic,adhd200_dx,99,0.005994842503189409,train,0.7835616438356164,0.021480153748672674,0.7762456448020858,0.022511469044813166,0.773096415704952,0.0222900131862803
+flat_mae,patch,logistic,adhd200_dx,99,0.005994842503189409,test,0.5846153846153846,0.055063505026458466,0.5411764705882354,0.06385498223982913,0.5526061776061776,0.05683970398529639
+flat_mae,patch,logistic,adhd200_dx,100,0.000774263682681127,train,0.673972602739726,0.022060997340963267,0.6497431638026272,0.02472617323407717,0.650897600293094,0.023015836470489808
+flat_mae,patch,logistic,adhd200_dx,100,0.000774263682681127,test,0.5692307692307692,0.057503494105310915,0.545,0.061292643112078674,0.5477799227799228,0.058572480030164895

data_scaling/n400_1/eval_v2/adhd200_dx__patch__logistic/log.txt

@@ -0,0 +1,241 @@
+fMRI foundation model logistic probe eval
+version: 0.1.dev66+g7ddd3aa04
+sha: 58906bf7243fb545e1349221e6921a1797e2e666, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-02-26 17:14:52
+config:
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (adhd200_dx patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_1/eval_v2/adhd200_dx__patch__logistic
+model: flat_mae
+representation: patch
+dataset: adhd200_dx
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/adhd200_dx__patch__logistic
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=True, reg_tokens=0, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: adhd200_dx (flat)
+train (n=301):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'site', 'gender', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 301
+}),
+    labels=['ADHD' 'Control'],
+    counts=[131 170]
+)
+
+validation (n=64):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'site', 'gender', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 64
+}),
+    labels=['ADHD' 'Control'],
+    counts=[28 36]
+)
+
+test (n=65):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'site', 'gender', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 65
+}),
+    labels=['ADHD' 'Control'],
+    counts=[28 37]
+)
+
+extracting features for all splits
+extract (train)  [  0/151]  eta: 0:13:18    time: 5.2900  data: 4.2381  max mem: 2698
+extract (train)  [ 20/151]  eta: 0:01:02    time: 0.2346  data: 0.0604  max mem: 2851
+extract (train)  [ 40/151]  eta: 0:00:39    time: 0.2212  data: 0.0568  max mem: 2851
+extract (train)  [ 60/151]  eta: 0:00:26    time: 0.1806  data: 0.0411  max mem: 2851
+extract (train)  [ 80/151]  eta: 0:00:19    time: 0.1849  data: 0.0446  max mem: 2851
+extract (train)  [100/151]  eta: 0:00:12    time: 0.1852  data: 0.0443  max mem: 2851
+extract (train)  [120/151]  eta: 0:00:07    time: 0.1945  data: 0.0506  max mem: 2851
+extract (train)  [140/151]  eta: 0:00:02    time: 0.1410  data: 0.0328  max mem: 2851
+extract (train)  [150/151]  eta: 0:00:00    time: 0.1364  data: 0.0320  max mem: 2851
+extract (train) Total time: 0:00:33 (0.2233 s / it)
+extract (validation)  [ 0/32]  eta: 0:01:56    time: 3.6309  data: 3.4524  max mem: 2851
+extract (validation)  [20/32]  eta: 0:00:04    time: 0.1815  data: 0.0476  max mem: 2851
+extract (validation)  [31/32]  eta: 0:00:00    time: 0.1419  data: 0.0320  max mem: 2851
+extract (validation) Total time: 0:00:09 (0.2842 s / it)
+extract (test)  [ 0/33]  eta: 0:01:45    time: 3.2064  data: 3.0594  max mem: 2851
+extract (test)  [20/33]  eta: 0:00:04    time: 0.1743  data: 0.0467  max mem: 2851
+extract (test)  [32/33]  eta: 0:00:00    time: 0.1434  data: 0.0357  max mem: 2851
+extract (test) Total time: 0:00:08 (0.2625 s / it)
+feature extraction time: 0:00:51
+train features: (301, 768)
+validation features: (64, 768)
+test features: (65, 768)
+evaluating fixed splits
+eval results (fixed splits):
+
+| model    | repr   | clf      | dataset    | trial   |        C | split   |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:-----------|:--------|---------:|:--------|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | adhd200_dx |         | 0.046416 | train   | 0.86575 |  0.017078 | 0.8615  | 0.017865 | 0.85739 |   0.018113 |
+| flat_mae | patch  | logistic | adhd200_dx |         | 0.046416 | test    | 0.67692 |  0.059505 | 0.67195 | 0.060473 | 0.67278 |   0.060382 |
+
+
+evaluating random splits (n=100)
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 1, "C": 0.000774263682681127, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.0572655890767544, "f1": 0.5250692869740489, "f1_std": 0.06076585750307687, "bacc": 0.5299227799227799, "bacc_std": 0.05785183843593267}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 2, "C": 0.046415888336127774, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.05671171570517729, "f1": 0.5376016260162602, "f1_std": 0.06225838701724851, "bacc": 0.5434362934362934, "bacc_std": 0.0579687388284011}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 3, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.05724673888027191, "f1": 0.5775, "f1_std": 0.06147789908968846, "bacc": 0.5791505791505791, "bacc_std": 0.058772705740167124}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 4, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.058280724431296206, "f1": 0.6018132810585641, "f1_std": 0.061335942045065636, "bacc": 0.6013513513513513, "bacc_std": 0.059766919147401495}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 5, "C": 0.005994842503189409, "split": "test", "acc": 0.5076923076923077, "acc_std": 0.06166973690315324, "f1": 0.5019157088122606, "f1_std": 0.06265548856678507, "bacc": 0.5024131274131274, "bacc_std": 0.06307831677995987}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 6, "C": 0.005994842503189409, "split": "test", "acc": 0.676923076923077, "acc_std": 0.05564570085522007, "f1": 0.6612062546537603, "f1_std": 0.06027805632457064, "bacc": 0.6597490347490347, "bacc_std": 0.058029226425179735}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 7, "C": 0.000774263682681127, "split": "test", "acc": 0.6, "acc_std": 0.053583809265342605, "f1": 0.570630081300813, "f1_std": 0.05875139503905504, "bacc": 0.5748069498069498, "bacc_std": 0.05497776935868238}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 8, "C": 0.000774263682681127, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.052420466665642984, "f1": 0.577922077922078, "f1_std": 0.06534726334508818, "bacc": 0.5931467181467182, "bacc_std": 0.05533943784487845}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 9, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.06166942986506083, "f1": 0.588206627680312, "f1_std": 0.06414252962711496, "bacc": 0.5878378378378378, "bacc_std": 0.0631285088032032}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 10, "C": 0.3593813663804626, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.059654952625754924, "f1": 0.5192307692307693, "f1_std": 0.06180309132185452, "bacc": 0.5207528957528957, "bacc_std": 0.060215730060988015}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 11, "C": 0.046415888336127774, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.05821271745394455, "f1": 0.5608108108108107, "f1_std": 0.05943390770577114, "bacc": 0.5608108108108107, "bacc_std": 0.05902766828279475}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 12, "C": 0.3593813663804626, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05862002808177161, "f1": 0.5830363506771205, "f1_std": 0.058889472645809154, "bacc": 0.5873552123552124, "bacc_std": 0.05932128623455292}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 13, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.06020512471061013, "f1": 0.5921814671814671, "f1_std": 0.06129344119325076, "bacc": 0.5921814671814671, "bacc_std": 0.060771826774628876}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 14, "C": 0.046415888336127774, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.05890901619743036, "f1": 0.6336682185738789, "f1_std": 0.062146457074397456, "bacc": 0.6327220077220077, "bacc_std": 0.060796763063037065}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 15, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05951964719625667, "f1": 0.6235521235521235, "f1_std": 0.06047272103557194, "bacc": 0.6235521235521235, "bacc_std": 0.060358604493063656}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 16, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06091686059971773, "f1": 0.5381034060279344, "f1_std": 0.06258601657950763, "bacc": 0.5386100386100386, "bacc_std": 0.06138737497131656}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 17, "C": 0.046415888336127774, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05871990680869055, "f1": 0.5699583435432491, "f1_std": 0.06145209366131349, "bacc": 0.5699806949806949, "bacc_std": 0.060034387646530685}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 18, "C": 0.046415888336127774, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.059894629368741695, "f1": 0.606060606060606, "f1_std": 0.06123124537723429, "bacc": 0.6056949806949807, "bacc_std": 0.060530000381377465}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 19, "C": 0.046415888336127774, "split": "test", "acc": 0.676923076923077, "acc_std": 0.06036583539420486, "f1": 0.6655231560891939, "f1_std": 0.06391764659581159, "bacc": 0.6640926640926641, "bacc_std": 0.06276648339768712}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 20, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05593973934395675, "f1": 0.5966741126830479, "f1_std": 0.05956773634875029, "bacc": 0.597007722007722, "bacc_std": 0.05736341308013188}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 21, "C": 0.005994842503189409, "split": "test", "acc": 0.7076923076923077, "acc_std": 0.05511287180916318, "f1": 0.6934723256391164, "f1_std": 0.05935428871932758, "bacc": 0.6911196911196911, "bacc_std": 0.05751842221042249}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 22, "C": 0.046415888336127774, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05756622119338048, "f1": 0.6094688776736361, "f1_std": 0.059000555408504046, "bacc": 0.61003861003861, "bacc_std": 0.05916352132013179}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 23, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05582575596967586, "f1": 0.5905769715293525, "f1_std": 0.06092880156668017, "bacc": 0.5926640926640927, "bacc_std": 0.05723420917198336}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 24, "C": 0.005994842503189409, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.05450640481338385, "f1": 0.6299171842650104, "f1_std": 0.06302193643219771, "bacc": 0.6332046332046332, "bacc_std": 0.057101951438083956}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 25, "C": 0.005994842503189409, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.05497279370140541, "f1": 0.6167649320687003, "f1_std": 0.0619366915297259, "bacc": 0.6196911196911197, "bacc_std": 0.05722116357302214}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 26, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.0551576546265822, "f1": 0.5834401435529352, "f1_std": 0.06222831061513817, "bacc": 0.5883204633204633, "bacc_std": 0.05713123455936296}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 27, "C": 0.005994842503189409, "split": "test", "acc": 0.7230769230769231, "acc_std": 0.05303444454024486, "f1": 0.7176640926640927, "f1_std": 0.05405444556692365, "bacc": 0.7176640926640927, "bacc_std": 0.05377801209791008}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 28, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.0541713258502937, "f1": 0.5834401435529352, "f1_std": 0.06113224527527147, "bacc": 0.5883204633204633, "bacc_std": 0.05613889901634252}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 29, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06031255082470189, "f1": 0.545, "f1_std": 0.06448818718183073, "bacc": 0.5477799227799228, "bacc_std": 0.06153564747621803}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 30, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.054693119111150615, "f1": 0.5833333333333333, "f1_std": 0.0571336811586876, "bacc": 0.5834942084942085, "bacc_std": 0.0555374175286818}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 31, "C": 0.3593813663804626, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.059494446063696034, "f1": 0.6235521235521235, "f1_std": 0.060752597031315435, "bacc": 0.6235521235521235, "bacc_std": 0.06019783739826568}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 32, "C": 0.046415888336127774, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.05741953155112201, "f1": 0.6289401836684041, "f1_std": 0.06168051084712195, "bacc": 0.6283783783783784, "bacc_std": 0.05967790280427032}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 33, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06055749676300353, "f1": 0.564176245210728, "f1_std": 0.06119702477144015, "bacc": 0.5651544401544402, "bacc_std": 0.061196796584670324}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 34, "C": 0.046415888336127774, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06031380659742999, "f1": 0.564176245210728, "f1_std": 0.06124768065242325, "bacc": 0.5651544401544402, "bacc_std": 0.06163327286376625}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 35, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.057498826332246115, "f1": 0.6036585365853658, "f1_std": 0.06332847127867035, "bacc": 0.6061776061776062, "bacc_std": 0.05911397772664288}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 36, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05808497346140251, "f1": 0.6018132810585641, "f1_std": 0.061038667690533496, "bacc": 0.6013513513513513, "bacc_std": 0.05966645370475177}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 37, "C": 0.000774263682681127, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.04606058269395328, "f1": 0.554672513017265, "f1_std": 0.05930753190875318, "bacc": 0.5752895752895753, "bacc_std": 0.04843900617142849}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 38, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.05746423842899906, "f1": 0.5512820512820513, "f1_std": 0.06012947737164356, "bacc": 0.5521235521235521, "bacc_std": 0.05820206915826109}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 39, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.0615423075721229, "f1": 0.6153846153846154, "f1_std": 0.06404714127278875, "bacc": 0.6148648648648649, "bacc_std": 0.06243158787689593}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 40, "C": 0.005994842503189409, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.04774166180799899, "f1": 0.6130952380952381, "f1_std": 0.06002177730440167, "bacc": 0.6245173745173745, "bacc_std": 0.05071997773002365}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 41, "C": 0.046415888336127774, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.05978969850515453, "f1": 0.644808743169399, "f1_std": 0.05982158801541909, "bacc": 0.6500965250965252, "bacc_std": 0.059381385241960516}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 42, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.06271468972810364, "f1": 0.6235521235521235, "f1_std": 0.06377761357442285, "bacc": 0.6235521235521235, "bacc_std": 0.06339578516973475}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 43, "C": 0.000774263682681127, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.060512243369694504, "f1": 0.5501153550371699, "f1_std": 0.06653549857759201, "bacc": 0.556949806949807, "bacc_std": 0.061573446806877014}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 44, "C": 0.046415888336127774, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05764464903293434, "f1": 0.606060606060606, "f1_std": 0.05921578486629486, "bacc": 0.6056949806949807, "bacc_std": 0.05872212325340343}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 45, "C": 0.005994842503189409, "split": "test", "acc": 0.5076923076923077, "acc_std": 0.05950008701699171, "f1": 0.4871794871794872, "f1_std": 0.06236568104858205, "bacc": 0.48938223938223935, "bacc_std": 0.06038077148270777}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 46, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.059431141598626364, "f1": 0.5966741126830479, "f1_std": 0.06273333213340375, "bacc": 0.597007722007722, "bacc_std": 0.06045994029541409}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 47, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06130707267325128, "f1": 0.5565302144249512, "f1_std": 0.06265016787306586, "bacc": 0.5564671814671815, "bacc_std": 0.06192038296137542}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 48, "C": 0.046415888336127774, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.060914148529912664, "f1": 0.6264367816091954, "f1_std": 0.06175308554170438, "bacc": 0.627895752895753, "bacc_std": 0.061938043329468624}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 49, "C": 0.046415888336127774, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06234080032061976, "f1": 0.5500119360229172, "f1_std": 0.06306192727957428, "bacc": 0.5516409266409266, "bacc_std": 0.06359814774523685}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 50, "C": 0.3593813663804626, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.058078836443265204, "f1": 0.5565302144249512, "f1_std": 0.06045777555896192, "bacc": 0.5564671814671815, "bacc_std": 0.05953511430094472}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 51, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.06084209069386094, "f1": 0.5921814671814671, "f1_std": 0.06265787776687684, "bacc": 0.5921814671814671, "bacc_std": 0.06242020102102711}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 52, "C": 166.81005372000556, "split": "test", "acc": 0.5230769230769231, "acc_std": 0.06048955311045832, "f1": 0.5115151515151515, "f1_std": 0.06166501453032125, "bacc": 0.5115830115830116, "bacc_std": 0.061061613368097144}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 53, "C": 0.046415888336127774, "split": "test", "acc": 0.5230769230769231, "acc_std": 0.05926792041680342, "f1": 0.49987589972697943, "f1_std": 0.06221983316200884, "bacc": 0.502895752895753, "bacc_std": 0.05999061139233182}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 54, "C": 0.000774263682681127, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.05798408269649392, "f1": 0.5125, "f1_std": 0.06153215536539919, "bacc": 0.5164092664092664, "bacc_std": 0.05870489862503548}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 55, "C": 0.046415888336127774, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06375048253675135, "f1": 0.5666666666666667, "f1_std": 0.06380417795717837, "bacc": 0.5694980694980695, "bacc_std": 0.06425679767514877}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 56, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05939442724840168, "f1": 0.5745454545454545, "f1_std": 0.06070472188012568, "bacc": 0.5743243243243243, "bacc_std": 0.060147054191353515}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 57, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.059952320305882836, "f1": 0.5905769715293525, "f1_std": 0.06525933957683168, "bacc": 0.5926640926640927, "bacc_std": 0.061561379891102906}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 58, "C": 0.005994842503189409, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.0565710194064628, "f1": 0.6515594541910331, "f1_std": 0.058946245373878714, "bacc": 0.6505791505791505, "bacc_std": 0.05777064500738074}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 59, "C": 0.046415888336127774, "split": "test", "acc": 0.5076923076923077, "acc_std": 0.05877225123128174, "f1": 0.48, "f1_std": 0.062216358119852586, "bacc": 0.48503861003861004, "bacc_std": 0.05948481241158543}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 60, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05591638728372842, "f1": 0.61, "f1_std": 0.06102715186358429, "bacc": 0.6105212355212355, "bacc_std": 0.05797635992295799}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 61, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.05920342225835084, "f1": 0.5250692869740489, "f1_std": 0.06373726101759931, "bacc": 0.5299227799227799, "bacc_std": 0.06050718983622143}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 62, "C": 0.046415888336127774, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.0586200603828582, "f1": 0.578226387887527, "f1_std": 0.0595344030671501, "bacc": 0.5786679536679536, "bacc_std": 0.05937859601291822}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 63, "C": 2.782559402207126, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.060720245105569795, "f1": 0.5330459770114943, "f1_std": 0.061116563758247026, "bacc": 0.5337837837837838, "bacc_std": 0.061119656790161986}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 64, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05838481605316686, "f1": 0.5966741126830479, "f1_std": 0.06258611262848957, "bacc": 0.597007722007722, "bacc_std": 0.060116132732405014}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 65, "C": 0.005994842503189409, "split": "test", "acc": 0.6923076923076923, "acc_std": 0.053164698880935946, "f1": 0.6697154471544715, "f1_std": 0.0600745987168088, "bacc": 0.6689189189189189, "bacc_std": 0.056129481992877664}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 66, "C": 0.005994842503189409, "split": "test", "acc": 0.676923076923077, "acc_std": 0.054646846842452725, "f1": 0.6612062546537603, "f1_std": 0.058706828966698724, "bacc": 0.6597490347490347, "bacc_std": 0.056793099000891065}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 67, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.05899682871830119, "f1": 0.5775, "f1_std": 0.06346597861177079, "bacc": 0.5791505791505791, "bacc_std": 0.06037054928764662}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 68, "C": 0.005994842503189409, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.0559226092457431, "f1": 0.6233308138070043, "f1_std": 0.061367562292631336, "bacc": 0.6240347490347491, "bacc_std": 0.05793770442040621}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 69, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06140196400305084, "f1": 0.5381034060279344, "f1_std": 0.06317904633758856, "bacc": 0.5386100386100386, "bacc_std": 0.061911850483780756}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 70, "C": 0.046415888336127774, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.06161260149289181, "f1": 0.6235521235521235, "f1_std": 0.06304160345581517, "bacc": 0.6235521235521235, "bacc_std": 0.06284827736084889}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 71, "C": 0.046415888336127774, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06153762307121098, "f1": 0.5469838981014179, "f1_std": 0.06248177938605528, "bacc": 0.5472972972972974, "bacc_std": 0.06253792977805431}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 72, "C": 0.005994842503189409, "split": "test", "acc": 0.7384615384615385, "acc_std": 0.05286726932309323, "f1": 0.7257383966244726, "f1_std": 0.05691980971726144, "bacc": 0.7224903474903475, "bacc_std": 0.05542362361576988}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 73, "C": 0.000774263682681127, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.0594706113090109, "f1": 0.5250692869740489, "f1_std": 0.06318208609007908, "bacc": 0.5299227799227799, "bacc_std": 0.0602026684615873}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 74, "C": 0.046415888336127774, "split": "test", "acc": 0.676923076923077, "acc_std": 0.060229179662687066, "f1": 0.6741465743614228, "f1_std": 0.06069938990241908, "bacc": 0.6771235521235521, "bacc_std": 0.060708661915514814}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 75, "C": 10000.0, "split": "test", "acc": 0.5076923076923077, "acc_std": 0.06054993733475372, "f1": 0.4980694980694981, "f1_std": 0.061499305600829515, "bacc": 0.4980694980694981, "bacc_std": 0.061220136756844686}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 76, "C": 0.005994842503189409, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.05453932732306175, "f1": 0.4846723044397463, "f1_std": 0.06020184380218335, "bacc": 0.5033783783783784, "bacc_std": 0.05457176405896156}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 77, "C": 0.046415888336127774, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.05900652055032241, "f1": 0.6289401836684041, "f1_std": 0.06280969185824659, "bacc": 0.6283783783783784, "bacc_std": 0.06058208559538486}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 78, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.050594047408129894, "f1": 0.48342011510002736, "f1_std": 0.06223730091284569, "bacc": 0.5125482625482626, "bacc_std": 0.05192318018313542}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 79, "C": 0.046415888336127774, "split": "test", "acc": 0.47692307692307695, "acc_std": 0.06153966152676176, "f1": 0.4615009746588694, "f1_std": 0.06252886367415189, "bacc": 0.4623552123552124, "bacc_std": 0.061813106623232986}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 80, "C": 0.3593813663804626, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.06052609589279378, "f1": 0.5374762808349146, "f1_std": 0.060904132016181785, "bacc": 0.5424710424710424, "bacc_std": 0.06113460717550788}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 81, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.05393571452919808, "f1": 0.5289855072463768, "f1_std": 0.06095492702591165, "bacc": 0.5390926640926641, "bacc_std": 0.055289992352924845}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 82, "C": 0.046415888336127774, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.060524406539694785, "f1": 0.6150201374081972, "f1_std": 0.06055672266851132, "bacc": 0.6230694980694981, "bacc_std": 0.06057061405854022}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 83, "C": 0.3593813663804626, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.056686669169553984, "f1": 0.6515594541910331, "f1_std": 0.058706671054663685, "bacc": 0.6505791505791505, "bacc_std": 0.05817856203597878}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 84, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.059838836807984026, "f1": 0.5644080416976918, "f1_std": 0.06381067756323514, "bacc": 0.5656370656370656, "bacc_std": 0.061261480436377226}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 85, "C": 0.046415888336127774, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.06100857248911634, "f1": 0.606060606060606, "f1_std": 0.062405489865537025, "bacc": 0.6056949806949807, "bacc_std": 0.06197521869229974}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 86, "C": 2.782559402207126, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.05874369964958548, "f1": 0.543030303030303, "f1_std": 0.06038817988207976, "bacc": 0.542953667953668, "bacc_std": 0.06003518819625919}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 87, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.051181903236738976, "f1": 0.5308740978348035, "f1_std": 0.061058728869043725, "bacc": 0.5482625482625483, "bacc_std": 0.05280501421462855}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 88, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05859980410735277, "f1": 0.6036585365853658, "f1_std": 0.06512271871877104, "bacc": 0.6061776061776062, "bacc_std": 0.06055959629257867}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 89, "C": 0.046415888336127774, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06292066198111723, "f1": 0.5381034060279344, "f1_std": 0.06561990331086198, "bacc": 0.5386100386100386, "bacc_std": 0.06416884384509056}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 90, "C": 0.005994842503189409, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.054909448858684574, "f1": 0.6425000000000001, "f1_std": 0.060229103255400214, "bacc": 0.6418918918918919, "bacc_std": 0.05716372688395238}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 91, "C": 0.046415888336127774, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05632554571943104, "f1": 0.5962732919254659, "f1_std": 0.06466910122485398, "bacc": 0.6018339768339769, "bacc_std": 0.05857590421093464}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 92, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.05649653076221301, "f1": 0.5626293995859213, "f1_std": 0.06274212693020575, "bacc": 0.5704633204633205, "bacc_std": 0.05757478748821781}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 93, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.046047580229263016, "f1": 0.5666666666666667, "f1_std": 0.062174406709198125, "bacc": 0.5888030888030888, "bacc_std": 0.04940116775797494}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 94, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.055571062240881104, "f1": 0.5834401435529352, "f1_std": 0.06236305682430676, "bacc": 0.5883204633204633, "bacc_std": 0.05741727355833475}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 95, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.056726780042273654, "f1": 0.5250692869740489, "f1_std": 0.06176104818236881, "bacc": 0.5299227799227799, "bacc_std": 0.05795150249710536}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 96, "C": 0.000774263682681127, "split": "test", "acc": 0.6, "acc_std": 0.0572460773571523, "f1": 0.5775, "f1_std": 0.06183627874640533, "bacc": 0.5791505791505791, "bacc_std": 0.058789709270174274}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 97, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.05417485606428317, "f1": 0.5190274841437632, "f1_std": 0.06157909892866873, "bacc": 0.5347490347490347, "bacc_std": 0.05486393676243797}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 98, "C": 0.005994842503189409, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.05957084390695656, "f1": 0.6289401836684041, "f1_std": 0.0631125443973189, "bacc": 0.6283783783783784, "bacc_std": 0.06077644959066037}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 99, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.055063505026458466, "f1": 0.5411764705882354, "f1_std": 0.06385498223982913, "bacc": 0.5526061776061776, "bacc_std": 0.05683970398529639}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 100, "C": 0.000774263682681127, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.057503494105310915, "f1": 0.545, "f1_std": 0.061292643112078674, "bacc": 0.5477799227799228, "bacc_std": 0.058572480030164895}
+eval results (random splits):
+
+| model    | repr   | clf      | dataset    | split   |   n_trials |      C |   C_std |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:-----------|:--------|-----------:|-------:|--------:|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | adhd200_dx | train   |        100 | 101.76 |  999.96 | 0.81411 |   0.08431 | 0.80593 | 0.090002 | 0.80322 |   0.089851 |
+| flat_mae | patch  | logistic | adhd200_dx | test    |        100 | 101.76 |  999.96 | 0.60062 |   0.04962 | 0.58054 | 0.052554 | 0.58356 |   0.05055  |
+
+
+done! total time: 0:04:31

data_scaling/n400_1/eval_v2/adni_ad_vs_cn__patch__logistic/config.yaml

@@ -0,0 +1,30 @@
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (adni_ad_vs_cn patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_1/eval_v2/adni_ad_vs_cn__patch__logistic
+model: flat_mae
+representation: patch
+dataset: adni_ad_vs_cn
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/adni_ad_vs_cn__patch__logistic
+remote_dir: null

data_scaling/n400_1/eval_v2/adni_ad_vs_cn__patch__logistic/eval_table.csv

@@ -0,0 +1,203 @@
+model,repr,clf,dataset,trial,C,split,acc,acc_std,f1,f1_std,bacc,bacc_std
+flat_mae,patch,logistic,adni_ad_vs_cn,,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,,166.81005372000556,test,0.5853658536585366,0.07386635197164572,0.4558938329430133,0.07245003296675631,0.4548611111111111,0.07947699372970114
+flat_mae,patch,logistic,adni_ad_vs_cn,1,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,1,21.54434690031882,test,0.8048780487804879,0.05391018595692995,0.7152777777777778,0.08597702690792597,0.7016129032258065,0.0834766569027455
+flat_mae,patch,logistic,adni_ad_vs_cn,2,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,2,21.54434690031882,test,0.7560975609756098,0.061981767002958564,0.6693548387096775,0.0847810777401297,0.6693548387096775,0.08654595352042534
+flat_mae,patch,logistic,adni_ad_vs_cn,3,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,3,166.81005372000556,test,0.7560975609756098,0.05325102462895257,0.6117424242424243,0.09015788579022892,0.6016129032258064,0.07713430210946796
+flat_mae,patch,logistic,adni_ad_vs_cn,4,0.046415888336127774,train,0.924119241192412,0.012935588153713175,0.8862014274385408,0.020871005067042907,0.861492316542033,0.024073938862968727
+flat_mae,patch,logistic,adni_ad_vs_cn,4,0.046415888336127774,test,0.8536585365853658,0.05470161339513611,0.8136363636363637,0.06720795718216024,0.8354838709677419,0.07056314719409477
+flat_mae,patch,logistic,adni_ad_vs_cn,5,0.046415888336127774,train,0.926829268292683,0.012334338164017772,0.8897471366126266,0.02002614236975282,0.8632591009943298,0.023785347749360104
+flat_mae,patch,logistic,adni_ad_vs_cn,5,0.046415888336127774,test,0.6829268292682927,0.04635265590127199,0.4696517412935323,0.065422983266204,0.4854838709677419,0.05345639630513192
+flat_mae,patch,logistic,adni_ad_vs_cn,6,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,6,2.782559402207126,test,0.7073170731707317,0.06410170012993327,0.5729166666666666,0.08631320225809522,0.5693548387096774,0.08067822186854992
+flat_mae,patch,logistic,adni_ad_vs_cn,7,0.3593813663804626,train,0.991869918699187,0.004641668090191123,0.9885825675299359,0.0065639564923437215,0.986605308570959,0.008639465232023502
+flat_mae,patch,logistic,adni_ad_vs_cn,7,0.3593813663804626,test,0.7073170731707317,0.05869847532392946,0.5729166666666666,0.08395517558019032,0.5693548387096774,0.07826616690763738
+flat_mae,patch,logistic,adni_ad_vs_cn,8,0.3593813663804626,train,0.991869918699187,0.00461317621811351,0.9885825675299359,0.006501085858554967,0.986605308570959,0.008000841641525334
+flat_mae,patch,logistic,adni_ad_vs_cn,8,0.3593813663804626,test,0.7560975609756098,0.06170290160622114,0.6693548387096775,0.08213433376507209,0.6693548387096775,0.08344518591907935
+flat_mae,patch,logistic,adni_ad_vs_cn,9,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,9,166.81005372000556,test,0.7073170731707317,0.06964300913228409,0.6272727272727273,0.08186535399146078,0.6370967741935484,0.0869084747678408
+flat_mae,patch,logistic,adni_ad_vs_cn,10,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,10,1291.5496650148827,test,0.7073170731707317,0.07018173061640827,0.6272727272727273,0.0858037201099534,0.6370967741935484,0.09160772015097365
+flat_mae,patch,logistic,adni_ad_vs_cn,11,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,11,166.81005372000556,test,0.7804878048780488,0.061380722538478046,0.6917293233082706,0.08542095474280705,0.685483870967742,0.08496856069118341
+flat_mae,patch,logistic,adni_ad_vs_cn,12,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,12,21.54434690031882,test,0.6585365853658537,0.06313378710611696,0.5017361111111112,0.08091185453150397,0.5032258064516129,0.07471761504437113
+flat_mae,patch,logistic,adni_ad_vs_cn,13,0.3593813663804626,train,0.997289972899729,0.002755280701176509,0.9961941891766453,0.0039028219143826424,0.9941860465116279,0.005911038248454255
+flat_mae,patch,logistic,adni_ad_vs_cn,13,0.3593813663804626,test,0.7804878048780488,0.054031607732385134,0.6660633484162897,0.08691143769468333,0.6516129032258065,0.08175133891987646
+flat_mae,patch,logistic,adni_ad_vs_cn,14,0.046415888336127774,train,0.9105691056910569,0.01353761047142437,0.8612481626234888,0.023349795026764732,0.8283753800640973,0.025975696495811385
+flat_mae,patch,logistic,adni_ad_vs_cn,14,0.046415888336127774,test,0.8292682926829268,0.0509403776016267,0.7402714932126697,0.08921050106114684,0.717741935483871,0.0842606625964113
+flat_mae,patch,logistic,adni_ad_vs_cn,15,0.3593813663804626,train,0.986449864498645,0.005850957923652251,0.9808134274809954,0.008401970857523429,0.9749774015942148,0.011271733415687477
+flat_mae,patch,logistic,adni_ad_vs_cn,15,0.3593813663804626,test,0.7804878048780488,0.05665143325588912,0.6660633484162897,0.09083147691761732,0.6516129032258065,0.08354267479538048
+flat_mae,patch,logistic,adni_ad_vs_cn,16,0.3593813663804626,train,0.991869918699187,0.004753838668632086,0.9885825675299359,0.006704088940573208,0.986605308570959,0.008353730666905649
+flat_mae,patch,logistic,adni_ad_vs_cn,16,0.3593813663804626,test,0.8536585365853658,0.036985904995716565,0.7415966386554622,0.09010351692403558,0.7,0.07582110524121896
+flat_mae,patch,logistic,adni_ad_vs_cn,17,0.046415888336127774,train,0.924119241192412,0.012287246136675815,0.8862014274385408,0.0195801745994442,0.861492316542033,0.02263412954812497
+flat_mae,patch,logistic,adni_ad_vs_cn,17,0.046415888336127774,test,0.7073170731707317,0.05518830454863606,0.5340909090909092,0.08348630064649379,0.535483870967742,0.06999235360348482
+flat_mae,patch,logistic,adni_ad_vs_cn,18,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,18,2.782559402207126,test,0.8048780487804879,0.06311846410995944,0.7515151515151515,0.07703548038342849,0.7693548387096774,0.08152995300018441
+flat_mae,patch,logistic,adni_ad_vs_cn,19,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,19,21.54434690031882,test,0.7560975609756098,0.06460590271082287,0.6693548387096775,0.08362620796164759,0.6693548387096775,0.08453113853423398
+flat_mae,patch,logistic,adni_ad_vs_cn,20,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,20,2.782559402207126,test,0.7804878048780488,0.06609250273148327,0.7410526315789474,0.0707631977471325,0.7870967741935484,0.07397149512867292
+flat_mae,patch,logistic,adni_ad_vs_cn,21,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,21,21.54434690031882,test,0.7804878048780488,0.05172572152902073,0.6660633484162897,0.08751576640814474,0.6516129032258065,0.07990633245682943
+flat_mae,patch,logistic,adni_ad_vs_cn,22,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,22,2.782559402207126,test,0.8536585365853658,0.047627220907025335,0.7670454545454546,0.08740481102412913,0.7338709677419355,0.0833820450831452
+flat_mae,patch,logistic,adni_ad_vs_cn,23,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,23,166.81005372000556,test,0.7317073170731707,0.06721340550688852,0.6479313036690086,0.08291295410040263,0.6532258064516129,0.08648400181674835
+flat_mae,patch,logistic,adni_ad_vs_cn,24,0.3593813663804626,train,0.989159891598916,0.0055988504024470686,0.9847141673570836,0.007967143085526405,0.9807913550825869,0.010281544556214377
+flat_mae,patch,logistic,adni_ad_vs_cn,24,0.3593813663804626,test,0.7317073170731707,0.059689360443455255,0.5918552036199095,0.09116953488845253,0.5854838709677419,0.08148273592548415
+flat_mae,patch,logistic,adni_ad_vs_cn,25,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,25,166.81005372000556,test,0.7560975609756098,0.06146873767116401,0.6693548387096775,0.08531968628628803,0.6693548387096775,0.08700309292766682
+flat_mae,patch,logistic,adni_ad_vs_cn,26,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,26,21.54434690031882,test,0.8048780487804879,0.05987373189942769,0.7354838709677419,0.07956328975010646,0.7354838709677419,0.08143188109767019
+flat_mae,patch,logistic,adni_ad_vs_cn,27,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,27,21.54434690031882,test,0.8292682926829268,0.053336416726256615,0.7402714932126697,0.08857955623317884,0.717741935483871,0.08549314490169431
+flat_mae,patch,logistic,adni_ad_vs_cn,28,0.3593813663804626,train,0.997289972899729,0.0026771501688070956,0.9961941891766453,0.0037882972519350617,0.9941860465116279,0.0057434210016849814
+flat_mae,patch,logistic,adni_ad_vs_cn,28,0.3593813663804626,test,0.7560975609756098,0.060003152802352695,0.6440972222222222,0.08843170797091617,0.635483870967742,0.08470727584242786
+flat_mae,patch,logistic,adni_ad_vs_cn,29,0.3593813663804626,train,0.991869918699187,0.004462834581547718,0.9885825675299359,0.006297733073871022,0.986605308570959,0.007968638096503606
+flat_mae,patch,logistic,adni_ad_vs_cn,29,0.3593813663804626,test,0.8048780487804879,0.056361542193498454,0.7354838709677419,0.07796743322014178,0.7354838709677419,0.08159971709108084
+flat_mae,patch,logistic,adni_ad_vs_cn,30,0.046415888336127774,train,0.9105691056910569,0.012992135604913918,0.8612481626234888,0.022285895763517225,0.8283753800640973,0.025010091535329418
+flat_mae,patch,logistic,adni_ad_vs_cn,30,0.046415888336127774,test,0.7560975609756098,0.04731797836700164,0.569327731092437,0.09281233422854243,0.567741935483871,0.06965098990984576
+flat_mae,patch,logistic,adni_ad_vs_cn,31,0.3593813663804626,train,0.994579945799458,0.0038069119557094374,0.9923570836785418,0.005437462520314869,0.9883720930232558,0.008167154137539464
+flat_mae,patch,logistic,adni_ad_vs_cn,31,0.3593813663804626,test,0.7317073170731707,0.05560599045872572,0.5918552036199095,0.08649327137234365,0.5854838709677419,0.07762647100725746
+flat_mae,patch,logistic,adni_ad_vs_cn,32,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,32,21.54434690031882,test,0.6585365853658537,0.06743637360820451,0.5651515151515152,0.07810186683862563,0.5709677419354839,0.0826876448939389
+flat_mae,patch,logistic,adni_ad_vs_cn,33,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,33,21.54434690031882,test,0.6097560975609756,0.07298609224308505,0.5287356321839081,0.0797788950736887,0.5387096774193548,0.08881179219496932
+flat_mae,patch,logistic,adni_ad_vs_cn,34,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,34,21.54434690031882,test,0.8292682926829268,0.05463841446742919,0.7885040530582166,0.0654277858466808,0.8193548387096774,0.06988346855973626
+flat_mae,patch,logistic,adni_ad_vs_cn,35,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,35,166.81005372000556,test,0.7317073170731707,0.06486409165844072,0.6232247284878863,0.0884525389368805,0.6193548387096774,0.08697701419892165
+flat_mae,patch,logistic,adni_ad_vs_cn,36,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,36,166.81005372000556,test,0.7073170731707317,0.05625426680889042,0.5340909090909092,0.08541830736460779,0.535483870967742,0.07257608150527094
+flat_mae,patch,logistic,adni_ad_vs_cn,37,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,37,2.782559402207126,test,0.6829268292682927,0.06537711952424433,0.5547201336675021,0.08495254629098649,0.5532258064516129,0.08168736880745213
+flat_mae,patch,logistic,adni_ad_vs_cn,38,0.3593813663804626,train,0.997289972899729,0.0026877130120796697,0.9961941891766453,0.0038054863671016896,0.9941860465116279,0.0057660819852174226
+flat_mae,patch,logistic,adni_ad_vs_cn,38,0.3593813663804626,test,0.7317073170731707,0.049525313653171074,0.5512437810945273,0.0841628563860447,0.5516129032258065,0.06724606305857803
+flat_mae,patch,logistic,adni_ad_vs_cn,39,0.3593813663804626,train,0.997289972899729,0.0025873819640399344,0.9961941891766453,0.0036602690193837874,0.9941860465116279,0.005550836887969385
+flat_mae,patch,logistic,adni_ad_vs_cn,39,0.3593813663804626,test,0.6341463414634146,0.06571821663451059,0.48621553884711777,0.07695666036352872,0.48709677419354835,0.07387955295935043
+flat_mae,patch,logistic,adni_ad_vs_cn,40,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,40,166.81005372000556,test,0.7317073170731707,0.0623072351548938,0.6232247284878863,0.0848774926356396,0.6193548387096774,0.08417209317529728
+flat_mae,patch,logistic,adni_ad_vs_cn,41,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,41,21.54434690031882,test,0.8536585365853658,0.05149470785951873,0.8016129032258064,0.07164623516925737,0.8016129032258064,0.07698034797937726
+flat_mae,patch,logistic,adni_ad_vs_cn,42,0.3593813663804626,train,0.991869918699187,0.004800647596485512,0.9885825675299359,0.00678634214967772,0.986605308570959,0.008688890813461672
+flat_mae,patch,logistic,adni_ad_vs_cn,42,0.3593813663804626,test,0.7804878048780488,0.05854737705401343,0.6917293233082706,0.08368967241507448,0.685483870967742,0.08413581038398465
+flat_mae,patch,logistic,adni_ad_vs_cn,43,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,43,166.81005372000556,test,0.6829268292682927,0.0726202872984208,0.5839188134270101,0.08901748428394309,0.5870967741935484,0.09259364520394478
+flat_mae,patch,logistic,adni_ad_vs_cn,44,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,44,21.54434690031882,test,0.8048780487804879,0.04308827732770946,0.6554621848739496,0.09582669010765707,0.6338709677419355,0.07503657769165208
+flat_mae,patch,logistic,adni_ad_vs_cn,45,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,45,166.81005372000556,test,0.8536585365853658,0.04911601678978526,0.7864583333333333,0.07799161096707341,0.7677419354838709,0.08022909730334937
+flat_mae,patch,logistic,adni_ad_vs_cn,46,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,46,166.81005372000556,test,0.7804878048780488,0.04019497395189424,0.5886287625418061,0.09275681193760191,0.5838709677419355,0.06625664312009948
+flat_mae,patch,logistic,adni_ad_vs_cn,47,0.3593813663804626,train,0.989159891598916,0.005433422538166638,0.9847141673570836,0.007736839807458125,0.9807913550825869,0.010034075272180615
+flat_mae,patch,logistic,adni_ad_vs_cn,47,0.3593813663804626,test,0.8048780487804879,0.05041137138839583,0.6893939393939394,0.08964154258823055,0.667741935483871,0.08086889969948656
+flat_mae,patch,logistic,adni_ad_vs_cn,48,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,48,166.81005372000556,test,0.6829268292682927,0.072823550328523,0.6072218128224024,0.08509614734598585,0.6209677419354839,0.09329840342238152
+flat_mae,patch,logistic,adni_ad_vs_cn,49,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,49,21.54434690031882,test,0.8048780487804879,0.05952169056594976,0.7354838709677419,0.08261690067169539,0.7354838709677419,0.0857223029053736
+flat_mae,patch,logistic,adni_ad_vs_cn,50,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,50,166.81005372000556,test,0.7317073170731707,0.06637237443426919,0.6479313036690086,0.08253963685421709,0.6532258064516129,0.08374469470825835
+flat_mae,patch,logistic,adni_ad_vs_cn,51,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,51,21.54434690031882,test,0.6341463414634146,0.06873549817308175,0.5199063231850116,0.0820297925501416,0.5209677419354839,0.08429176424331608
+flat_mae,patch,logistic,adni_ad_vs_cn,52,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,52,21.54434690031882,test,0.6829268292682927,0.0695735458830013,0.5839188134270101,0.08550833222013729,0.5870967741935484,0.08871766972794255
+flat_mae,patch,logistic,adni_ad_vs_cn,53,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,53,2.782559402207126,test,0.8048780487804879,0.042022061569330284,0.6554621848739496,0.09436140665069609,0.6338709677419355,0.07330387170306715
+flat_mae,patch,logistic,adni_ad_vs_cn,54,0.046415888336127774,train,0.926829268292683,0.011968598981472596,0.887597730067579,0.020169343428268784,0.8551647629221792,0.0235200137614075
+flat_mae,patch,logistic,adni_ad_vs_cn,54,0.046415888336127774,test,0.7804878048780488,0.05957970980413577,0.6660633484162897,0.09420053828039614,0.6516129032258065,0.08611389259230211
+flat_mae,patch,logistic,adni_ad_vs_cn,55,0.3593813663804626,train,0.991869918699187,0.004925442755702336,0.9885825675299359,0.006955378621758862,0.986605308570959,0.008713627881419795
+flat_mae,patch,logistic,adni_ad_vs_cn,55,0.3593813663804626,test,0.8048780487804879,0.05776472826087685,0.7354838709677419,0.08057535054223001,0.7354838709677419,0.08376707560405186
+flat_mae,patch,logistic,adni_ad_vs_cn,56,0.3593813663804626,train,0.997289972899729,0.0026833374080187755,0.9961941891766453,0.0038008554975543407,0.9941860465116279,0.005756694788133296
+flat_mae,patch,logistic,adni_ad_vs_cn,56,0.3593813663804626,test,0.7073170731707317,0.054980846808360684,0.5340909090909092,0.08040178397247043,0.535483870967742,0.0686516922523801
+flat_mae,patch,logistic,adni_ad_vs_cn,57,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,57,21.54434690031882,test,0.7560975609756098,0.06530341029071804,0.6693548387096775,0.0903257246699847,0.6693548387096775,0.09142580713451595
+flat_mae,patch,logistic,adni_ad_vs_cn,58,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,58,2.782559402207126,test,0.8048780487804879,0.05515593599451731,0.7152777777777778,0.08583845250229447,0.7016129032258065,0.08385552115824027
+flat_mae,patch,logistic,adni_ad_vs_cn,59,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,59,166.81005372000556,test,0.7804878048780488,0.049204281043947334,0.6328358208955224,0.09244720167071581,0.6177419354838709,0.07575783004901666
+flat_mae,patch,logistic,adni_ad_vs_cn,60,0.046415888336127774,train,0.9186991869918699,0.012684106910101056,0.8744897959183674,0.02154979871481047,0.8417700714931383,0.02473750607127376
+flat_mae,patch,logistic,adni_ad_vs_cn,60,0.046415888336127774,test,0.7804878048780488,0.05625765067939033,0.6660633484162897,0.09245433308035948,0.6516129032258065,0.0844890670168005
+flat_mae,patch,logistic,adni_ad_vs_cn,61,0.3593813663804626,train,0.991869918699187,0.00466337318352436,0.9885825675299359,0.006602379486343818,0.986605308570959,0.008649567506473712
+flat_mae,patch,logistic,adni_ad_vs_cn,61,0.3593813663804626,test,0.7317073170731707,0.05640214238550867,0.5918552036199095,0.08738230100677158,0.5854838709677419,0.0781524575308169
+flat_mae,patch,logistic,adni_ad_vs_cn,62,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,62,21.54434690031882,test,0.6585365853658537,0.06223019575361796,0.5017361111111112,0.07837861446662327,0.5032258064516129,0.07363439587572747
+flat_mae,patch,logistic,adni_ad_vs_cn,63,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,63,21.54434690031882,test,0.7560975609756098,0.060238747681116124,0.6693548387096775,0.083269184975022,0.6693548387096775,0.08488400262820564
+flat_mae,patch,logistic,adni_ad_vs_cn,64,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,64,166.81005372000556,test,0.7560975609756098,0.06029511006244203,0.6440972222222222,0.08617357999023835,0.635483870967742,0.08060846907217784
+flat_mae,patch,logistic,adni_ad_vs_cn,65,0.046415888336127774,train,0.926829268292683,0.012239690915627453,0.887597730067579,0.020758968378944132,0.8551647629221792,0.024296919374212883
+flat_mae,patch,logistic,adni_ad_vs_cn,65,0.046415888336127774,test,0.7560975609756098,0.06233596677932339,0.6440972222222222,0.09274211761519914,0.635483870967742,0.08992350842424744
+flat_mae,patch,logistic,adni_ad_vs_cn,66,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,66,2.782559402207126,test,0.7073170731707317,0.06096304260738657,0.5729166666666666,0.0871052016458449,0.5693548387096774,0.08034549511209382
+flat_mae,patch,logistic,adni_ad_vs_cn,67,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,67,166.81005372000556,test,0.8048780487804879,0.056125210577278394,0.7152777777777778,0.08812599671116532,0.7016129032258065,0.08497914318392903
+flat_mae,patch,logistic,adni_ad_vs_cn,68,0.046415888336127774,train,0.926829268292683,0.012950299408582198,0.8897471366126266,0.02096213295936407,0.8632591009943298,0.024241733144302547
+flat_mae,patch,logistic,adni_ad_vs_cn,68,0.046415888336127774,test,0.7560975609756098,0.06491940731982912,0.6893939393939394,0.07727275671658856,0.7032258064516128,0.0831321600548176
+flat_mae,patch,logistic,adni_ad_vs_cn,69,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,69,166.81005372000556,test,0.7804878048780488,0.06038297367483309,0.6917293233082706,0.08572546880720444,0.685483870967742,0.08539629256940474
+flat_mae,patch,logistic,adni_ad_vs_cn,70,0.046415888336127774,train,0.9132791327913279,0.013213014977961608,0.8661224489795918,0.023014966495215682,0.8341893335524694,0.026255686930354082
+flat_mae,patch,logistic,adni_ad_vs_cn,70,0.046415888336127774,test,0.7804878048780488,0.037534326941996675,0.5886287625418061,0.08634322435858609,0.5838709677419355,0.06174752148863623
+flat_mae,patch,logistic,adni_ad_vs_cn,71,0.046415888336127774,train,0.9295392953929539,0.01249628024647272,0.8922893838692294,0.02091377701784035,0.8609787164105513,0.024753402067189283
+flat_mae,patch,logistic,adni_ad_vs_cn,71,0.046415888336127774,test,0.7073170731707317,0.05568542135155785,0.5340909090909092,0.08611944502630993,0.535483870967742,0.07329033244831828
+flat_mae,patch,logistic,adni_ad_vs_cn,72,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,72,166.81005372000556,test,0.7804878048780488,0.05474594397684203,0.6660633484162897,0.08821744239812374,0.6516129032258065,0.08136827413298017
+flat_mae,patch,logistic,adni_ad_vs_cn,73,0.3593813663804626,train,0.991869918699187,0.0046030877304059775,0.9885825675299359,0.006494663373041181,0.986605308570959,0.008107414530145076
+flat_mae,patch,logistic,adni_ad_vs_cn,73,0.3593813663804626,test,0.8292682926829268,0.05580359882552229,0.7602339181286549,0.08268777083551411,0.7516129032258064,0.08496527263906467
+flat_mae,patch,logistic,adni_ad_vs_cn,74,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,74,166.81005372000556,test,0.7560975609756098,0.05823777391461187,0.6440972222222222,0.0856649423359408,0.635483870967742,0.08176734405366246
+flat_mae,patch,logistic,adni_ad_vs_cn,75,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,75,166.81005372000556,test,0.8536585365853658,0.053442113435003324,0.8016129032258064,0.07369994504183223,0.8016129032258064,0.07947899047647251
+flat_mae,patch,logistic,adni_ad_vs_cn,76,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,76,21.54434690031882,test,0.6585365853658537,0.04357176279185874,0.39705882352941174,0.015967585899545497,0.43548387096774194,0.028813585072035616
+flat_mae,patch,logistic,adni_ad_vs_cn,77,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,77,166.81005372000556,test,0.7317073170731707,0.05990148565222108,0.6232247284878863,0.08474017033244868,0.6193548387096774,0.0828688440692745
+flat_mae,patch,logistic,adni_ad_vs_cn,78,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,78,166.81005372000556,test,0.7317073170731707,0.06827916242467993,0.6232247284878863,0.09537404928292492,0.6193548387096774,0.0922470177798827
+flat_mae,patch,logistic,adni_ad_vs_cn,79,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,79,166.81005372000556,test,0.9024390243902439,0.046890102762323665,0.8757575757575757,0.05690882441074128,0.9016129032258065,0.055331638400796954
+flat_mae,patch,logistic,adni_ad_vs_cn,80,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,80,166.81005372000556,test,0.6341463414634146,0.06349930557452856,0.48621553884711777,0.07700010581437944,0.48709677419354835,0.0742686326943568
+flat_mae,patch,logistic,adni_ad_vs_cn,81,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,81,21.54434690031882,test,0.7804878048780488,0.0541150425909677,0.6660633484162897,0.09247558794321739,0.6516129032258065,0.08325339627712096
+flat_mae,patch,logistic,adni_ad_vs_cn,82,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,82,2.782559402207126,test,0.7073170731707317,0.06313203447944146,0.603225806451613,0.08263421425543989,0.603225806451613,0.08322627056634782
+flat_mae,patch,logistic,adni_ad_vs_cn,83,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,83,2.782559402207126,test,0.7317073170731707,0.06509956681152483,0.6232247284878863,0.09001409711109588,0.6193548387096774,0.08814094497486052
+flat_mae,patch,logistic,adni_ad_vs_cn,84,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,84,2.782559402207126,test,0.6829268292682927,0.06655880824792237,0.5547201336675021,0.08524929216672607,0.5532258064516129,0.08219143006334956
+flat_mae,patch,logistic,adni_ad_vs_cn,85,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,85,166.81005372000556,test,0.8048780487804879,0.05791398573919672,0.7354838709677419,0.07732404140369112,0.7354838709677419,0.08040211446476905
+flat_mae,patch,logistic,adni_ad_vs_cn,86,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,86,21.54434690031882,test,0.6097560975609756,0.06897945620075109,0.5030303030303029,0.07609141856558463,0.5048387096774194,0.08024937582622983
+flat_mae,patch,logistic,adni_ad_vs_cn,87,0.3593813663804626,train,0.991869918699187,0.004531823320695356,0.9885825675299359,0.006399847326545155,0.986605308570959,0.008215456002119838
+flat_mae,patch,logistic,adni_ad_vs_cn,87,0.3593813663804626,test,0.8780487804878049,0.04664659492996655,0.8144796380090498,0.08183768703725718,0.7838709677419355,0.08338934644747208
+flat_mae,patch,logistic,adni_ad_vs_cn,88,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,88,2.782559402207126,test,0.7560975609756098,0.06213722795680428,0.6693548387096775,0.08127940765455609,0.6693548387096775,0.08238885516096608
+flat_mae,patch,logistic,adni_ad_vs_cn,89,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,89,166.81005372000556,test,0.7804878048780488,0.06501681473812182,0.7280766396462786,0.07623836691415395,0.7532258064516129,0.08150183338399926
+flat_mae,patch,logistic,adni_ad_vs_cn,90,0.3593813663804626,train,0.983739837398374,0.006601374588596787,0.9768796992481203,0.009570674458254123,0.9691634481058427,0.012887383869980625
+flat_mae,patch,logistic,adni_ad_vs_cn,90,0.3593813663804626,test,0.7560975609756098,0.0619282461725146,0.6440972222222222,0.09439593471415984,0.635483870967742,0.08958584324017421
+flat_mae,patch,logistic,adni_ad_vs_cn,91,0.3593813663804626,train,0.994579945799458,0.003758470529218197,0.9924192620593311,0.005257372648629068,0.9924192620593311,0.005985416770038435
+flat_mae,patch,logistic,adni_ad_vs_cn,91,0.3593813663804626,test,0.7073170731707317,0.06478526282724706,0.603225806451613,0.08404953414526098,0.603225806451613,0.08625785527274546
+flat_mae,patch,logistic,adni_ad_vs_cn,92,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,92,2.782559402207126,test,0.6585365853658537,0.056398746088727233,0.4564393939393939,0.06952246989062644,0.4693548387096774,0.061472894376107486
+flat_mae,patch,logistic,adni_ad_vs_cn,93,0.046415888336127774,train,0.940379403794038,0.010816505469148143,0.9088602478893479,0.018096460568637056,0.8761401922918892,0.022387356718131784
+flat_mae,patch,logistic,adni_ad_vs_cn,93,0.046415888336127774,test,0.7073170731707317,0.06771329136501665,0.603225806451613,0.0861114929051163,0.603225806451613,0.0858741970624694
+flat_mae,patch,logistic,adni_ad_vs_cn,94,10000.0,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,94,10000.0,test,0.6829268292682927,0.06245980980255294,0.5547201336675021,0.08223894148352015,0.5532258064516129,0.07924385177226353
+flat_mae,patch,logistic,adni_ad_vs_cn,95,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,95,21.54434690031882,test,0.7317073170731707,0.05993055668048907,0.6232247284878863,0.08259845078475109,0.6193548387096774,0.0803052274945738
+flat_mae,patch,logistic,adni_ad_vs_cn,96,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,96,166.81005372000556,test,0.7560975609756098,0.061250424612157114,0.6893939393939394,0.0754613459593449,0.7032258064516128,0.08066997063549053
+flat_mae,patch,logistic,adni_ad_vs_cn,97,0.3593813663804626,train,0.991869918699187,0.0048159735975308576,0.9885825675299359,0.006804175807948498,0.986605308570959,0.008667941043457045
+flat_mae,patch,logistic,adni_ad_vs_cn,97,0.3593813663804626,test,0.7804878048780488,0.040695207832269396,0.5886287625418061,0.09217418692311527,0.5838709677419355,0.06582024357369012
+flat_mae,patch,logistic,adni_ad_vs_cn,98,0.005994842503189409,train,0.8319783197831978,0.013373674842603676,0.695576964019587,0.030307143767250615,0.6638178979373819,0.02508040687479159
+flat_mae,patch,logistic,adni_ad_vs_cn,98,0.005994842503189409,test,0.7804878048780488,0.048083285864591145,0.6328358208955224,0.09401663382734154,0.6177419354838709,0.07778912842782644
+flat_mae,patch,logistic,adni_ad_vs_cn,99,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,99,166.81005372000556,test,0.7317073170731707,0.0640586249934384,0.6232247284878863,0.08839824034097298,0.6193548387096774,0.08582491015383371
+flat_mae,patch,logistic,adni_ad_vs_cn,100,0.046415888336127774,train,0.9186991869918699,0.013166452961292803,0.8732249198350893,0.023130337592401583,0.837722902457063,0.02681260064564242
+flat_mae,patch,logistic,adni_ad_vs_cn,100,0.046415888336127774,test,0.6829268292682927,0.0563032496317555,0.5176470588235295,0.07654181812916153,0.5193548387096775,0.06803311085594083

data_scaling/n400_1/eval_v2/adni_ad_vs_cn__patch__logistic/log.txt

@@ -0,0 +1,240 @@
+fMRI foundation model logistic probe eval
+version: 0.1.dev66+g7ddd3aa04
+sha: 58906bf7243fb545e1349221e6921a1797e2e666, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-02-26 17:14:46
+config:
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (adni_ad_vs_cn patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_1/eval_v2/adni_ad_vs_cn__patch__logistic
+model: flat_mae
+representation: patch
+dataset: adni_ad_vs_cn
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/adni_ad_vs_cn__patch__logistic
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=True, reg_tokens=0, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: adni_ad_vs_cn (flat)
+train (n=328):
+ADNIDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 525
+}),
+    labels=[0 1],
+    counts=[251  77]
+)
+
+validation (n=41):
+ADNIDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 66
+}),
+    labels=[0 1],
+    counts=[31 10]
+)
+
+test (n=41):
+ADNIDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 66
+}),
+    labels=[0 1],
+    counts=[32  9]
+)
+
+extracting features for all splits
+extract (train)  [  0/164]  eta: 0:11:34    time: 4.2346  data: 3.4810  max mem: 2698
+extract (train)  [ 20/164]  eta: 0:00:57    time: 0.2070  data: 0.0702  max mem: 2851
+extract (train)  [ 40/164]  eta: 0:00:36    time: 0.1929  data: 0.0554  max mem: 2851
+extract (train)  [ 60/164]  eta: 0:00:27    time: 0.1948  data: 0.0513  max mem: 2851
+extract (train)  [ 80/164]  eta: 0:00:21    time: 0.2073  data: 0.0564  max mem: 2851
+extract (train)  [100/164]  eta: 0:00:15    time: 0.2011  data: 0.0535  max mem: 2851
+extract (train)  [120/164]  eta: 0:00:10    time: 0.2009  data: 0.0547  max mem: 2851
+extract (train)  [140/164]  eta: 0:00:05    time: 0.1686  data: 0.0426  max mem: 2851
+extract (train)  [160/164]  eta: 0:00:00    time: 0.1476  data: 0.0401  max mem: 2851
+extract (train)  [163/164]  eta: 0:00:00    time: 0.1497  data: 0.0406  max mem: 2851
+extract (train) Total time: 0:00:35 (0.2168 s / it)
+extract (validation)  [ 0/21]  eta: 0:01:35    time: 4.5584  data: 4.4337  max mem: 2851
+extract (validation)  [20/21]  eta: 0:00:00    time: 0.1444  data: 0.0385  max mem: 2851
+extract (validation) Total time: 0:00:07 (0.3697 s / it)
+extract (test)  [ 0/21]  eta: 0:01:33    time: 4.4379  data: 4.3180  max mem: 2851
+extract (test)  [20/21]  eta: 0:00:00    time: 0.1477  data: 0.0367  max mem: 2851
+extract (test) Total time: 0:00:07 (0.3679 s / it)
+feature extraction time: 0:00:51
+train features: (328, 768)
+validation features: (41, 768)
+test features: (41, 768)
+evaluating fixed splits
+eval results (fixed splits):
+
+| model    | repr   | clf      | dataset       | trial   |      C | split   |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:--------------|:--------|-------:|:--------|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | adni_ad_vs_cn |         | 166.81 | train   | 1       |  0        | 1       |  0       | 1       |   0        |
+| flat_mae | patch  | logistic | adni_ad_vs_cn |         | 166.81 | test    | 0.58537 |  0.073866 | 0.45589 |  0.07245 | 0.45486 |   0.079477 |
+
+
+evaluating random splits (n=100)
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 1, "C": 21.54434690031882, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05391018595692995, "f1": 0.7152777777777778, "f1_std": 0.08597702690792597, "bacc": 0.7016129032258065, "bacc_std": 0.0834766569027455}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 2, "C": 21.54434690031882, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.061981767002958564, "f1": 0.6693548387096775, "f1_std": 0.0847810777401297, "bacc": 0.6693548387096775, "bacc_std": 0.08654595352042534}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 3, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.05325102462895257, "f1": 0.6117424242424243, "f1_std": 0.09015788579022892, "bacc": 0.6016129032258064, "bacc_std": 0.07713430210946796}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 4, "C": 0.046415888336127774, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.05470161339513611, "f1": 0.8136363636363637, "f1_std": 0.06720795718216024, "bacc": 0.8354838709677419, "bacc_std": 0.07056314719409477}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 5, "C": 0.046415888336127774, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.04635265590127199, "f1": 0.4696517412935323, "f1_std": 0.065422983266204, "bacc": 0.4854838709677419, "bacc_std": 0.05345639630513192}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 6, "C": 2.782559402207126, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06410170012993327, "f1": 0.5729166666666666, "f1_std": 0.08631320225809522, "bacc": 0.5693548387096774, "bacc_std": 0.08067822186854992}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 7, "C": 0.3593813663804626, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.05869847532392946, "f1": 0.5729166666666666, "f1_std": 0.08395517558019032, "bacc": 0.5693548387096774, "bacc_std": 0.07826616690763738}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 8, "C": 0.3593813663804626, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06170290160622114, "f1": 0.6693548387096775, "f1_std": 0.08213433376507209, "bacc": 0.6693548387096775, "bacc_std": 0.08344518591907935}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 9, "C": 166.81005372000556, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06964300913228409, "f1": 0.6272727272727273, "f1_std": 0.08186535399146078, "bacc": 0.6370967741935484, "bacc_std": 0.0869084747678408}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 10, "C": 1291.5496650148827, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.07018173061640827, "f1": 0.6272727272727273, "f1_std": 0.0858037201099534, "bacc": 0.6370967741935484, "bacc_std": 0.09160772015097365}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 11, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.061380722538478046, "f1": 0.6917293233082706, "f1_std": 0.08542095474280705, "bacc": 0.685483870967742, "bacc_std": 0.08496856069118341}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 12, "C": 21.54434690031882, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.06313378710611696, "f1": 0.5017361111111112, "f1_std": 0.08091185453150397, "bacc": 0.5032258064516129, "bacc_std": 0.07471761504437113}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 13, "C": 0.3593813663804626, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.054031607732385134, "f1": 0.6660633484162897, "f1_std": 0.08691143769468333, "bacc": 0.6516129032258065, "bacc_std": 0.08175133891987646}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 14, "C": 0.046415888336127774, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.0509403776016267, "f1": 0.7402714932126697, "f1_std": 0.08921050106114684, "bacc": 0.717741935483871, "bacc_std": 0.0842606625964113}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 15, "C": 0.3593813663804626, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05665143325588912, "f1": 0.6660633484162897, "f1_std": 0.09083147691761732, "bacc": 0.6516129032258065, "bacc_std": 0.08354267479538048}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 16, "C": 0.3593813663804626, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.036985904995716565, "f1": 0.7415966386554622, "f1_std": 0.09010351692403558, "bacc": 0.7, "bacc_std": 0.07582110524121896}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 17, "C": 0.046415888336127774, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.05518830454863606, "f1": 0.5340909090909092, "f1_std": 0.08348630064649379, "bacc": 0.535483870967742, "bacc_std": 0.06999235360348482}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 18, "C": 2.782559402207126, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.06311846410995944, "f1": 0.7515151515151515, "f1_std": 0.07703548038342849, "bacc": 0.7693548387096774, "bacc_std": 0.08152995300018441}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 19, "C": 21.54434690031882, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06460590271082287, "f1": 0.6693548387096775, "f1_std": 0.08362620796164759, "bacc": 0.6693548387096775, "bacc_std": 0.08453113853423398}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 20, "C": 2.782559402207126, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.06609250273148327, "f1": 0.7410526315789474, "f1_std": 0.0707631977471325, "bacc": 0.7870967741935484, "bacc_std": 0.07397149512867292}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 21, "C": 21.54434690031882, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05172572152902073, "f1": 0.6660633484162897, "f1_std": 0.08751576640814474, "bacc": 0.6516129032258065, "bacc_std": 0.07990633245682943}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 22, "C": 2.782559402207126, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.047627220907025335, "f1": 0.7670454545454546, "f1_std": 0.08740481102412913, "bacc": 0.7338709677419355, "bacc_std": 0.0833820450831452}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 23, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06721340550688852, "f1": 0.6479313036690086, "f1_std": 0.08291295410040263, "bacc": 0.6532258064516129, "bacc_std": 0.08648400181674835}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 24, "C": 0.3593813663804626, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.059689360443455255, "f1": 0.5918552036199095, "f1_std": 0.09116953488845253, "bacc": 0.5854838709677419, "bacc_std": 0.08148273592548415}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 25, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06146873767116401, "f1": 0.6693548387096775, "f1_std": 0.08531968628628803, "bacc": 0.6693548387096775, "bacc_std": 0.08700309292766682}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 26, "C": 21.54434690031882, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05987373189942769, "f1": 0.7354838709677419, "f1_std": 0.07956328975010646, "bacc": 0.7354838709677419, "bacc_std": 0.08143188109767019}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 27, "C": 21.54434690031882, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.053336416726256615, "f1": 0.7402714932126697, "f1_std": 0.08857955623317884, "bacc": 0.717741935483871, "bacc_std": 0.08549314490169431}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 28, "C": 0.3593813663804626, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.060003152802352695, "f1": 0.6440972222222222, "f1_std": 0.08843170797091617, "bacc": 0.635483870967742, "bacc_std": 0.08470727584242786}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 29, "C": 0.3593813663804626, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.056361542193498454, "f1": 0.7354838709677419, "f1_std": 0.07796743322014178, "bacc": 0.7354838709677419, "bacc_std": 0.08159971709108084}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 30, "C": 0.046415888336127774, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.04731797836700164, "f1": 0.569327731092437, "f1_std": 0.09281233422854243, "bacc": 0.567741935483871, "bacc_std": 0.06965098990984576}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 31, "C": 0.3593813663804626, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.05560599045872572, "f1": 0.5918552036199095, "f1_std": 0.08649327137234365, "bacc": 0.5854838709677419, "bacc_std": 0.07762647100725746}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 32, "C": 21.54434690031882, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.06743637360820451, "f1": 0.5651515151515152, "f1_std": 0.07810186683862563, "bacc": 0.5709677419354839, "bacc_std": 0.0826876448939389}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 33, "C": 21.54434690031882, "split": "test", "acc": 0.6097560975609756, "acc_std": 0.07298609224308505, "f1": 0.5287356321839081, "f1_std": 0.0797788950736887, "bacc": 0.5387096774193548, "bacc_std": 0.08881179219496932}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 34, "C": 21.54434690031882, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.05463841446742919, "f1": 0.7885040530582166, "f1_std": 0.0654277858466808, "bacc": 0.8193548387096774, "bacc_std": 0.06988346855973626}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 35, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06486409165844072, "f1": 0.6232247284878863, "f1_std": 0.0884525389368805, "bacc": 0.6193548387096774, "bacc_std": 0.08697701419892165}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 36, "C": 166.81005372000556, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.05625426680889042, "f1": 0.5340909090909092, "f1_std": 0.08541830736460779, "bacc": 0.535483870967742, "bacc_std": 0.07257608150527094}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 37, "C": 2.782559402207126, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06537711952424433, "f1": 0.5547201336675021, "f1_std": 0.08495254629098649, "bacc": 0.5532258064516129, "bacc_std": 0.08168736880745213}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 38, "C": 0.3593813663804626, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.049525313653171074, "f1": 0.5512437810945273, "f1_std": 0.0841628563860447, "bacc": 0.5516129032258065, "bacc_std": 0.06724606305857803}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 39, "C": 0.3593813663804626, "split": "test", "acc": 0.6341463414634146, "acc_std": 0.06571821663451059, "f1": 0.48621553884711777, "f1_std": 0.07695666036352872, "bacc": 0.48709677419354835, "bacc_std": 0.07387955295935043}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 40, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.0623072351548938, "f1": 0.6232247284878863, "f1_std": 0.0848774926356396, "bacc": 0.6193548387096774, "bacc_std": 0.08417209317529728}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 41, "C": 21.54434690031882, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.05149470785951873, "f1": 0.8016129032258064, "f1_std": 0.07164623516925737, "bacc": 0.8016129032258064, "bacc_std": 0.07698034797937726}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 42, "C": 0.3593813663804626, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05854737705401343, "f1": 0.6917293233082706, "f1_std": 0.08368967241507448, "bacc": 0.685483870967742, "bacc_std": 0.08413581038398465}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 43, "C": 166.81005372000556, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.0726202872984208, "f1": 0.5839188134270101, "f1_std": 0.08901748428394309, "bacc": 0.5870967741935484, "bacc_std": 0.09259364520394478}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 44, "C": 21.54434690031882, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.04308827732770946, "f1": 0.6554621848739496, "f1_std": 0.09582669010765707, "bacc": 0.6338709677419355, "bacc_std": 0.07503657769165208}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 45, "C": 166.81005372000556, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.04911601678978526, "f1": 0.7864583333333333, "f1_std": 0.07799161096707341, "bacc": 0.7677419354838709, "bacc_std": 0.08022909730334937}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 46, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.04019497395189424, "f1": 0.5886287625418061, "f1_std": 0.09275681193760191, "bacc": 0.5838709677419355, "bacc_std": 0.06625664312009948}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 47, "C": 0.3593813663804626, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05041137138839583, "f1": 0.6893939393939394, "f1_std": 0.08964154258823055, "bacc": 0.667741935483871, "bacc_std": 0.08086889969948656}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 48, "C": 166.81005372000556, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.072823550328523, "f1": 0.6072218128224024, "f1_std": 0.08509614734598585, "bacc": 0.6209677419354839, "bacc_std": 0.09329840342238152}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 49, "C": 21.54434690031882, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05952169056594976, "f1": 0.7354838709677419, "f1_std": 0.08261690067169539, "bacc": 0.7354838709677419, "bacc_std": 0.0857223029053736}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 50, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06637237443426919, "f1": 0.6479313036690086, "f1_std": 0.08253963685421709, "bacc": 0.6532258064516129, "bacc_std": 0.08374469470825835}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 51, "C": 21.54434690031882, "split": "test", "acc": 0.6341463414634146, "acc_std": 0.06873549817308175, "f1": 0.5199063231850116, "f1_std": 0.0820297925501416, "bacc": 0.5209677419354839, "bacc_std": 0.08429176424331608}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 52, "C": 21.54434690031882, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.0695735458830013, "f1": 0.5839188134270101, "f1_std": 0.08550833222013729, "bacc": 0.5870967741935484, "bacc_std": 0.08871766972794255}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 53, "C": 2.782559402207126, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.042022061569330284, "f1": 0.6554621848739496, "f1_std": 0.09436140665069609, "bacc": 0.6338709677419355, "bacc_std": 0.07330387170306715}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 54, "C": 0.046415888336127774, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05957970980413577, "f1": 0.6660633484162897, "f1_std": 0.09420053828039614, "bacc": 0.6516129032258065, "bacc_std": 0.08611389259230211}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 55, "C": 0.3593813663804626, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05776472826087685, "f1": 0.7354838709677419, "f1_std": 0.08057535054223001, "bacc": 0.7354838709677419, "bacc_std": 0.08376707560405186}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 56, "C": 0.3593813663804626, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.054980846808360684, "f1": 0.5340909090909092, "f1_std": 0.08040178397247043, "bacc": 0.535483870967742, "bacc_std": 0.0686516922523801}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 57, "C": 21.54434690031882, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06530341029071804, "f1": 0.6693548387096775, "f1_std": 0.0903257246699847, "bacc": 0.6693548387096775, "bacc_std": 0.09142580713451595}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 58, "C": 2.782559402207126, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05515593599451731, "f1": 0.7152777777777778, "f1_std": 0.08583845250229447, "bacc": 0.7016129032258065, "bacc_std": 0.08385552115824027}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 59, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.049204281043947334, "f1": 0.6328358208955224, "f1_std": 0.09244720167071581, "bacc": 0.6177419354838709, "bacc_std": 0.07575783004901666}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 60, "C": 0.046415888336127774, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05625765067939033, "f1": 0.6660633484162897, "f1_std": 0.09245433308035948, "bacc": 0.6516129032258065, "bacc_std": 0.0844890670168005}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 61, "C": 0.3593813663804626, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.05640214238550867, "f1": 0.5918552036199095, "f1_std": 0.08738230100677158, "bacc": 0.5854838709677419, "bacc_std": 0.0781524575308169}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 62, "C": 21.54434690031882, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.06223019575361796, "f1": 0.5017361111111112, "f1_std": 0.07837861446662327, "bacc": 0.5032258064516129, "bacc_std": 0.07363439587572747}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 63, "C": 21.54434690031882, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.060238747681116124, "f1": 0.6693548387096775, "f1_std": 0.083269184975022, "bacc": 0.6693548387096775, "bacc_std": 0.08488400262820564}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 64, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06029511006244203, "f1": 0.6440972222222222, "f1_std": 0.08617357999023835, "bacc": 0.635483870967742, "bacc_std": 0.08060846907217784}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 65, "C": 0.046415888336127774, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06233596677932339, "f1": 0.6440972222222222, "f1_std": 0.09274211761519914, "bacc": 0.635483870967742, "bacc_std": 0.08992350842424744}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 66, "C": 2.782559402207126, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06096304260738657, "f1": 0.5729166666666666, "f1_std": 0.0871052016458449, "bacc": 0.5693548387096774, "bacc_std": 0.08034549511209382}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 67, "C": 166.81005372000556, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.056125210577278394, "f1": 0.7152777777777778, "f1_std": 0.08812599671116532, "bacc": 0.7016129032258065, "bacc_std": 0.08497914318392903}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 68, "C": 0.046415888336127774, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06491940731982912, "f1": 0.6893939393939394, "f1_std": 0.07727275671658856, "bacc": 0.7032258064516128, "bacc_std": 0.0831321600548176}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 69, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.06038297367483309, "f1": 0.6917293233082706, "f1_std": 0.08572546880720444, "bacc": 0.685483870967742, "bacc_std": 0.08539629256940474}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 70, "C": 0.046415888336127774, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.037534326941996675, "f1": 0.5886287625418061, "f1_std": 0.08634322435858609, "bacc": 0.5838709677419355, "bacc_std": 0.06174752148863623}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 71, "C": 0.046415888336127774, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.05568542135155785, "f1": 0.5340909090909092, "f1_std": 0.08611944502630993, "bacc": 0.535483870967742, "bacc_std": 0.07329033244831828}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 72, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05474594397684203, "f1": 0.6660633484162897, "f1_std": 0.08821744239812374, "bacc": 0.6516129032258065, "bacc_std": 0.08136827413298017}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 73, "C": 0.3593813663804626, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.05580359882552229, "f1": 0.7602339181286549, "f1_std": 0.08268777083551411, "bacc": 0.7516129032258064, "bacc_std": 0.08496527263906467}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 74, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.05823777391461187, "f1": 0.6440972222222222, "f1_std": 0.0856649423359408, "bacc": 0.635483870967742, "bacc_std": 0.08176734405366246}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 75, "C": 166.81005372000556, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.053442113435003324, "f1": 0.8016129032258064, "f1_std": 0.07369994504183223, "bacc": 0.8016129032258064, "bacc_std": 0.07947899047647251}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 76, "C": 21.54434690031882, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.04357176279185874, "f1": 0.39705882352941174, "f1_std": 0.015967585899545497, "bacc": 0.43548387096774194, "bacc_std": 0.028813585072035616}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 77, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.05990148565222108, "f1": 0.6232247284878863, "f1_std": 0.08474017033244868, "bacc": 0.6193548387096774, "bacc_std": 0.0828688440692745}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 78, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06827916242467993, "f1": 0.6232247284878863, "f1_std": 0.09537404928292492, "bacc": 0.6193548387096774, "bacc_std": 0.0922470177798827}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 79, "C": 166.81005372000556, "split": "test", "acc": 0.9024390243902439, "acc_std": 0.046890102762323665, "f1": 0.8757575757575757, "f1_std": 0.05690882441074128, "bacc": 0.9016129032258065, "bacc_std": 0.055331638400796954}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 80, "C": 166.81005372000556, "split": "test", "acc": 0.6341463414634146, "acc_std": 0.06349930557452856, "f1": 0.48621553884711777, "f1_std": 0.07700010581437944, "bacc": 0.48709677419354835, "bacc_std": 0.0742686326943568}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 81, "C": 21.54434690031882, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.0541150425909677, "f1": 0.6660633484162897, "f1_std": 0.09247558794321739, "bacc": 0.6516129032258065, "bacc_std": 0.08325339627712096}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 82, "C": 2.782559402207126, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06313203447944146, "f1": 0.603225806451613, "f1_std": 0.08263421425543989, "bacc": 0.603225806451613, "bacc_std": 0.08322627056634782}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 83, "C": 2.782559402207126, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06509956681152483, "f1": 0.6232247284878863, "f1_std": 0.09001409711109588, "bacc": 0.6193548387096774, "bacc_std": 0.08814094497486052}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 84, "C": 2.782559402207126, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06655880824792237, "f1": 0.5547201336675021, "f1_std": 0.08524929216672607, "bacc": 0.5532258064516129, "bacc_std": 0.08219143006334956}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 85, "C": 166.81005372000556, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05791398573919672, "f1": 0.7354838709677419, "f1_std": 0.07732404140369112, "bacc": 0.7354838709677419, "bacc_std": 0.08040211446476905}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 86, "C": 21.54434690031882, "split": "test", "acc": 0.6097560975609756, "acc_std": 0.06897945620075109, "f1": 0.5030303030303029, "f1_std": 0.07609141856558463, "bacc": 0.5048387096774194, "bacc_std": 0.08024937582622983}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 87, "C": 0.3593813663804626, "split": "test", "acc": 0.8780487804878049, "acc_std": 0.04664659492996655, "f1": 0.8144796380090498, "f1_std": 0.08183768703725718, "bacc": 0.7838709677419355, "bacc_std": 0.08338934644747208}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 88, "C": 2.782559402207126, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06213722795680428, "f1": 0.6693548387096775, "f1_std": 0.08127940765455609, "bacc": 0.6693548387096775, "bacc_std": 0.08238885516096608}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 89, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.06501681473812182, "f1": 0.7280766396462786, "f1_std": 0.07623836691415395, "bacc": 0.7532258064516129, "bacc_std": 0.08150183338399926}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 90, "C": 0.3593813663804626, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.0619282461725146, "f1": 0.6440972222222222, "f1_std": 0.09439593471415984, "bacc": 0.635483870967742, "bacc_std": 0.08958584324017421}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 91, "C": 0.3593813663804626, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06478526282724706, "f1": 0.603225806451613, "f1_std": 0.08404953414526098, "bacc": 0.603225806451613, "bacc_std": 0.08625785527274546}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 92, "C": 2.782559402207126, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.056398746088727233, "f1": 0.4564393939393939, "f1_std": 0.06952246989062644, "bacc": 0.4693548387096774, "bacc_std": 0.061472894376107486}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 93, "C": 0.046415888336127774, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06771329136501665, "f1": 0.603225806451613, "f1_std": 0.0861114929051163, "bacc": 0.603225806451613, "bacc_std": 0.0858741970624694}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 94, "C": 10000.0, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06245980980255294, "f1": 0.5547201336675021, "f1_std": 0.08223894148352015, "bacc": 0.5532258064516129, "bacc_std": 0.07924385177226353}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 95, "C": 21.54434690031882, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.05993055668048907, "f1": 0.6232247284878863, "f1_std": 0.08259845078475109, "bacc": 0.6193548387096774, "bacc_std": 0.0803052274945738}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 96, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.061250424612157114, "f1": 0.6893939393939394, "f1_std": 0.0754613459593449, "bacc": 0.7032258064516128, "bacc_std": 0.08066997063549053}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 97, "C": 0.3593813663804626, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.040695207832269396, "f1": 0.5886287625418061, "f1_std": 0.09217418692311527, "bacc": 0.5838709677419355, "bacc_std": 0.06582024357369012}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 98, "C": 0.005994842503189409, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.048083285864591145, "f1": 0.6328358208955224, "f1_std": 0.09401663382734154, "bacc": 0.6177419354838709, "bacc_std": 0.07778912842782644}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 99, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.0640586249934384, "f1": 0.6232247284878863, "f1_std": 0.08839824034097298, "bacc": 0.6193548387096774, "bacc_std": 0.08582491015383371}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 100, "C": 0.046415888336127774, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.0563032496317555, "f1": 0.5176470588235295, "f1_std": 0.07654181812916153, "bacc": 0.5193548387096775, "bacc_std": 0.06803311085594083}
+eval results (random splits):
+
+| model    | repr   | clf      | dataset       | split   |   n_trials |      C |   C_std |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:--------------|:--------|-----------:|-------:|--------:|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | adni_ad_vs_cn | train   |        100 | 164.81 |  1003.7 | 0.98672 |  0.030189 | 0.97939 | 0.048896 | 0.97455 |   0.05885  |
+| flat_mae | patch  | logistic | adni_ad_vs_cn | test    |        100 | 164.81 |  1003.7 | 0.75122 |  0.060741 | 0.63988 | 0.089256 | 0.63768 |   0.087864 |
+
+
+done! total time: 0:04:37

data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/config.yaml

@@ -0,0 +1,96 @@
+output_root: experiments/data_scaling/output
+name_prefix: eval_probe
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (hcpya_task21 patch attn)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: true
+  norm: true
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn
+model: flat_mae
+representation: patch
+classifier: attn
+dataset: hcpya_task21
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn
+remote_dir: null

data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/eval_log.json

@@ -0,0 +1 @@
+{"eval/epoch": 8, "eval/id_best": 32, "eval/lr_best": 0.0011099999999999999, "eval/wd_best": 0.05, "eval/train/loss": 0.0005137641564942896, "eval/train/acc": 1.0, "eval/train/acc_std": 0.0, "eval/train/f1": 1.0, "eval/train/f1_std": 0.0, "eval/validation/loss": 0.0740802213549614, "eval/validation/acc": 0.9818948412698413, "eval/validation/acc_std": 0.002097045764527702, "eval/validation/f1": 0.9781444390549351, "eval/validation/f1_std": 0.002774058814578552, "eval/test/loss": 0.0744621604681015, "eval/test/acc": 0.9805555555555555, "eval/test/acc_std": 0.001995141609403247, "eval/test/f1": 0.9784635221118831, "eval/test/f1_std": 0.0024018065907962607}

data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/eval_log_best.json

@@ -0,0 +1 @@
+{"eval/best/epoch": 8, "eval/best/id_best": 32, "eval/best/lr_best": 0.0011099999999999999, "eval/best/wd_best": 0.05, "eval/best/train/loss": 0.0005137641564942896, "eval/best/train/acc": 1.0, "eval/best/train/acc_std": 0.0, "eval/best/train/f1": 1.0, "eval/best/train/f1_std": 0.0, "eval/best/validation/loss": 0.0740802213549614, "eval/best/validation/acc": 0.9818948412698413, "eval/best/validation/acc_std": 0.002097045764527702, "eval/best/validation/f1": 0.9781444390549351, "eval/best/validation/f1_std": 0.002774058814578552, "eval/best/test/loss": 0.0744621604681015, "eval/best/test/acc": 0.9805555555555555, "eval/best/test/acc_std": 0.001995141609403247, "eval/best/test/f1": 0.9784635221118831, "eval/best/test/f1_std": 0.0024018065907962607}

data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/eval_log_last.json

@@ -0,0 +1 @@
+{"eval/last/epoch": 19, "eval/last/id_best": 32, "eval/last/lr_best": 0.0011099999999999999, "eval/last/wd_best": 0.05, "eval/last/train/loss": 0.00024033591034822166, "eval/last/train/acc": 1.0, "eval/last/train/acc_std": 0.0, "eval/last/train/f1": 1.0, "eval/last/train/f1_std": 0.0, "eval/last/validation/loss": 0.07630305737257004, "eval/last/validation/acc": 0.9811507936507936, "eval/last/validation/acc_std": 0.002150529316895912, "eval/last/validation/f1": 0.9772682189701773, "eval/last/validation/f1_std": 0.002873631591748449, "eval/last/test/loss": 0.07681722193956375, "eval/last/test/acc": 0.9807539682539682, "eval/last/test/acc_std": 0.00203266797658856, "eval/last/test/f1": 0.9782791448250526, "eval/last/test/f1_std": 0.0025320041470026336}

data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/eval_table.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,attn,hcpya_task21,best,8,0.0011099999999999999,0.05,32,"[3.7, 1.0]",train,0.0005137641564942896,1.0,0.0,1.0,0.0
+flat_mae,patch,attn,hcpya_task21,best,8,0.0011099999999999999,0.05,32,"[3.7, 1.0]",validation,0.0740802213549614,0.9818948412698413,0.002097045764527702,0.9781444390549351,0.002774058814578552
+flat_mae,patch,attn,hcpya_task21,best,8,0.0011099999999999999,0.05,32,"[3.7, 1.0]",test,0.0744621604681015,0.9805555555555555,0.001995141609403247,0.9784635221118831,0.0024018065907962607

data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/eval_table_best.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,attn,hcpya_task21,best,8,0.0011099999999999999,0.05,32,"[3.7, 1.0]",train,0.0005137641564942896,1.0,0.0,1.0,0.0
+flat_mae,patch,attn,hcpya_task21,best,8,0.0011099999999999999,0.05,32,"[3.7, 1.0]",validation,0.0740802213549614,0.9818948412698413,0.002097045764527702,0.9781444390549351,0.002774058814578552
+flat_mae,patch,attn,hcpya_task21,best,8,0.0011099999999999999,0.05,32,"[3.7, 1.0]",test,0.0744621604681015,0.9805555555555555,0.001995141609403247,0.9784635221118831,0.0024018065907962607

data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/eval_table_last.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,attn,hcpya_task21,last,19,0.0011099999999999999,0.05,32,"[3.7, 1.0]",train,0.00024033591034822166,1.0,0.0,1.0,0.0
+flat_mae,patch,attn,hcpya_task21,last,19,0.0011099999999999999,0.05,32,"[3.7, 1.0]",validation,0.07630305737257004,0.9811507936507936,0.002150529316895912,0.9772682189701773,0.002873631591748449
+flat_mae,patch,attn,hcpya_task21,last,19,0.0011099999999999999,0.05,32,"[3.7, 1.0]",test,0.07681722193956375,0.9807539682539682,0.00203266797658856,0.9782791448250526,0.0025320041470026336

data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/log.txt

@@ -0,0 +1,885 @@
+fMRI foundation model probe eval
+version: 0.1.dev65+g4003a1397
+sha: 6c01b606db98add5848cecd23e5d599250c0bf86, status: clean, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-02-24 19:38:29
+config:
+output_root: experiments/data_scaling/output
+name_prefix: eval_probe
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (hcpya_task21 patch attn)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: true
+  norm: true
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn
+model: flat_mae
+representation: patch
+classifier: attn
+dataset: hcpya_task21
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=True, reg_tokens=0, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: hcpya_task21 (flat)
+train (n=18999):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'task', 'cond', 'cond_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 18999
+}),
+    labels=[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20],
+    counts=[ 832 1248 3201 1660  832  832  832  832  832 1248 1247 1243  832  416
+  416  416  416  416  416  416  416]
+)
+
+validation (n=4032):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'task', 'cond', 'cond_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 4032
+}),
+    labels=[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20],
+    counts=[176 264 688 352 176 176 176 176 176 264 264 264 176  88  88  88  88  88
+  88  88  88]
+)
+
+test (n=5040):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'task', 'cond', 'cond_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 5040
+}),
+    labels=[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20],
+    counts=[220 330 860 440 220 220 220 220 220 330 330 330 220 110 110 110 110 110
+ 110 110 110]
+)
+
+running backbone on example batch to get embedding dim
+embedding feature dim (patch): 768
+initializing sweep of classifier heads
+classifiers:
+ModuleList(
+  (0-48): 49 x AttnPoolClassifier(
+    (kv): Linear(in_features=768, out_features=1536, bias=True)
+    (linear): Linear(in_features=768, out_features=21, bias=True)
+  )
+)
+classifier params (train): 58.7M (58.7M)
+setting up optimizer
+total batch size: 128 = 64 bs per gpu x 2 accum
+lr: 3.00e-04
+full schedule: epochs = 20 (steps = 4000) (decay = True)
+warmup: epochs = 5 (steps = 1000)
+start training for 20 epochs
+train: [0]  [  0/400]  eta: 0:20:57  lr: nan  time: 3.1436  data: 2.6849  max mem: 21740
+train: [0]  [ 20/400]  eta: 0:03:32  lr: 0.000003  loss: 3.0531 (3.0542)  grad: 0.3047 (0.3217)  time: 0.4308  data: 0.0034  max mem: 22446
+train: [0]  [ 40/400]  eta: 0:02:58  lr: 0.000006  loss: 3.0230 (3.0085)  grad: 0.3167 (0.3213)  time: 0.4294  data: 0.0039  max mem: 22446
+train: [0]  [ 60/400]  eta: 0:02:42  lr: 0.000009  loss: 2.8746 (2.9508)  grad: 0.3167 (0.3154)  time: 0.4358  data: 0.0043  max mem: 22446
+train: [0]  [ 80/400]  eta: 0:02:29  lr: 0.000012  loss: 2.7749 (2.8922)  grad: 0.2926 (0.3064)  time: 0.4344  data: 0.0044  max mem: 22446
+train: [0]  [100/400]  eta: 0:02:18  lr: 0.000015  loss: 2.6184 (2.8263)  grad: 0.2714 (0.3010)  time: 0.4396  data: 0.0042  max mem: 22446
+train: [0]  [120/400]  eta: 0:02:07  lr: 0.000018  loss: 2.5304 (2.7613)  grad: 0.2760 (0.2954)  time: 0.4374  data: 0.0042  max mem: 22446
+train: [0]  [140/400]  eta: 0:01:59  lr: 0.000021  loss: 2.3794 (2.6989)  grad: 0.2716 (0.2932)  time: 0.4887  data: 0.0046  max mem: 22446
+train: [0]  [160/400]  eta: 0:01:50  lr: 0.000024  loss: 2.2654 (2.6431)  grad: 0.2602 (0.2874)  time: 0.4487  data: 0.0045  max mem: 22446
+train: [0]  [180/400]  eta: 0:01:40  lr: 0.000027  loss: 2.1755 (2.5867)  grad: 0.2373 (0.2824)  time: 0.4353  data: 0.0043  max mem: 22446
+train: [0]  [200/400]  eta: 0:01:31  lr: 0.000030  loss: 2.1218 (2.5337)  grad: 0.2471 (0.2787)  time: 0.4514  data: 0.0044  max mem: 22446
+train: [0]  [220/400]  eta: 0:01:21  lr: 0.000033  loss: 2.0214 (2.4838)  grad: 0.2284 (0.2744)  time: 0.4390  data: 0.0043  max mem: 22446
+train: [0]  [240/400]  eta: 0:01:12  lr: 0.000036  loss: 1.9066 (2.4323)  grad: 0.2362 (0.2718)  time: 0.4459  data: 0.0041  max mem: 22446
+train: [0]  [260/400]  eta: 0:01:03  lr: 0.000039  loss: 1.8422 (2.3855)  grad: 0.2378 (0.2693)  time: 0.4764  data: 0.0044  max mem: 22446
+train: [0]  [280/400]  eta: 0:00:54  lr: 0.000042  loss: 1.8158 (2.3448)  grad: 0.2212 (0.2655)  time: 0.4587  data: 0.0044  max mem: 22446
+train: [0]  [300/400]  eta: 0:00:46  lr: 0.000045  loss: 1.7632 (2.3043)  grad: 0.2057 (0.2617)  time: 0.5915  data: 0.1594  max mem: 22446
+train: [0]  [320/400]  eta: 0:00:37  lr: 0.000048  loss: 1.7180 (2.2654)  grad: 0.2026 (0.2585)  time: 0.4490  data: 0.0031  max mem: 22446
+train: [0]  [340/400]  eta: 0:00:27  lr: 0.000051  loss: 1.6443 (2.2281)  grad: 0.2148 (0.2563)  time: 0.4376  data: 0.0041  max mem: 22446
+train: [0]  [360/400]  eta: 0:00:18  lr: 0.000054  loss: 1.6321 (2.1939)  grad: 0.2140 (0.2536)  time: 0.4456  data: 0.0041  max mem: 22446
+train: [0]  [380/400]  eta: 0:00:09  lr: 0.000057  loss: 1.5819 (2.1609)  grad: 0.2014 (0.2509)  time: 0.4398  data: 0.0042  max mem: 22446
+train: [0]  [399/400]  eta: 0:00:00  lr: 0.000060  loss: 1.5303 (2.1277)  grad: 0.2038 (0.2488)  time: 0.4358  data: 0.0041  max mem: 22446
+train: [0] Total time: 0:03:03 (0.4598 s / it)
+train: [0] Summary: lr: 0.000060  loss: 1.5303 (2.1277)  grad: 0.2038 (0.2488)
+eval (validation): [0]  [ 0/63]  eta: 0:03:12    time: 3.0570  data: 2.7695  max mem: 22446
+eval (validation): [0]  [20/63]  eta: 0:00:20    time: 0.3361  data: 0.0037  max mem: 22446
+eval (validation): [0]  [40/63]  eta: 0:00:09    time: 0.3327  data: 0.0033  max mem: 22446
+eval (validation): [0]  [60/63]  eta: 0:00:01    time: 0.3138  data: 0.0033  max mem: 22446
+eval (validation): [0]  [62/63]  eta: 0:00:00    time: 0.3125  data: 0.0033  max mem: 22446
+eval (validation): [0] Total time: 0:00:23 (0.3748 s / it)
+cv: [0]  best hparam: (36, 1.0) (046) ('046_lr3.6e+01_wd1.0e+00')  loss: 0.099  acc: 0.967  f1: 0.960
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-best.pth
+train: [1]  [  0/400]  eta: 0:20:47  lr: nan  time: 3.1191  data: 2.7742  max mem: 22446
+train: [1]  [ 20/400]  eta: 0:03:37  lr: 0.000063  loss: 1.4729 (1.4830)  grad: 0.1973 (0.1991)  time: 0.4437  data: 0.0038  max mem: 22446
+train: [1]  [ 40/400]  eta: 0:03:01  lr: 0.000066  loss: 1.4703 (1.4600)  grad: 0.1973 (0.1986)  time: 0.4342  data: 0.0035  max mem: 22446
+train: [1]  [ 60/400]  eta: 0:02:44  lr: 0.000069  loss: 1.4114 (1.4368)  grad: 0.1922 (0.1955)  time: 0.4375  data: 0.0042  max mem: 22446
+train: [1]  [ 80/400]  eta: 0:02:30  lr: 0.000072  loss: 1.3820 (1.4217)  grad: 0.1880 (0.1951)  time: 0.4394  data: 0.0041  max mem: 22446
+train: [1]  [100/400]  eta: 0:02:19  lr: 0.000075  loss: 1.3580 (1.4107)  grad: 0.1924 (0.1949)  time: 0.4351  data: 0.0040  max mem: 22446
+train: [1]  [120/400]  eta: 0:02:08  lr: 0.000078  loss: 1.3267 (1.3920)  grad: 0.1829 (0.1935)  time: 0.4414  data: 0.0040  max mem: 22446
+train: [1]  [140/400]  eta: 0:01:59  lr: 0.000081  loss: 1.2830 (1.3784)  grad: 0.1787 (0.1917)  time: 0.4645  data: 0.0045  max mem: 22446
+train: [1]  [160/400]  eta: 0:01:49  lr: 0.000084  loss: 1.2620 (1.3614)  grad: 0.1761 (0.1902)  time: 0.4343  data: 0.0043  max mem: 22446
+train: [1]  [180/400]  eta: 0:01:40  lr: 0.000087  loss: 1.2333 (1.3479)  grad: 0.1774 (0.1889)  time: 0.4383  data: 0.0042  max mem: 22446
+train: [1]  [200/400]  eta: 0:01:30  lr: 0.000090  loss: 1.2006 (1.3330)  grad: 0.1712 (0.1875)  time: 0.4459  data: 0.0043  max mem: 22446
+train: [1]  [220/400]  eta: 0:01:21  lr: 0.000093  loss: 1.1805 (1.3182)  grad: 0.1735 (0.1874)  time: 0.4426  data: 0.0042  max mem: 22446
+train: [1]  [240/400]  eta: 0:01:12  lr: 0.000096  loss: 1.1629 (1.3043)  grad: 0.1782 (0.1861)  time: 0.4564  data: 0.0043  max mem: 22446
+train: [1]  [260/400]  eta: 0:01:03  lr: 0.000099  loss: 1.1477 (1.2924)  grad: 0.1699 (0.1848)  time: 0.4667  data: 0.0043  max mem: 22446
+train: [1]  [280/400]  eta: 0:00:54  lr: 0.000102  loss: 1.1322 (1.2791)  grad: 0.1676 (0.1842)  time: 0.4364  data: 0.0042  max mem: 22446
+train: [1]  [300/400]  eta: 0:00:46  lr: 0.000105  loss: 1.0939 (1.2661)  grad: 0.1666 (0.1826)  time: 0.6084  data: 0.1670  max mem: 22446
+train: [1]  [320/400]  eta: 0:00:36  lr: 0.000108  loss: 1.0816 (1.2543)  grad: 0.1603 (0.1814)  time: 0.4423  data: 0.0033  max mem: 22446
+train: [1]  [340/400]  eta: 0:00:27  lr: 0.000111  loss: 1.0647 (1.2426)  grad: 0.1597 (0.1798)  time: 0.4381  data: 0.0038  max mem: 22446
+train: [1]  [360/400]  eta: 0:00:18  lr: 0.000114  loss: 1.0543 (1.2326)  grad: 0.1533 (0.1783)  time: 0.4405  data: 0.0043  max mem: 22446
+train: [1]  [380/400]  eta: 0:00:09  lr: 0.000117  loss: 1.0288 (1.2223)  grad: 0.1554 (0.1772)  time: 0.4402  data: 0.0039  max mem: 22446
+train: [1]  [399/400]  eta: 0:00:00  lr: 0.000120  loss: 1.0181 (1.2125)  grad: 0.1588 (0.1762)  time: 0.4351  data: 0.0042  max mem: 22446
+train: [1] Total time: 0:03:03 (0.4583 s / it)
+train: [1] Summary: lr: 0.000120  loss: 1.0181 (1.2125)  grad: 0.1588 (0.1762)
+eval (validation): [1]  [ 0/63]  eta: 0:03:10    time: 3.0239  data: 2.7472  max mem: 22446
+eval (validation): [1]  [20/63]  eta: 0:00:19    time: 0.3371  data: 0.0043  max mem: 22446
+eval (validation): [1]  [40/63]  eta: 0:00:09    time: 0.3303  data: 0.0029  max mem: 22446
+eval (validation): [1]  [60/63]  eta: 0:00:01    time: 0.3165  data: 0.0034  max mem: 22446
+eval (validation): [1]  [62/63]  eta: 0:00:00    time: 0.3135  data: 0.0034  max mem: 22446
+eval (validation): [1] Total time: 0:00:23 (0.3745 s / it)
+cv: [1]  best hparam: (8.3, 1.0) (037) ('037_lr8.3e+00_wd1.0e+00')  loss: 0.088  acc: 0.973  f1: 0.968
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-best.pth
+train: [2]  [  0/400]  eta: 0:20:37  lr: nan  time: 3.0935  data: 2.7522  max mem: 22446
+train: [2]  [ 20/400]  eta: 0:03:35  lr: 0.000123  loss: 0.9465 (0.9676)  grad: 0.1614 (0.1666)  time: 0.4394  data: 0.0034  max mem: 22446
+train: [2]  [ 40/400]  eta: 0:03:00  lr: 0.000126  loss: 0.9697 (0.9741)  grad: 0.1614 (0.1702)  time: 0.4359  data: 0.0045  max mem: 22446
+train: [2]  [ 60/400]  eta: 0:02:43  lr: 0.000129  loss: 0.9697 (0.9703)  grad: 0.1695 (0.1735)  time: 0.4342  data: 0.0041  max mem: 22446
+train: [2]  [ 80/400]  eta: 0:02:30  lr: 0.000132  loss: 0.9590 (0.9698)  grad: 0.1760 (0.1754)  time: 0.4386  data: 0.0044  max mem: 22446
+train: [2]  [100/400]  eta: 0:02:19  lr: 0.000135  loss: 0.9565 (0.9664)  grad: 0.1815 (0.1782)  time: 0.4490  data: 0.0042  max mem: 22446
+train: [2]  [120/400]  eta: 0:02:10  lr: 0.000138  loss: 0.9375 (0.9636)  grad: 0.1847 (0.1821)  time: 0.4630  data: 0.0041  max mem: 22446
+train: [2]  [140/400]  eta: 0:02:00  lr: 0.000141  loss: 0.8941 (0.9528)  grad: 0.1843 (0.1815)  time: 0.4586  data: 0.0043  max mem: 22446
+train: [2]  [160/400]  eta: 0:01:50  lr: 0.000144  loss: 0.9158 (0.9536)  grad: 0.1961 (0.1843)  time: 0.4435  data: 0.0041  max mem: 22446
+train: [2]  [180/400]  eta: 0:01:41  lr: 0.000147  loss: 0.9429 (0.9492)  grad: 0.1982 (0.1854)  time: 0.4622  data: 0.0045  max mem: 22446
+train: [2]  [200/400]  eta: 0:01:32  lr: 0.000150  loss: 0.8883 (0.9412)  grad: 0.1821 (0.1850)  time: 0.4466  data: 0.0042  max mem: 22446
+train: [2]  [220/400]  eta: 0:01:22  lr: 0.000153  loss: 0.9120 (0.9445)  grad: 0.1821 (0.1869)  time: 0.4433  data: 0.0043  max mem: 22446
+train: [2]  [240/400]  eta: 0:01:13  lr: 0.000156  loss: 0.9036 (0.9387)  grad: 0.1870 (0.1873)  time: 0.4735  data: 0.0044  max mem: 22446
+train: [2]  [260/400]  eta: 0:01:04  lr: 0.000159  loss: 0.8502 (0.9353)  grad: 0.1989 (0.1882)  time: 0.4637  data: 0.0045  max mem: 22446
+train: [2]  [280/400]  eta: 0:00:55  lr: 0.000162  loss: 0.8502 (0.9312)  grad: 0.1991 (0.1893)  time: 0.4351  data: 0.0042  max mem: 22446
+train: [2]  [300/400]  eta: 0:00:46  lr: 0.000165  loss: 0.8414 (0.9253)  grad: 0.1986 (0.1907)  time: 0.6139  data: 0.1762  max mem: 22446
+train: [2]  [320/400]  eta: 0:00:37  lr: 0.000168  loss: 0.8750 (0.9226)  grad: 0.2121 (0.1924)  time: 0.4396  data: 0.0032  max mem: 22446
+train: [2]  [340/400]  eta: 0:00:27  lr: 0.000171  loss: 0.8643 (0.9180)  grad: 0.2064 (0.1932)  time: 0.4370  data: 0.0041  max mem: 22446
+train: [2]  [360/400]  eta: 0:00:18  lr: 0.000174  loss: 0.8691 (0.9173)  grad: 0.2054 (0.1953)  time: 0.4330  data: 0.0040  max mem: 22446
+train: [2]  [380/400]  eta: 0:00:09  lr: 0.000177  loss: 0.8440 (0.9115)  grad: 0.2232 (0.1975)  time: 0.4351  data: 0.0036  max mem: 22446
+train: [2]  [399/400]  eta: 0:00:00  lr: 0.000180  loss: 0.7762 (0.9034)  grad: 0.2134 (0.1983)  time: 0.4381  data: 0.0042  max mem: 22446
+train: [2] Total time: 0:03:04 (0.4614 s / it)
+train: [2] Summary: lr: 0.000180  loss: 0.7762 (0.9034)  grad: 0.2134 (0.1983)
+eval (validation): [2]  [ 0/63]  eta: 0:03:07    time: 2.9738  data: 2.7421  max mem: 22446
+eval (validation): [2]  [20/63]  eta: 0:00:21    time: 0.3697  data: 0.0043  max mem: 22446
+eval (validation): [2]  [40/63]  eta: 0:00:09    time: 0.3210  data: 0.0033  max mem: 22446
+eval (validation): [2]  [60/63]  eta: 0:00:01    time: 0.3065  data: 0.0031  max mem: 22446
+eval (validation): [2]  [62/63]  eta: 0:00:00    time: 0.3041  data: 0.0031  max mem: 22446
+eval (validation): [2] Total time: 0:00:23 (0.3778 s / it)
+cv: [2]  best hparam: (8.3, 1.0) (037) ('037_lr8.3e+00_wd1.0e+00')  loss: 0.070  acc: 0.977  f1: 0.974
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-best.pth
+train: [3]  [  0/400]  eta: 0:20:47  lr: nan  time: 3.1186  data: 2.7865  max mem: 22446
+train: [3]  [ 20/400]  eta: 0:03:30  lr: 0.000183  loss: 0.7132 (0.7295)  grad: 0.2116 (0.2214)  time: 0.4254  data: 0.0027  max mem: 22446
+train: [3]  [ 40/400]  eta: 0:02:58  lr: 0.000186  loss: 0.7585 (0.7649)  grad: 0.2116 (0.2163)  time: 0.4347  data: 0.0037  max mem: 22446
+train: [3]  [ 60/400]  eta: 0:02:42  lr: 0.000189  loss: 0.7748 (0.7669)  grad: 0.2046 (0.2177)  time: 0.4398  data: 0.0041  max mem: 22446
+train: [3]  [ 80/400]  eta: 0:02:29  lr: 0.000192  loss: 0.7406 (0.7747)  grad: 0.2104 (0.2175)  time: 0.4339  data: 0.0042  max mem: 22446
+train: [3]  [100/400]  eta: 0:02:18  lr: 0.000195  loss: 0.7518 (0.7761)  grad: 0.2225 (0.2173)  time: 0.4407  data: 0.0041  max mem: 22446
+train: [3]  [120/400]  eta: 0:02:09  lr: 0.000198  loss: 0.8042 (0.7873)  grad: 0.2325 (0.2261)  time: 0.4583  data: 0.0042  max mem: 22446
+train: [3]  [140/400]  eta: 0:01:59  lr: 0.000201  loss: 0.8078 (0.7905)  grad: 0.2490 (0.2335)  time: 0.4415  data: 0.0039  max mem: 22446
+train: [3]  [160/400]  eta: 0:01:49  lr: 0.000204  loss: 0.7841 (0.7919)  grad: 0.2644 (0.2420)  time: 0.4368  data: 0.0042  max mem: 22446
+train: [3]  [180/400]  eta: 0:01:40  lr: 0.000207  loss: 0.7948 (0.7961)  grad: 0.2694 (0.2456)  time: 0.4500  data: 0.0041  max mem: 22446
+train: [3]  [200/400]  eta: 0:01:30  lr: 0.000210  loss: 0.7948 (0.7992)  grad: 0.2694 (0.2529)  time: 0.4349  data: 0.0042  max mem: 22446
+train: [3]  [220/400]  eta: 0:01:21  lr: 0.000213  loss: 0.7764 (0.7969)  grad: 0.2772 (0.2595)  time: 0.4408  data: 0.0045  max mem: 22446
+train: [3]  [240/400]  eta: 0:01:12  lr: 0.000216  loss: 0.7941 (0.8082)  grad: 0.2945 (0.2637)  time: 0.4650  data: 0.0043  max mem: 22446
+train: [3]  [260/400]  eta: 0:01:03  lr: 0.000219  loss: 0.7279 (0.8066)  grad: 0.3013 (0.2667)  time: 0.4637  data: 0.0042  max mem: 22446
+train: [3]  [280/400]  eta: 0:00:54  lr: 0.000222  loss: 0.7771 (0.8089)  grad: 0.3013 (0.2710)  time: 0.4362  data: 0.0042  max mem: 22446
+train: [3]  [300/400]  eta: 0:00:46  lr: 0.000225  loss: 0.8356 (0.8148)  grad: 0.3274 (0.2761)  time: 0.6181  data: 0.1733  max mem: 22446
+train: [3]  [320/400]  eta: 0:00:37  lr: 0.000228  loss: 0.8388 (0.8171)  grad: 0.3286 (0.2797)  time: 0.4483  data: 0.0036  max mem: 22446
+train: [3]  [340/400]  eta: 0:00:27  lr: 0.000231  loss: 0.6930 (0.8102)  grad: 0.3257 (0.2831)  time: 0.4306  data: 0.0042  max mem: 22446
+train: [3]  [360/400]  eta: 0:00:18  lr: 0.000234  loss: 0.6930 (0.8048)  grad: 0.3299 (0.2866)  time: 0.4439  data: 0.0042  max mem: 22446
+train: [3]  [380/400]  eta: 0:00:09  lr: 0.000237  loss: 0.7603 (0.8078)  grad: 0.3635 (0.2925)  time: 0.4413  data: 0.0042  max mem: 22446
+train: [3]  [399/400]  eta: 0:00:00  lr: 0.000240  loss: 0.8656 (0.8122)  grad: 0.3866 (0.2980)  time: 0.4388  data: 0.0041  max mem: 22446
+train: [3] Total time: 0:03:03 (0.4583 s / it)
+train: [3] Summary: lr: 0.000240  loss: 0.8656 (0.8122)  grad: 0.3866 (0.2980)
+eval (validation): [3]  [ 0/63]  eta: 0:03:15    time: 3.1036  data: 2.8714  max mem: 22446
+eval (validation): [3]  [20/63]  eta: 0:00:19    time: 0.3220  data: 0.0024  max mem: 22446
+eval (validation): [3]  [40/63]  eta: 0:00:09    time: 0.3393  data: 0.0034  max mem: 22446
+eval (validation): [3]  [60/63]  eta: 0:00:01    time: 0.3187  data: 0.0033  max mem: 22446
+eval (validation): [3]  [62/63]  eta: 0:00:00    time: 0.3178  data: 0.0033  max mem: 22446
+eval (validation): [3] Total time: 0:00:23 (0.3751 s / it)
+cv: [3]  best hparam: (5.1, 1.0) (034) ('034_lr5.1e+00_wd1.0e+00')  loss: 0.073  acc: 0.976  f1: 0.971
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [4]  [  0/400]  eta: 0:20:36  lr: nan  time: 3.0907  data: 2.7527  max mem: 22446
+train: [4]  [ 20/400]  eta: 0:03:35  lr: 0.000243  loss: 0.9643 (0.9557)  grad: 0.3612 (0.4278)  time: 0.4398  data: 0.0032  max mem: 22446
+train: [4]  [ 40/400]  eta: 0:03:01  lr: 0.000246  loss: 0.9889 (0.9976)  grad: 0.4106 (0.4343)  time: 0.4394  data: 0.0038  max mem: 22446
+train: [4]  [ 60/400]  eta: 0:02:44  lr: 0.000249  loss: 0.9760 (0.9690)  grad: 0.4364 (0.4570)  time: 0.4421  data: 0.0039  max mem: 22446
+train: [4]  [ 80/400]  eta: 0:02:31  lr: 0.000252  loss: 0.9021 (0.9447)  grad: 0.4579 (0.4487)  time: 0.4397  data: 0.0042  max mem: 22446
+train: [4]  [100/400]  eta: 0:02:20  lr: 0.000255  loss: 0.9147 (0.9351)  grad: 0.4264 (0.4465)  time: 0.4457  data: 0.0041  max mem: 22446
+train: [4]  [120/400]  eta: 0:02:11  lr: 0.000258  loss: 0.9205 (0.9408)  grad: 0.4632 (0.4744)  time: 0.4744  data: 0.0043  max mem: 22446
+train: [4]  [140/400]  eta: 0:02:00  lr: 0.000261  loss: 0.8545 (0.9356)  grad: 0.4770 (0.4776)  time: 0.4327  data: 0.0042  max mem: 22446
+train: [4]  [160/400]  eta: 0:01:50  lr: 0.000264  loss: 0.9504 (0.9578)  grad: 0.5102 (0.4861)  time: 0.4500  data: 0.0040  max mem: 22446
+train: [4]  [180/400]  eta: 0:01:41  lr: 0.000267  loss: 0.9934 (0.9550)  grad: 0.4518 (0.4781)  time: 0.4382  data: 0.0041  max mem: 22446
+train: [4]  [200/400]  eta: 0:01:31  lr: 0.000270  loss: 0.9215 (0.9518)  grad: 0.4407 (0.4849)  time: 0.4323  data: 0.0041  max mem: 22446
+train: [4]  [220/400]  eta: 0:01:21  lr: 0.000273  loss: 0.9215 (0.9583)  grad: 0.5042 (0.4902)  time: 0.4331  data: 0.0042  max mem: 22446
+train: [4]  [240/400]  eta: 0:01:12  lr: 0.000276  loss: 0.9176 (0.9535)  grad: 0.5135 (0.4932)  time: 0.4741  data: 0.0040  max mem: 22446
+train: [4]  [260/400]  eta: 0:01:03  lr: 0.000279  loss: 1.0125 (0.9826)  grad: 0.5466 (0.5008)  time: 0.4421  data: 0.0040  max mem: 22446
+train: [4]  [280/400]  eta: 0:00:54  lr: 0.000282  loss: 1.2209 (0.9997)  grad: 0.5545 (0.5068)  time: 0.4292  data: 0.0038  max mem: 22446
+train: [4]  [300/400]  eta: 0:00:46  lr: 0.000285  loss: 1.0280 (1.0037)  grad: 0.5581 (0.5111)  time: 0.6330  data: 0.1704  max mem: 22446
+train: [4]  [320/400]  eta: 0:00:37  lr: 0.000288  loss: 0.8179 (0.9939)  grad: 0.5613 (0.5185)  time: 0.4400  data: 0.0033  max mem: 22446
+train: [4]  [340/400]  eta: 0:00:27  lr: 0.000291  loss: 0.7244 (0.9830)  grad: 0.5613 (0.5179)  time: 0.4454  data: 0.0042  max mem: 22446
+train: [4]  [360/400]  eta: 0:00:18  lr: 0.000294  loss: 0.7919 (0.9891)  grad: 0.5021 (0.5196)  time: 0.4447  data: 0.0042  max mem: 22446
+train: [4]  [380/400]  eta: 0:00:09  lr: 0.000297  loss: 1.0565 (1.0032)  grad: 0.5789 (0.5284)  time: 0.4385  data: 0.0042  max mem: 22446
+train: [4]  [399/400]  eta: 0:00:00  lr: 0.000300  loss: 1.0951 (1.0070)  grad: 0.5575 (0.5294)  time: 0.4397  data: 0.0042  max mem: 22446
+train: [4] Total time: 0:03:03 (0.4599 s / it)
+train: [4] Summary: lr: 0.000300  loss: 1.0951 (1.0070)  grad: 0.5575 (0.5294)
+eval (validation): [4]  [ 0/63]  eta: 0:03:11    time: 3.0468  data: 2.8075  max mem: 22446
+eval (validation): [4]  [20/63]  eta: 0:00:20    time: 0.3444  data: 0.0037  max mem: 22446
+eval (validation): [4]  [40/63]  eta: 0:00:09    time: 0.3175  data: 0.0028  max mem: 22446
+eval (validation): [4]  [60/63]  eta: 0:00:01    time: 0.3092  data: 0.0032  max mem: 22446
+eval (validation): [4]  [62/63]  eta: 0:00:00    time: 0.3077  data: 0.0032  max mem: 22446
+eval (validation): [4] Total time: 0:00:23 (0.3707 s / it)
+cv: [4]  best hparam: (3.1, 1.0) (031) ('031_lr3.1e+00_wd1.0e+00')  loss: 0.077  acc: 0.977  f1: 0.974
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [5]  [  0/400]  eta: 0:21:37  lr: nan  time: 3.2436  data: 2.8674  max mem: 22446
+train: [5]  [ 20/400]  eta: 0:03:37  lr: 0.000300  loss: 0.8394 (0.9301)  grad: 0.4893 (0.5218)  time: 0.4382  data: 0.0027  max mem: 22446
+train: [5]  [ 40/400]  eta: 0:03:03  lr: 0.000300  loss: 0.8472 (0.9751)  grad: 0.4893 (0.5056)  time: 0.4429  data: 0.0041  max mem: 22446
+train: [5]  [ 60/400]  eta: 0:02:45  lr: 0.000300  loss: 0.9470 (0.9662)  grad: 0.5366 (0.5536)  time: 0.4432  data: 0.0040  max mem: 22446
+train: [5]  [ 80/400]  eta: 0:02:32  lr: 0.000300  loss: 0.9775 (0.9922)  grad: 0.6238 (0.5771)  time: 0.4478  data: 0.0044  max mem: 22446
+train: [5]  [100/400]  eta: 0:02:21  lr: 0.000300  loss: 1.0824 (1.0281)  grad: 0.5746 (0.5785)  time: 0.4453  data: 0.0042  max mem: 22446
+train: [5]  [120/400]  eta: 0:02:11  lr: 0.000300  loss: 1.1668 (1.0659)  grad: 0.5726 (0.5778)  time: 0.4694  data: 0.0043  max mem: 22446
+train: [5]  [140/400]  eta: 0:02:01  lr: 0.000300  loss: 1.2481 (1.0868)  grad: 0.6159 (0.5948)  time: 0.4425  data: 0.0040  max mem: 22446
+train: [5]  [160/400]  eta: 0:01:51  lr: 0.000299  loss: 1.0626 (1.1083)  grad: 0.6435 (0.5969)  time: 0.4599  data: 0.0041  max mem: 22446
+train: [5]  [180/400]  eta: 0:01:42  lr: 0.000299  loss: 0.9686 (1.1057)  grad: 0.6140 (0.5999)  time: 0.4467  data: 0.0042  max mem: 22446
+train: [5]  [200/400]  eta: 0:01:32  lr: 0.000299  loss: 1.1373 (1.1241)  grad: 0.6351 (0.6059)  time: 0.4433  data: 0.0042  max mem: 22446
+train: [5]  [220/400]  eta: 0:01:22  lr: 0.000299  loss: 1.1373 (1.1392)  grad: 0.6351 (0.6059)  time: 0.4296  data: 0.0043  max mem: 22446
+train: [5]  [240/400]  eta: 0:01:13  lr: 0.000299  loss: 1.0353 (1.1387)  grad: 0.6323 (0.6106)  time: 0.4678  data: 0.0044  max mem: 22446
+train: [5]  [260/400]  eta: 0:01:04  lr: 0.000299  loss: 1.2034 (1.1690)  grad: 0.7213 (0.6246)  time: 0.4518  data: 0.0043  max mem: 22446
+train: [5]  [280/400]  eta: 0:00:54  lr: 0.000298  loss: 1.3694 (1.1915)  grad: 0.7522 (0.6305)  time: 0.4309  data: 0.0040  max mem: 22446
+train: [5]  [300/400]  eta: 0:00:46  lr: 0.000298  loss: 1.2287 (1.1863)  grad: 0.6535 (0.6330)  time: 0.6135  data: 0.1726  max mem: 22446
+train: [5]  [320/400]  eta: 0:00:37  lr: 0.000298  loss: 0.8613 (1.1634)  grad: 0.5854 (0.6275)  time: 0.4461  data: 0.0033  max mem: 22446
+train: [5]  [340/400]  eta: 0:00:27  lr: 0.000298  loss: 0.8847 (1.1620)  grad: 0.5679 (0.6270)  time: 0.4312  data: 0.0043  max mem: 22446
+train: [5]  [360/400]  eta: 0:00:18  lr: 0.000297  loss: 0.9092 (1.1522)  grad: 0.5630 (0.6229)  time: 0.4370  data: 0.0043  max mem: 22446
+train: [5]  [380/400]  eta: 0:00:09  lr: 0.000297  loss: 0.9830 (1.1528)  grad: 0.5630 (0.6241)  time: 0.4346  data: 0.0039  max mem: 22446
+train: [5]  [399/400]  eta: 0:00:00  lr: 0.000297  loss: 1.0659 (1.1418)  grad: 0.5790 (0.6204)  time: 0.4393  data: 0.0036  max mem: 22446
+train: [5] Total time: 0:03:04 (0.4606 s / it)
+train: [5] Summary: lr: 0.000297  loss: 1.0659 (1.1418)  grad: 0.5790 (0.6204)
+eval (validation): [5]  [ 0/63]  eta: 0:03:17    time: 3.1284  data: 2.9031  max mem: 22446
+eval (validation): [5]  [20/63]  eta: 0:00:20    time: 0.3338  data: 0.0031  max mem: 22446
+eval (validation): [5]  [40/63]  eta: 0:00:09    time: 0.3261  data: 0.0028  max mem: 22446
+eval (validation): [5]  [60/63]  eta: 0:00:01    time: 0.3220  data: 0.0033  max mem: 22446
+eval (validation): [5]  [62/63]  eta: 0:00:00    time: 0.3193  data: 0.0033  max mem: 22446
+eval (validation): [5] Total time: 0:00:23 (0.3759 s / it)
+cv: [5]  best hparam: (3.1, 1.0) (031) ('031_lr3.1e+00_wd1.0e+00')  loss: 0.070  acc: 0.978  f1: 0.974
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-best.pth
+train: [6]  [  0/400]  eta: 0:20:33  lr: nan  time: 3.0846  data: 2.7545  max mem: 22446
+train: [6]  [ 20/400]  eta: 0:03:31  lr: 0.000296  loss: 0.8134 (0.8569)  grad: 0.5114 (0.5680)  time: 0.4301  data: 0.0039  max mem: 22446
+train: [6]  [ 40/400]  eta: 0:03:00  lr: 0.000296  loss: 0.8033 (0.8525)  grad: 0.5087 (0.5540)  time: 0.4430  data: 0.0037  max mem: 22446
+train: [6]  [ 60/400]  eta: 0:02:43  lr: 0.000296  loss: 0.7791 (0.8867)  grad: 0.5128 (0.5885)  time: 0.4432  data: 0.0041  max mem: 22446
+train: [6]  [ 80/400]  eta: 0:02:32  lr: 0.000295  loss: 0.8987 (0.8965)  grad: 0.6187 (0.5868)  time: 0.4584  data: 0.0044  max mem: 22446
+train: [6]  [100/400]  eta: 0:02:21  lr: 0.000295  loss: 0.8192 (0.8703)  grad: 0.5400 (0.5738)  time: 0.4447  data: 0.0042  max mem: 22446
+train: [6]  [120/400]  eta: 0:02:11  lr: 0.000295  loss: 0.6692 (0.8490)  grad: 0.5304 (0.5647)  time: 0.4636  data: 0.0042  max mem: 22446
+train: [6]  [140/400]  eta: 0:02:00  lr: 0.000294  loss: 0.6475 (0.8390)  grad: 0.5447 (0.5648)  time: 0.4431  data: 0.0040  max mem: 22446
+train: [6]  [160/400]  eta: 0:01:51  lr: 0.000294  loss: 0.6030 (0.8283)  grad: 0.4871 (0.5559)  time: 0.4491  data: 0.0046  max mem: 22446
+train: [6]  [180/400]  eta: 0:01:41  lr: 0.000293  loss: 0.6583 (0.8380)  grad: 0.4832 (0.5518)  time: 0.4359  data: 0.0034  max mem: 22446
+train: [6]  [200/400]  eta: 0:01:31  lr: 0.000293  loss: 0.8725 (0.8674)  grad: 0.5034 (0.5508)  time: 0.4470  data: 0.0039  max mem: 22446
+train: [6]  [220/400]  eta: 0:01:22  lr: 0.000292  loss: 0.8235 (0.8633)  grad: 0.5522 (0.5474)  time: 0.4482  data: 0.0041  max mem: 22446
+train: [6]  [240/400]  eta: 0:01:13  lr: 0.000292  loss: 0.7251 (0.8497)  grad: 0.5192 (0.5437)  time: 0.4719  data: 0.0045  max mem: 22446
+train: [6]  [260/400]  eta: 0:01:04  lr: 0.000291  loss: 0.6410 (0.8366)  grad: 0.5154 (0.5443)  time: 0.4520  data: 0.0042  max mem: 22446
+train: [6]  [280/400]  eta: 0:00:54  lr: 0.000291  loss: 0.7245 (0.8414)  grad: 0.5154 (0.5411)  time: 0.4464  data: 0.0041  max mem: 22446
+train: [6]  [300/400]  eta: 0:00:46  lr: 0.000290  loss: 0.7194 (0.8323)  grad: 0.4744 (0.5382)  time: 0.6071  data: 0.1746  max mem: 22446
+train: [6]  [320/400]  eta: 0:00:37  lr: 0.000290  loss: 0.6389 (0.8201)  grad: 0.4121 (0.5319)  time: 0.4353  data: 0.0034  max mem: 22446
+train: [6]  [340/400]  eta: 0:00:27  lr: 0.000289  loss: 0.5418 (0.8109)  grad: 0.4121 (0.5262)  time: 0.4354  data: 0.0042  max mem: 22446
+train: [6]  [360/400]  eta: 0:00:18  lr: 0.000288  loss: 0.4957 (0.7976)  grad: 0.4108 (0.5195)  time: 0.4423  data: 0.0044  max mem: 22446
+train: [6]  [380/400]  eta: 0:00:09  lr: 0.000288  loss: 0.4772 (0.7841)  grad: 0.4160 (0.5147)  time: 0.4450  data: 0.0044  max mem: 22446
+train: [6]  [399/400]  eta: 0:00:00  lr: 0.000287  loss: 0.4356 (0.7659)  grad: 0.3698 (0.5070)  time: 0.4485  data: 0.0043  max mem: 22446
+train: [6] Total time: 0:03:04 (0.4615 s / it)
+train: [6] Summary: lr: 0.000287  loss: 0.4356 (0.7659)  grad: 0.3698 (0.5070)
+eval (validation): [6]  [ 0/63]  eta: 0:03:18    time: 3.1517  data: 2.8718  max mem: 22446
+eval (validation): [6]  [20/63]  eta: 0:00:20    time: 0.3468  data: 0.0024  max mem: 22446
+eval (validation): [6]  [40/63]  eta: 0:00:09    time: 0.3243  data: 0.0034  max mem: 22446
+eval (validation): [6]  [60/63]  eta: 0:00:01    time: 0.3166  data: 0.0035  max mem: 22446
+eval (validation): [6]  [62/63]  eta: 0:00:00    time: 0.3152  data: 0.0035  max mem: 22446
+eval (validation): [6] Total time: 0:00:23 (0.3780 s / it)
+cv: [6]  best hparam: (2.3, 1.0) (029) ('029_lr2.3e+00_wd1.0e+00')  loss: 0.065  acc: 0.980  f1: 0.977
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-best.pth
+train: [7]  [  0/400]  eta: 0:21:01  lr: nan  time: 3.1537  data: 2.7809  max mem: 22446
+train: [7]  [ 20/400]  eta: 0:03:39  lr: 0.000286  loss: 0.4046 (0.5245)  grad: 0.3406 (0.3528)  time: 0.4476  data: 0.0030  max mem: 22446
+train: [7]  [ 40/400]  eta: 0:03:03  lr: 0.000286  loss: 0.4046 (0.4939)  grad: 0.3406 (0.3568)  time: 0.4411  data: 0.0039  max mem: 22446
+train: [7]  [ 60/400]  eta: 0:02:46  lr: 0.000285  loss: 0.4689 (0.5311)  grad: 0.3761 (0.3762)  time: 0.4514  data: 0.0042  max mem: 22446
+train: [7]  [ 80/400]  eta: 0:02:32  lr: 0.000284  loss: 0.4730 (0.5142)  grad: 0.3775 (0.3720)  time: 0.4334  data: 0.0041  max mem: 22446
+train: [7]  [100/400]  eta: 0:02:22  lr: 0.000284  loss: 0.4220 (0.5158)  grad: 0.3775 (0.3783)  time: 0.4636  data: 0.0044  max mem: 22446
+train: [7]  [120/400]  eta: 0:02:11  lr: 0.000283  loss: 0.3948 (0.5082)  grad: 0.3989 (0.3757)  time: 0.4430  data: 0.0042  max mem: 22446
+train: [7]  [140/400]  eta: 0:02:01  lr: 0.000282  loss: 0.4676 (0.5142)  grad: 0.3946 (0.3794)  time: 0.4526  data: 0.0043  max mem: 22446
+train: [7]  [160/400]  eta: 0:01:51  lr: 0.000282  loss: 0.4756 (0.5068)  grad: 0.3946 (0.3787)  time: 0.4405  data: 0.0044  max mem: 22446
+train: [7]  [180/400]  eta: 0:01:41  lr: 0.000281  loss: 0.4595 (0.5145)  grad: 0.3751 (0.3801)  time: 0.4367  data: 0.0041  max mem: 22446
+train: [7]  [200/400]  eta: 0:01:31  lr: 0.000280  loss: 0.3448 (0.5139)  grad: 0.3843 (0.3818)  time: 0.4356  data: 0.0041  max mem: 22446
+train: [7]  [220/400]  eta: 0:01:22  lr: 0.000279  loss: 0.4909 (0.5142)  grad: 0.3783 (0.3796)  time: 0.4629  data: 0.0041  max mem: 22446
+train: [7]  [240/400]  eta: 0:01:13  lr: 0.000278  loss: 0.4909 (0.5150)  grad: 0.3696 (0.3799)  time: 0.4554  data: 0.0043  max mem: 22446
+train: [7]  [260/400]  eta: 0:01:03  lr: 0.000278  loss: 0.4530 (0.5078)  grad: 0.3834 (0.3793)  time: 0.4342  data: 0.0039  max mem: 22446
+train: [7]  [280/400]  eta: 0:00:54  lr: 0.000277  loss: 0.4537 (0.5214)  grad: 0.3834 (0.3793)  time: 0.4645  data: 0.0043  max mem: 22446
+train: [7]  [300/400]  eta: 0:00:46  lr: 0.000276  loss: 0.5993 (0.5278)  grad: 0.4108 (0.3802)  time: 0.5942  data: 0.1643  max mem: 22446
+train: [7]  [320/400]  eta: 0:00:37  lr: 0.000275  loss: 0.4802 (0.5301)  grad: 0.3752 (0.3791)  time: 0.4357  data: 0.0032  max mem: 22446
+train: [7]  [340/400]  eta: 0:00:27  lr: 0.000274  loss: 0.4218 (0.5237)  grad: 0.3567 (0.3797)  time: 0.4389  data: 0.0040  max mem: 22446
+train: [7]  [360/400]  eta: 0:00:18  lr: 0.000273  loss: 0.3742 (0.5192)  grad: 0.3287 (0.3779)  time: 0.4369  data: 0.0043  max mem: 22446
+train: [7]  [380/400]  eta: 0:00:09  lr: 0.000272  loss: 0.3505 (0.5087)  grad: 0.3232 (0.3764)  time: 0.4352  data: 0.0040  max mem: 22446
+train: [7]  [399/400]  eta: 0:00:00  lr: 0.000271  loss: 0.3413 (0.5029)  grad: 0.3221 (0.3747)  time: 0.4374  data: 0.0040  max mem: 22446
+train: [7] Total time: 0:03:03 (0.4594 s / it)
+train: [7] Summary: lr: 0.000271  loss: 0.3413 (0.5029)  grad: 0.3221 (0.3747)
+eval (validation): [7]  [ 0/63]  eta: 0:03:10    time: 3.0294  data: 2.7579  max mem: 22446
+eval (validation): [7]  [20/63]  eta: 0:00:19    time: 0.3187  data: 0.0029  max mem: 22446
+eval (validation): [7]  [40/63]  eta: 0:00:08    time: 0.3296  data: 0.0031  max mem: 22446
+eval (validation): [7]  [60/63]  eta: 0:00:01    time: 0.3140  data: 0.0035  max mem: 22446
+eval (validation): [7]  [62/63]  eta: 0:00:00    time: 0.3093  data: 0.0034  max mem: 22446
+eval (validation): [7] Total time: 0:00:23 (0.3676 s / it)
+cv: [7]  best hparam: (2.7, 1.0) (030) ('030_lr2.7e+00_wd1.0e+00')  loss: 0.067  acc: 0.980  f1: 0.977
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [8]  [  0/400]  eta: 0:20:53  lr: nan  time: 3.1349  data: 2.7415  max mem: 22446
+train: [8]  [ 20/400]  eta: 0:03:36  lr: 0.000270  loss: 0.2602 (0.3000)  grad: 0.2681 (0.2795)  time: 0.4402  data: 0.0027  max mem: 22446
+train: [8]  [ 40/400]  eta: 0:03:01  lr: 0.000270  loss: 0.2972 (0.3357)  grad: 0.2707 (0.2790)  time: 0.4381  data: 0.0040  max mem: 22446
+train: [8]  [ 60/400]  eta: 0:02:46  lr: 0.000269  loss: 0.3205 (0.3371)  grad: 0.2707 (0.2852)  time: 0.4552  data: 0.0043  max mem: 22446
+train: [8]  [ 80/400]  eta: 0:02:32  lr: 0.000268  loss: 0.3052 (0.3481)  grad: 0.2970 (0.2892)  time: 0.4429  data: 0.0038  max mem: 22446
+train: [8]  [100/400]  eta: 0:02:21  lr: 0.000267  loss: 0.3052 (0.3427)  grad: 0.2970 (0.2904)  time: 0.4557  data: 0.0041  max mem: 22446
+train: [8]  [120/400]  eta: 0:02:10  lr: 0.000266  loss: 0.3250 (0.3673)  grad: 0.3130 (0.2985)  time: 0.4346  data: 0.0041  max mem: 22446
+train: [8]  [140/400]  eta: 0:02:00  lr: 0.000265  loss: 0.3672 (0.3732)  grad: 0.3130 (0.3022)  time: 0.4512  data: 0.0043  max mem: 22446
+train: [8]  [160/400]  eta: 0:01:50  lr: 0.000264  loss: 0.3518 (0.3763)  grad: 0.2954 (0.3017)  time: 0.4370  data: 0.0042  max mem: 22446
+train: [8]  [180/400]  eta: 0:01:40  lr: 0.000263  loss: 0.3080 (0.3718)  grad: 0.2715 (0.2995)  time: 0.4380  data: 0.0042  max mem: 22446
+train: [8]  [200/400]  eta: 0:01:31  lr: 0.000262  loss: 0.2857 (0.3648)  grad: 0.2660 (0.2964)  time: 0.4368  data: 0.0040  max mem: 22446
+train: [8]  [220/400]  eta: 0:01:22  lr: 0.000260  loss: 0.2857 (0.3650)  grad: 0.2791 (0.2964)  time: 0.4557  data: 0.0041  max mem: 22446
+train: [8]  [240/400]  eta: 0:01:13  lr: 0.000259  loss: 0.3381 (0.3648)  grad: 0.2951 (0.2956)  time: 0.4719  data: 0.0046  max mem: 22446
+train: [8]  [260/400]  eta: 0:01:03  lr: 0.000258  loss: 0.3237 (0.3667)  grad: 0.2804 (0.2955)  time: 0.4406  data: 0.0039  max mem: 22446
+train: [8]  [280/400]  eta: 0:00:54  lr: 0.000257  loss: 0.3280 (0.3677)  grad: 0.2804 (0.2956)  time: 0.4756  data: 0.0043  max mem: 22446
+train: [8]  [300/400]  eta: 0:00:46  lr: 0.000256  loss: 0.3314 (0.3676)  grad: 0.3038 (0.2963)  time: 0.6202  data: 0.1696  max mem: 22446
+train: [8]  [320/400]  eta: 0:00:37  lr: 0.000255  loss: 0.2994 (0.3603)  grad: 0.2538 (0.2911)  time: 0.4478  data: 0.0035  max mem: 22446
+train: [8]  [340/400]  eta: 0:00:27  lr: 0.000254  loss: 0.2498 (0.3559)  grad: 0.2556 (0.2900)  time: 0.4461  data: 0.0042  max mem: 22446
+train: [8]  [360/400]  eta: 0:00:18  lr: 0.000253  loss: 0.2602 (0.3501)  grad: 0.2556 (0.2869)  time: 0.4564  data: 0.0043  max mem: 22446
+train: [8]  [380/400]  eta: 0:00:09  lr: 0.000252  loss: 0.2426 (0.3458)  grad: 0.2095 (0.2840)  time: 0.4475  data: 0.0041  max mem: 22446
+train: [8]  [399/400]  eta: 0:00:00  lr: 0.000250  loss: 0.2426 (0.3427)  grad: 0.1924 (0.2810)  time: 0.4497  data: 0.0043  max mem: 22446
+train: [8] Total time: 0:03:05 (0.4643 s / it)
+train: [8] Summary: lr: 0.000250  loss: 0.2426 (0.3427)  grad: 0.1924 (0.2810)
+eval (validation): [8]  [ 0/63]  eta: 0:03:12    time: 3.0485  data: 2.7769  max mem: 22446
+eval (validation): [8]  [20/63]  eta: 0:00:21    time: 0.3620  data: 0.0041  max mem: 22446
+eval (validation): [8]  [40/63]  eta: 0:00:09    time: 0.3225  data: 0.0029  max mem: 22446
+eval (validation): [8]  [60/63]  eta: 0:00:01    time: 0.3189  data: 0.0033  max mem: 22446
+eval (validation): [8]  [62/63]  eta: 0:00:00    time: 0.3173  data: 0.0032  max mem: 22446
+eval (validation): [8] Total time: 0:00:24 (0.3814 s / it)
+cv: [8]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.074  acc: 0.982  f1: 0.978
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-best.pth
+train: [9]  [  0/400]  eta: 0:20:35  lr: nan  time: 3.0894  data: 2.7079  max mem: 22446
+train: [9]  [ 20/400]  eta: 0:03:41  lr: 0.000249  loss: 0.2693 (0.3088)  grad: 0.2029 (0.2103)  time: 0.4585  data: 0.0025  max mem: 22446
+train: [9]  [ 40/400]  eta: 0:03:08  lr: 0.000248  loss: 0.2693 (0.2900)  grad: 0.2113 (0.2205)  time: 0.4623  data: 0.0041  max mem: 22446
+train: [9]  [ 60/400]  eta: 0:02:50  lr: 0.000247  loss: 0.2441 (0.2768)  grad: 0.2370 (0.2298)  time: 0.4502  data: 0.0040  max mem: 22446
+train: [9]  [ 80/400]  eta: 0:02:36  lr: 0.000246  loss: 0.2449 (0.2711)  grad: 0.2078 (0.2242)  time: 0.4535  data: 0.0042  max mem: 22446
+train: [9]  [100/400]  eta: 0:02:24  lr: 0.000244  loss: 0.2474 (0.2697)  grad: 0.2099 (0.2218)  time: 0.4582  data: 0.0042  max mem: 22446
+train: [9]  [120/400]  eta: 0:02:13  lr: 0.000243  loss: 0.2633 (0.2736)  grad: 0.2142 (0.2251)  time: 0.4478  data: 0.0044  max mem: 22446
+train: [9]  [140/400]  eta: 0:02:03  lr: 0.000242  loss: 0.2185 (0.2688)  grad: 0.2196 (0.2225)  time: 0.4537  data: 0.0042  max mem: 22446
+train: [9]  [160/400]  eta: 0:01:53  lr: 0.000241  loss: 0.1907 (0.2630)  grad: 0.2068 (0.2227)  time: 0.4641  data: 0.0046  max mem: 22446
+train: [9]  [180/400]  eta: 0:01:43  lr: 0.000240  loss: 0.2112 (0.2578)  grad: 0.2239 (0.2207)  time: 0.4523  data: 0.0043  max mem: 22446
+train: [9]  [200/400]  eta: 0:01:33  lr: 0.000238  loss: 0.2162 (0.2572)  grad: 0.2165 (0.2196)  time: 0.4456  data: 0.0042  max mem: 22446
+train: [9]  [220/400]  eta: 0:01:24  lr: 0.000237  loss: 0.2335 (0.2572)  grad: 0.2066 (0.2174)  time: 0.4658  data: 0.0041  max mem: 22446
+train: [9]  [240/400]  eta: 0:01:14  lr: 0.000236  loss: 0.2105 (0.2538)  grad: 0.1801 (0.2159)  time: 0.4653  data: 0.0044  max mem: 22446
+train: [9]  [260/400]  eta: 0:01:05  lr: 0.000234  loss: 0.1910 (0.2524)  grad: 0.2211 (0.2160)  time: 0.4406  data: 0.0042  max mem: 22446
+train: [9]  [280/400]  eta: 0:00:55  lr: 0.000233  loss: 0.2292 (0.2538)  grad: 0.2297 (0.2168)  time: 0.4718  data: 0.0045  max mem: 22446
+train: [9]  [300/400]  eta: 0:00:47  lr: 0.000232  loss: 0.2349 (0.2528)  grad: 0.2168 (0.2173)  time: 0.6276  data: 0.1821  max mem: 22446
+train: [9]  [320/400]  eta: 0:00:38  lr: 0.000230  loss: 0.1755 (0.2474)  grad: 0.1653 (0.2126)  time: 0.4540  data: 0.0035  max mem: 22446
+train: [9]  [340/400]  eta: 0:00:28  lr: 0.000229  loss: 0.1865 (0.2463)  grad: 0.1646 (0.2113)  time: 0.4630  data: 0.0044  max mem: 22446
+train: [9]  [360/400]  eta: 0:00:18  lr: 0.000228  loss: 0.2090 (0.2436)  grad: 0.1730 (0.2090)  time: 0.4575  data: 0.0042  max mem: 22446
+train: [9]  [380/400]  eta: 0:00:09  lr: 0.000226  loss: 0.1696 (0.2406)  grad: 0.1605 (0.2063)  time: 0.4676  data: 0.0041  max mem: 22446
+train: [9]  [399/400]  eta: 0:00:00  lr: 0.000225  loss: 0.1852 (0.2381)  grad: 0.1523 (0.2041)  time: 0.4777  data: 0.0044  max mem: 22446
+train: [9] Total time: 0:03:09 (0.4739 s / it)
+train: [9] Summary: lr: 0.000225  loss: 0.1852 (0.2381)  grad: 0.1523 (0.2041)
+eval (validation): [9]  [ 0/63]  eta: 0:03:18    time: 3.1479  data: 2.9038  max mem: 22446
+eval (validation): [9]  [20/63]  eta: 0:00:20    time: 0.3349  data: 0.0036  max mem: 22446
+eval (validation): [9]  [40/63]  eta: 0:00:09    time: 0.3477  data: 0.0028  max mem: 22446
+eval (validation): [9]  [60/63]  eta: 0:00:01    time: 0.3397  data: 0.0037  max mem: 22446
+eval (validation): [9]  [62/63]  eta: 0:00:00    time: 0.3391  data: 0.0037  max mem: 22446
+eval (validation): [9] Total time: 0:00:24 (0.3904 s / it)
+cv: [9]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.074  acc: 0.981  f1: 0.978
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [10]  [  0/400]  eta: 0:24:56  lr: nan  time: 3.7400  data: 3.3894  max mem: 22446
+train: [10]  [ 20/400]  eta: 0:03:44  lr: 0.000224  loss: 0.2301 (0.2314)  grad: 0.1539 (0.1563)  time: 0.4344  data: 0.0018  max mem: 22446
+train: [10]  [ 40/400]  eta: 0:03:08  lr: 0.000222  loss: 0.2035 (0.2162)  grad: 0.1539 (0.1559)  time: 0.4525  data: 0.0039  max mem: 22446
+train: [10]  [ 60/400]  eta: 0:02:50  lr: 0.000221  loss: 0.2024 (0.2105)  grad: 0.1444 (0.1525)  time: 0.4524  data: 0.0042  max mem: 22446
+train: [10]  [ 80/400]  eta: 0:02:35  lr: 0.000220  loss: 0.1809 (0.2028)  grad: 0.1421 (0.1523)  time: 0.4476  data: 0.0040  max mem: 22446
+train: [10]  [100/400]  eta: 0:02:24  lr: 0.000218  loss: 0.1754 (0.1999)  grad: 0.1448 (0.1545)  time: 0.4562  data: 0.0042  max mem: 22446
+train: [10]  [120/400]  eta: 0:02:13  lr: 0.000217  loss: 0.1823 (0.1957)  grad: 0.1525 (0.1549)  time: 0.4447  data: 0.0041  max mem: 22446
+train: [10]  [140/400]  eta: 0:02:02  lr: 0.000215  loss: 0.1873 (0.1963)  grad: 0.1525 (0.1540)  time: 0.4416  data: 0.0042  max mem: 22446
+train: [10]  [160/400]  eta: 0:01:52  lr: 0.000214  loss: 0.1714 (0.1923)  grad: 0.1518 (0.1531)  time: 0.4589  data: 0.0044  max mem: 22446
+train: [10]  [180/400]  eta: 0:01:42  lr: 0.000213  loss: 0.1693 (0.1906)  grad: 0.1518 (0.1529)  time: 0.4362  data: 0.0042  max mem: 22446
+train: [10]  [200/400]  eta: 0:01:32  lr: 0.000211  loss: 0.1963 (0.1898)  grad: 0.1543 (0.1550)  time: 0.4499  data: 0.0043  max mem: 22446
+train: [10]  [220/400]  eta: 0:01:23  lr: 0.000210  loss: 0.1783 (0.1882)  grad: 0.1179 (0.1505)  time: 0.4499  data: 0.0044  max mem: 22446
+train: [10]  [240/400]  eta: 0:01:13  lr: 0.000208  loss: 0.1624 (0.1882)  grad: 0.1096 (0.1492)  time: 0.4487  data: 0.0043  max mem: 22446
+train: [10]  [260/400]  eta: 0:01:04  lr: 0.000207  loss: 0.1672 (0.1880)  grad: 0.1315 (0.1484)  time: 0.4580  data: 0.0042  max mem: 22446
+train: [10]  [280/400]  eta: 0:00:55  lr: 0.000205  loss: 0.1537 (0.1860)  grad: 0.1194 (0.1455)  time: 0.4478  data: 0.0044  max mem: 22446
+train: [10]  [300/400]  eta: 0:00:47  lr: 0.000204  loss: 0.1547 (0.1854)  grad: 0.1061 (0.1443)  time: 0.6186  data: 0.1789  max mem: 22446
+train: [10]  [320/400]  eta: 0:00:37  lr: 0.000202  loss: 0.1584 (0.1840)  grad: 0.1250 (0.1444)  time: 0.4580  data: 0.0029  max mem: 22446
+train: [10]  [340/400]  eta: 0:00:28  lr: 0.000201  loss: 0.1563 (0.1823)  grad: 0.1235 (0.1432)  time: 0.4610  data: 0.0040  max mem: 22446
+train: [10]  [360/400]  eta: 0:00:18  lr: 0.000199  loss: 0.1403 (0.1808)  grad: 0.1099 (0.1407)  time: 0.4464  data: 0.0039  max mem: 22446
+train: [10]  [380/400]  eta: 0:00:09  lr: 0.000198  loss: 0.1393 (0.1800)  grad: 0.1005 (0.1390)  time: 0.4538  data: 0.0041  max mem: 22446
+train: [10]  [399/400]  eta: 0:00:00  lr: 0.000196  loss: 0.1561 (0.1790)  grad: 0.1006 (0.1375)  time: 0.4499  data: 0.0044  max mem: 22446
+train: [10] Total time: 0:03:06 (0.4671 s / it)
+train: [10] Summary: lr: 0.000196  loss: 0.1561 (0.1790)  grad: 0.1006 (0.1375)
+eval (validation): [10]  [ 0/63]  eta: 0:03:20    time: 3.1841  data: 2.9498  max mem: 22446
+eval (validation): [10]  [20/63]  eta: 0:00:19    time: 0.3291  data: 0.0073  max mem: 22446
+eval (validation): [10]  [40/63]  eta: 0:00:09    time: 0.3465  data: 0.0027  max mem: 22446
+eval (validation): [10]  [60/63]  eta: 0:00:01    time: 0.3271  data: 0.0035  max mem: 22446
+eval (validation): [10]  [62/63]  eta: 0:00:00    time: 0.3260  data: 0.0036  max mem: 22446
+eval (validation): [10] Total time: 0:00:24 (0.3846 s / it)
+cv: [10]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.074  acc: 0.981  f1: 0.978
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [11]  [  0/400]  eta: 0:21:43  lr: nan  time: 3.2596  data: 2.9195  max mem: 22446
+train: [11]  [ 20/400]  eta: 0:03:40  lr: 0.000195  loss: 0.1531 (0.1593)  grad: 0.0647 (0.0809)  time: 0.4453  data: 0.0081  max mem: 22446
+train: [11]  [ 40/400]  eta: 0:03:03  lr: 0.000193  loss: 0.1531 (0.1555)  grad: 0.0858 (0.0865)  time: 0.4380  data: 0.0036  max mem: 22446
+train: [11]  [ 60/400]  eta: 0:02:46  lr: 0.000192  loss: 0.1466 (0.1589)  grad: 0.0902 (0.0902)  time: 0.4455  data: 0.0042  max mem: 22446
+train: [11]  [ 80/400]  eta: 0:02:32  lr: 0.000190  loss: 0.1554 (0.1593)  grad: 0.1013 (0.0951)  time: 0.4442  data: 0.0042  max mem: 22446
+train: [11]  [100/400]  eta: 0:02:21  lr: 0.000189  loss: 0.1502 (0.1554)  grad: 0.1039 (0.0978)  time: 0.4399  data: 0.0043  max mem: 22446
+train: [11]  [120/400]  eta: 0:02:10  lr: 0.000187  loss: 0.1437 (0.1543)  grad: 0.0987 (0.0989)  time: 0.4401  data: 0.0042  max mem: 22446
+train: [11]  [140/400]  eta: 0:02:00  lr: 0.000186  loss: 0.1440 (0.1547)  grad: 0.0969 (0.0989)  time: 0.4392  data: 0.0041  max mem: 22446
+train: [11]  [160/400]  eta: 0:01:50  lr: 0.000184  loss: 0.1566 (0.1561)  grad: 0.0867 (0.0988)  time: 0.4377  data: 0.0040  max mem: 22446
+train: [11]  [180/400]  eta: 0:01:40  lr: 0.000183  loss: 0.1549 (0.1553)  grad: 0.1042 (0.1005)  time: 0.4401  data: 0.0041  max mem: 22446
+train: [11]  [200/400]  eta: 0:01:31  lr: 0.000181  loss: 0.1378 (0.1531)  grad: 0.1042 (0.1007)  time: 0.4509  data: 0.0043  max mem: 22446
+train: [11]  [220/400]  eta: 0:01:21  lr: 0.000180  loss: 0.1323 (0.1520)  grad: 0.0979 (0.1007)  time: 0.4451  data: 0.0042  max mem: 22446
+train: [11]  [240/400]  eta: 0:01:12  lr: 0.000178  loss: 0.1472 (0.1511)  grad: 0.1023 (0.1006)  time: 0.4532  data: 0.0042  max mem: 22446
+train: [11]  [260/400]  eta: 0:01:03  lr: 0.000177  loss: 0.1437 (0.1518)  grad: 0.1042 (0.1008)  time: 0.4515  data: 0.0041  max mem: 22446
+train: [11]  [280/400]  eta: 0:00:54  lr: 0.000175  loss: 0.1431 (0.1522)  grad: 0.1156 (0.1019)  time: 0.4529  data: 0.0044  max mem: 22446
+train: [11]  [300/400]  eta: 0:00:46  lr: 0.000174  loss: 0.1526 (0.1526)  grad: 0.1102 (0.1027)  time: 0.5940  data: 0.1672  max mem: 22446
+train: [11]  [320/400]  eta: 0:00:36  lr: 0.000172  loss: 0.1542 (0.1524)  grad: 0.0998 (0.1034)  time: 0.4370  data: 0.0032  max mem: 22446
+train: [11]  [340/400]  eta: 0:00:27  lr: 0.000170  loss: 0.1449 (0.1521)  grad: 0.0949 (0.1026)  time: 0.4512  data: 0.0044  max mem: 22446
+train: [11]  [360/400]  eta: 0:00:18  lr: 0.000169  loss: 0.1449 (0.1515)  grad: 0.0879 (0.1025)  time: 0.4426  data: 0.0045  max mem: 22446
+train: [11]  [380/400]  eta: 0:00:09  lr: 0.000167  loss: 0.1364 (0.1507)  grad: 0.0868 (0.1016)  time: 0.4422  data: 0.0043  max mem: 22446
+train: [11]  [399/400]  eta: 0:00:00  lr: 0.000166  loss: 0.1387 (0.1507)  grad: 0.0867 (0.1008)  time: 0.4266  data: 0.0039  max mem: 22446
+train: [11] Total time: 0:03:03 (0.4584 s / it)
+train: [11] Summary: lr: 0.000166  loss: 0.1387 (0.1507)  grad: 0.0867 (0.1008)
+eval (validation): [11]  [ 0/63]  eta: 0:03:05    time: 2.9469  data: 2.6801  max mem: 22446
+eval (validation): [11]  [20/63]  eta: 0:00:20    time: 0.3430  data: 0.0040  max mem: 22446
+eval (validation): [11]  [40/63]  eta: 0:00:09    time: 0.3238  data: 0.0033  max mem: 22446
+eval (validation): [11]  [60/63]  eta: 0:00:01    time: 0.3076  data: 0.0034  max mem: 22446
+eval (validation): [11]  [62/63]  eta: 0:00:00    time: 0.3082  data: 0.0034  max mem: 22446
+eval (validation): [11] Total time: 0:00:23 (0.3705 s / it)
+cv: [11]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.075  acc: 0.981  f1: 0.978
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [12]  [  0/400]  eta: 0:20:23  lr: nan  time: 3.0582  data: 2.6969  max mem: 22446
+train: [12]  [ 20/400]  eta: 0:03:31  lr: 0.000164  loss: 0.1337 (0.1331)  grad: 0.0670 (0.0790)  time: 0.4322  data: 0.0029  max mem: 22446
+train: [12]  [ 40/400]  eta: 0:02:59  lr: 0.000163  loss: 0.1337 (0.1335)  grad: 0.0712 (0.0775)  time: 0.4382  data: 0.0041  max mem: 22446
+train: [12]  [ 60/400]  eta: 0:02:42  lr: 0.000161  loss: 0.1323 (0.1319)  grad: 0.0750 (0.0789)  time: 0.4312  data: 0.0044  max mem: 22446
+train: [12]  [ 80/400]  eta: 0:02:29  lr: 0.000160  loss: 0.1350 (0.1328)  grad: 0.0811 (0.0799)  time: 0.4326  data: 0.0042  max mem: 22446
+train: [12]  [100/400]  eta: 0:02:17  lr: 0.000158  loss: 0.1394 (0.1360)  grad: 0.0757 (0.0800)  time: 0.4351  data: 0.0042  max mem: 22446
+train: [12]  [120/400]  eta: 0:02:07  lr: 0.000156  loss: 0.1346 (0.1364)  grad: 0.0812 (0.0830)  time: 0.4319  data: 0.0041  max mem: 22446
+train: [12]  [140/400]  eta: 0:01:57  lr: 0.000155  loss: 0.1271 (0.1356)  grad: 0.0928 (0.0838)  time: 0.4348  data: 0.0045  max mem: 22446
+train: [12]  [160/400]  eta: 0:01:48  lr: 0.000153  loss: 0.1343 (0.1361)  grad: 0.0849 (0.0833)  time: 0.4516  data: 0.0044  max mem: 22446
+train: [12]  [180/400]  eta: 0:01:39  lr: 0.000152  loss: 0.1348 (0.1367)  grad: 0.0760 (0.0837)  time: 0.4475  data: 0.0045  max mem: 22446
+train: [12]  [200/400]  eta: 0:01:29  lr: 0.000150  loss: 0.1314 (0.1364)  grad: 0.0764 (0.0837)  time: 0.4339  data: 0.0042  max mem: 22446
+train: [12]  [220/400]  eta: 0:01:20  lr: 0.000149  loss: 0.1312 (0.1372)  grad: 0.0791 (0.0839)  time: 0.4368  data: 0.0043  max mem: 22446
+train: [12]  [240/400]  eta: 0:01:11  lr: 0.000147  loss: 0.1399 (0.1378)  grad: 0.0843 (0.0842)  time: 0.4458  data: 0.0044  max mem: 22446
+train: [12]  [260/400]  eta: 0:01:02  lr: 0.000145  loss: 0.1388 (0.1370)  grad: 0.0730 (0.0832)  time: 0.4598  data: 0.0046  max mem: 22446
+train: [12]  [280/400]  eta: 0:00:53  lr: 0.000144  loss: 0.1222 (0.1365)  grad: 0.0654 (0.0821)  time: 0.4526  data: 0.0044  max mem: 22446
+train: [12]  [300/400]  eta: 0:00:45  lr: 0.000142  loss: 0.1301 (0.1368)  grad: 0.0690 (0.0815)  time: 0.5839  data: 0.1629  max mem: 22446
+train: [12]  [320/400]  eta: 0:00:36  lr: 0.000141  loss: 0.1301 (0.1361)  grad: 0.0656 (0.0803)  time: 0.4422  data: 0.0044  max mem: 22446
+train: [12]  [340/400]  eta: 0:00:27  lr: 0.000139  loss: 0.1241 (0.1360)  grad: 0.0609 (0.0799)  time: 0.4587  data: 0.0043  max mem: 22446
+train: [12]  [360/400]  eta: 0:00:18  lr: 0.000138  loss: 0.1300 (0.1356)  grad: 0.0711 (0.0792)  time: 0.4377  data: 0.0042  max mem: 22446
+train: [12]  [380/400]  eta: 0:00:09  lr: 0.000136  loss: 0.1297 (0.1357)  grad: 0.0700 (0.0789)  time: 0.4420  data: 0.0041  max mem: 22446
+train: [12]  [399/400]  eta: 0:00:00  lr: 0.000134  loss: 0.1258 (0.1352)  grad: 0.0582 (0.0781)  time: 0.4455  data: 0.0042  max mem: 22446
+train: [12] Total time: 0:03:02 (0.4558 s / it)
+train: [12] Summary: lr: 0.000134  loss: 0.1258 (0.1352)  grad: 0.0582 (0.0781)
+eval (validation): [12]  [ 0/63]  eta: 0:03:15    time: 3.1056  data: 2.8294  max mem: 22446
+eval (validation): [12]  [20/63]  eta: 0:00:20    time: 0.3512  data: 0.0030  max mem: 22446
+eval (validation): [12]  [40/63]  eta: 0:00:09    time: 0.3214  data: 0.0035  max mem: 22446
+eval (validation): [12]  [60/63]  eta: 0:00:01    time: 0.3154  data: 0.0035  max mem: 22446
+eval (validation): [12]  [62/63]  eta: 0:00:00    time: 0.3164  data: 0.0036  max mem: 22446
+eval (validation): [12] Total time: 0:00:23 (0.3780 s / it)
+cv: [12]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.075  acc: 0.981  f1: 0.977
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [13]  [  0/400]  eta: 0:21:21  lr: nan  time: 3.2037  data: 2.7996  max mem: 22446
+train: [13]  [ 20/400]  eta: 0:03:36  lr: 0.000133  loss: 0.1323 (0.1311)  grad: 0.0746 (0.0749)  time: 0.4374  data: 0.0024  max mem: 22446
+train: [13]  [ 40/400]  eta: 0:03:00  lr: 0.000131  loss: 0.1220 (0.1236)  grad: 0.0689 (0.0694)  time: 0.4301  data: 0.0042  max mem: 22446
+train: [13]  [ 60/400]  eta: 0:02:43  lr: 0.000130  loss: 0.1144 (0.1216)  grad: 0.0605 (0.0694)  time: 0.4364  data: 0.0042  max mem: 22446
+train: [13]  [ 80/400]  eta: 0:02:30  lr: 0.000128  loss: 0.1174 (0.1235)  grad: 0.0615 (0.0678)  time: 0.4373  data: 0.0039  max mem: 22446
+train: [13]  [100/400]  eta: 0:02:19  lr: 0.000127  loss: 0.1232 (0.1235)  grad: 0.0620 (0.0671)  time: 0.4454  data: 0.0043  max mem: 22446
+train: [13]  [120/400]  eta: 0:02:09  lr: 0.000125  loss: 0.1232 (0.1234)  grad: 0.0607 (0.0669)  time: 0.4523  data: 0.0045  max mem: 22446
+train: [13]  [140/400]  eta: 0:01:59  lr: 0.000124  loss: 0.1137 (0.1247)  grad: 0.0617 (0.0668)  time: 0.4431  data: 0.0042  max mem: 22446
+train: [13]  [160/400]  eta: 0:01:49  lr: 0.000122  loss: 0.1162 (0.1247)  grad: 0.0586 (0.0655)  time: 0.4436  data: 0.0042  max mem: 22446
+train: [13]  [180/400]  eta: 0:01:40  lr: 0.000120  loss: 0.1240 (0.1258)  grad: 0.0532 (0.0651)  time: 0.4479  data: 0.0043  max mem: 22446
+train: [13]  [200/400]  eta: 0:01:30  lr: 0.000119  loss: 0.1240 (0.1264)  grad: 0.0571 (0.0651)  time: 0.4367  data: 0.0042  max mem: 22446
+train: [13]  [220/400]  eta: 0:01:21  lr: 0.000117  loss: 0.1179 (0.1256)  grad: 0.0628 (0.0658)  time: 0.4331  data: 0.0042  max mem: 22446
+train: [13]  [240/400]  eta: 0:01:12  lr: 0.000116  loss: 0.1122 (0.1253)  grad: 0.0538 (0.0653)  time: 0.4464  data: 0.0043  max mem: 22446
+train: [13]  [260/400]  eta: 0:01:03  lr: 0.000114  loss: 0.1230 (0.1258)  grad: 0.0585 (0.0655)  time: 0.4445  data: 0.0043  max mem: 22446
+train: [13]  [280/400]  eta: 0:00:54  lr: 0.000113  loss: 0.1183 (0.1251)  grad: 0.0618 (0.0653)  time: 0.4606  data: 0.0047  max mem: 22446
+train: [13]  [300/400]  eta: 0:00:46  lr: 0.000111  loss: 0.1047 (0.1251)  grad: 0.0606 (0.0653)  time: 0.6179  data: 0.1708  max mem: 22446
+train: [13]  [320/400]  eta: 0:00:36  lr: 0.000110  loss: 0.1215 (0.1251)  grad: 0.0573 (0.0648)  time: 0.4241  data: 0.0032  max mem: 22446
+train: [13]  [340/400]  eta: 0:00:27  lr: 0.000108  loss: 0.1215 (0.1244)  grad: 0.0544 (0.0643)  time: 0.4468  data: 0.0040  max mem: 22446
+train: [13]  [360/400]  eta: 0:00:18  lr: 0.000107  loss: 0.1013 (0.1234)  grad: 0.0516 (0.0637)  time: 0.4538  data: 0.0044  max mem: 22446
+train: [13]  [380/400]  eta: 0:00:09  lr: 0.000105  loss: 0.1123 (0.1236)  grad: 0.0516 (0.0634)  time: 0.4408  data: 0.0041  max mem: 22446
+train: [13]  [399/400]  eta: 0:00:00  lr: 0.000104  loss: 0.1202 (0.1237)  grad: 0.0530 (0.0632)  time: 0.4378  data: 0.0040  max mem: 22446
+train: [13] Total time: 0:03:03 (0.4582 s / it)
+train: [13] Summary: lr: 0.000104  loss: 0.1202 (0.1237)  grad: 0.0530 (0.0632)
+eval (validation): [13]  [ 0/63]  eta: 0:03:09    time: 3.0102  data: 2.7695  max mem: 22446
+eval (validation): [13]  [20/63]  eta: 0:00:19    time: 0.3332  data: 0.0076  max mem: 22446
+eval (validation): [13]  [40/63]  eta: 0:00:09    time: 0.3324  data: 0.0032  max mem: 22446
+eval (validation): [13]  [60/63]  eta: 0:00:01    time: 0.3066  data: 0.0031  max mem: 22446
+eval (validation): [13]  [62/63]  eta: 0:00:00    time: 0.3056  data: 0.0031  max mem: 22446
+eval (validation): [13] Total time: 0:00:23 (0.3702 s / it)
+cv: [13]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.075  acc: 0.981  f1: 0.978
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [14]  [  0/400]  eta: 0:21:00  lr: nan  time: 3.1514  data: 2.7627  max mem: 22446
+train: [14]  [ 20/400]  eta: 0:03:39  lr: 0.000102  loss: 0.1287 (0.1277)  grad: 0.0548 (0.0578)  time: 0.4491  data: 0.0036  max mem: 22446
+train: [14]  [ 40/400]  eta: 0:03:02  lr: 0.000101  loss: 0.1145 (0.1199)  grad: 0.0555 (0.0583)  time: 0.4304  data: 0.0041  max mem: 22446
+train: [14]  [ 60/400]  eta: 0:02:43  lr: 0.000099  loss: 0.1115 (0.1177)  grad: 0.0559 (0.0578)  time: 0.4319  data: 0.0044  max mem: 22446
+train: [14]  [ 80/400]  eta: 0:02:30  lr: 0.000098  loss: 0.1132 (0.1163)  grad: 0.0572 (0.0583)  time: 0.4379  data: 0.0042  max mem: 22446
+train: [14]  [100/400]  eta: 0:02:19  lr: 0.000096  loss: 0.1044 (0.1153)  grad: 0.0566 (0.0572)  time: 0.4414  data: 0.0042  max mem: 22446
+train: [14]  [120/400]  eta: 0:02:09  lr: 0.000095  loss: 0.1227 (0.1182)  grad: 0.0514 (0.0574)  time: 0.4554  data: 0.0042  max mem: 22446
+train: [14]  [140/400]  eta: 0:01:59  lr: 0.000093  loss: 0.1321 (0.1214)  grad: 0.0589 (0.0581)  time: 0.4475  data: 0.0041  max mem: 22446
+train: [14]  [160/400]  eta: 0:01:49  lr: 0.000092  loss: 0.1111 (0.1200)  grad: 0.0553 (0.0575)  time: 0.4300  data: 0.0042  max mem: 22446
+train: [14]  [180/400]  eta: 0:01:40  lr: 0.000090  loss: 0.1077 (0.1195)  grad: 0.0536 (0.0574)  time: 0.4447  data: 0.0043  max mem: 22446
+train: [14]  [200/400]  eta: 0:01:30  lr: 0.000089  loss: 0.1101 (0.1194)  grad: 0.0531 (0.0569)  time: 0.4414  data: 0.0042  max mem: 22446
+train: [14]  [220/400]  eta: 0:01:21  lr: 0.000088  loss: 0.1101 (0.1189)  grad: 0.0531 (0.0574)  time: 0.4439  data: 0.0044  max mem: 22446
+train: [14]  [240/400]  eta: 0:01:12  lr: 0.000086  loss: 0.1087 (0.1178)  grad: 0.0593 (0.0575)  time: 0.4392  data: 0.0042  max mem: 22446
+train: [14]  [260/400]  eta: 0:01:03  lr: 0.000085  loss: 0.1087 (0.1172)  grad: 0.0526 (0.0572)  time: 0.4427  data: 0.0043  max mem: 22446
+train: [14]  [280/400]  eta: 0:00:54  lr: 0.000083  loss: 0.1164 (0.1184)  grad: 0.0526 (0.0571)  time: 0.4407  data: 0.0044  max mem: 22446
+train: [14]  [300/400]  eta: 0:00:46  lr: 0.000082  loss: 0.1222 (0.1183)  grad: 0.0551 (0.0571)  time: 0.6230  data: 0.1741  max mem: 22446
+train: [14]  [320/400]  eta: 0:00:36  lr: 0.000081  loss: 0.1115 (0.1180)  grad: 0.0523 (0.0568)  time: 0.4604  data: 0.0038  max mem: 22446
+train: [14]  [340/400]  eta: 0:00:27  lr: 0.000079  loss: 0.1159 (0.1185)  grad: 0.0515 (0.0564)  time: 0.4342  data: 0.0039  max mem: 22446
+train: [14]  [360/400]  eta: 0:00:18  lr: 0.000078  loss: 0.1214 (0.1188)  grad: 0.0541 (0.0566)  time: 0.4559  data: 0.0043  max mem: 22446
+train: [14]  [380/400]  eta: 0:00:09  lr: 0.000076  loss: 0.1219 (0.1190)  grad: 0.0586 (0.0566)  time: 0.4566  data: 0.0043  max mem: 22446
+train: [14]  [399/400]  eta: 0:00:00  lr: 0.000075  loss: 0.1111 (0.1190)  grad: 0.0530 (0.0565)  time: 0.4393  data: 0.0039  max mem: 22446
+train: [14] Total time: 0:03:03 (0.4596 s / it)
+train: [14] Summary: lr: 0.000075  loss: 0.1111 (0.1190)  grad: 0.0530 (0.0565)
+eval (validation): [14]  [ 0/63]  eta: 0:03:10    time: 3.0224  data: 2.7785  max mem: 22446
+eval (validation): [14]  [20/63]  eta: 0:00:19    time: 0.3294  data: 0.0036  max mem: 22446
+eval (validation): [14]  [40/63]  eta: 0:00:09    time: 0.3358  data: 0.0032  max mem: 22446
+eval (validation): [14]  [60/63]  eta: 0:00:01    time: 0.3115  data: 0.0034  max mem: 22446
+eval (validation): [14]  [62/63]  eta: 0:00:00    time: 0.3105  data: 0.0033  max mem: 22446
+eval (validation): [14] Total time: 0:00:23 (0.3723 s / it)
+cv: [14]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.075  acc: 0.981  f1: 0.977
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [15]  [  0/400]  eta: 0:21:52  lr: nan  time: 3.2822  data: 2.8906  max mem: 22446
+train: [15]  [ 20/400]  eta: 0:03:44  lr: 0.000074  loss: 0.1200 (0.1236)  grad: 0.0471 (0.0511)  time: 0.4573  data: 0.0035  max mem: 22446
+train: [15]  [ 40/400]  eta: 0:03:06  lr: 0.000072  loss: 0.1200 (0.1216)  grad: 0.0505 (0.0538)  time: 0.4419  data: 0.0043  max mem: 22446
+train: [15]  [ 60/400]  eta: 0:02:47  lr: 0.000071  loss: 0.1147 (0.1171)  grad: 0.0520 (0.0536)  time: 0.4382  data: 0.0043  max mem: 22446
+train: [15]  [ 80/400]  eta: 0:02:34  lr: 0.000070  loss: 0.1080 (0.1154)  grad: 0.0498 (0.0534)  time: 0.4480  data: 0.0043  max mem: 22446
+train: [15]  [100/400]  eta: 0:02:21  lr: 0.000068  loss: 0.1082 (0.1139)  grad: 0.0503 (0.0533)  time: 0.4310  data: 0.0042  max mem: 22446
+train: [15]  [120/400]  eta: 0:02:11  lr: 0.000067  loss: 0.0985 (0.1123)  grad: 0.0491 (0.0533)  time: 0.4526  data: 0.0041  max mem: 22446
+train: [15]  [140/400]  eta: 0:02:01  lr: 0.000066  loss: 0.1047 (0.1124)  grad: 0.0508 (0.0530)  time: 0.4485  data: 0.0042  max mem: 22446
+train: [15]  [160/400]  eta: 0:01:50  lr: 0.000064  loss: 0.1111 (0.1119)  grad: 0.0546 (0.0535)  time: 0.4349  data: 0.0042  max mem: 22446
+train: [15]  [180/400]  eta: 0:01:41  lr: 0.000063  loss: 0.1102 (0.1123)  grad: 0.0514 (0.0532)  time: 0.4457  data: 0.0041  max mem: 22446
+train: [15]  [200/400]  eta: 0:01:31  lr: 0.000062  loss: 0.1102 (0.1127)  grad: 0.0461 (0.0530)  time: 0.4561  data: 0.0042  max mem: 22446
+train: [15]  [220/400]  eta: 0:01:22  lr: 0.000061  loss: 0.1078 (0.1123)  grad: 0.0502 (0.0530)  time: 0.4520  data: 0.0041  max mem: 22446
+train: [15]  [240/400]  eta: 0:01:13  lr: 0.000059  loss: 0.1003 (0.1118)  grad: 0.0541 (0.0532)  time: 0.4450  data: 0.0042  max mem: 22446
+train: [15]  [260/400]  eta: 0:01:03  lr: 0.000058  loss: 0.1150 (0.1122)  grad: 0.0504 (0.0528)  time: 0.4426  data: 0.0042  max mem: 22446
+train: [15]  [280/400]  eta: 0:00:54  lr: 0.000057  loss: 0.1159 (0.1124)  grad: 0.0499 (0.0528)  time: 0.4414  data: 0.0043  max mem: 22446
+train: [15]  [300/400]  eta: 0:00:46  lr: 0.000056  loss: 0.1031 (0.1120)  grad: 0.0499 (0.0527)  time: 0.6411  data: 0.1819  max mem: 22446
+train: [15]  [320/400]  eta: 0:00:37  lr: 0.000054  loss: 0.1030 (0.1120)  grad: 0.0536 (0.0528)  time: 0.4436  data: 0.0047  max mem: 22446
+train: [15]  [340/400]  eta: 0:00:27  lr: 0.000053  loss: 0.1075 (0.1123)  grad: 0.0552 (0.0529)  time: 0.4404  data: 0.0043  max mem: 22446
+train: [15]  [360/400]  eta: 0:00:18  lr: 0.000052  loss: 0.1083 (0.1123)  grad: 0.0552 (0.0530)  time: 0.4406  data: 0.0042  max mem: 22446
+train: [15]  [380/400]  eta: 0:00:09  lr: 0.000051  loss: 0.0992 (0.1115)  grad: 0.0489 (0.0528)  time: 0.4510  data: 0.0042  max mem: 22446
+train: [15]  [399/400]  eta: 0:00:00  lr: 0.000050  loss: 0.1103 (0.1124)  grad: 0.0486 (0.0527)  time: 0.4332  data: 0.0041  max mem: 22446
+train: [15] Total time: 0:03:04 (0.4619 s / it)
+train: [15] Summary: lr: 0.000050  loss: 0.1103 (0.1124)  grad: 0.0486 (0.0527)
+eval (validation): [15]  [ 0/63]  eta: 0:03:17    time: 3.1330  data: 2.8552  max mem: 22446
+eval (validation): [15]  [20/63]  eta: 0:00:19    time: 0.3209  data: 0.0033  max mem: 22446
+eval (validation): [15]  [40/63]  eta: 0:00:08    time: 0.3178  data: 0.0030  max mem: 22446
+eval (validation): [15]  [60/63]  eta: 0:00:01    time: 0.3072  data: 0.0034  max mem: 22446
+eval (validation): [15]  [62/63]  eta: 0:00:00    time: 0.3061  data: 0.0034  max mem: 22446
+eval (validation): [15] Total time: 0:00:22 (0.3643 s / it)
+cv: [15]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.076  acc: 0.981  f1: 0.977
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [16]  [  0/400]  eta: 0:20:40  lr: nan  time: 3.1007  data: 2.7595  max mem: 22446
+train: [16]  [ 20/400]  eta: 0:03:35  lr: 0.000048  loss: 0.1067 (0.1106)  grad: 0.0513 (0.0527)  time: 0.4411  data: 0.0029  max mem: 22446
+train: [16]  [ 40/400]  eta: 0:03:02  lr: 0.000047  loss: 0.1109 (0.1128)  grad: 0.0513 (0.0527)  time: 0.4439  data: 0.0041  max mem: 22446
+train: [16]  [ 60/400]  eta: 0:02:45  lr: 0.000046  loss: 0.1127 (0.1128)  grad: 0.0479 (0.0514)  time: 0.4440  data: 0.0041  max mem: 22446
+train: [16]  [ 80/400]  eta: 0:02:31  lr: 0.000045  loss: 0.1139 (0.1139)  grad: 0.0479 (0.0515)  time: 0.4382  data: 0.0041  max mem: 22446
+train: [16]  [100/400]  eta: 0:02:20  lr: 0.000044  loss: 0.1139 (0.1140)  grad: 0.0498 (0.0515)  time: 0.4399  data: 0.0042  max mem: 22446
+train: [16]  [120/400]  eta: 0:02:10  lr: 0.000043  loss: 0.1017 (0.1129)  grad: 0.0490 (0.0517)  time: 0.4481  data: 0.0041  max mem: 22446
+train: [16]  [140/400]  eta: 0:02:00  lr: 0.000042  loss: 0.1045 (0.1125)  grad: 0.0513 (0.0521)  time: 0.4486  data: 0.0043  max mem: 22446
+train: [16]  [160/400]  eta: 0:01:50  lr: 0.000041  loss: 0.1073 (0.1121)  grad: 0.0539 (0.0522)  time: 0.4324  data: 0.0042  max mem: 22446
+train: [16]  [180/400]  eta: 0:01:40  lr: 0.000040  loss: 0.1090 (0.1120)  grad: 0.0523 (0.0520)  time: 0.4404  data: 0.0042  max mem: 22446
+train: [16]  [200/400]  eta: 0:01:31  lr: 0.000039  loss: 0.1101 (0.1122)  grad: 0.0512 (0.0523)  time: 0.4474  data: 0.0043  max mem: 22446
+train: [16]  [220/400]  eta: 0:01:21  lr: 0.000038  loss: 0.1046 (0.1118)  grad: 0.0495 (0.0517)  time: 0.4393  data: 0.0042  max mem: 22446
+train: [16]  [240/400]  eta: 0:01:12  lr: 0.000036  loss: 0.1046 (0.1117)  grad: 0.0495 (0.0517)  time: 0.4428  data: 0.0041  max mem: 22446
+train: [16]  [260/400]  eta: 0:01:03  lr: 0.000035  loss: 0.1070 (0.1125)  grad: 0.0506 (0.0517)  time: 0.4466  data: 0.0041  max mem: 22446
+train: [16]  [280/400]  eta: 0:00:54  lr: 0.000034  loss: 0.1147 (0.1126)  grad: 0.0538 (0.0521)  time: 0.4411  data: 0.0041  max mem: 22446
+train: [16]  [300/400]  eta: 0:00:46  lr: 0.000033  loss: 0.1118 (0.1128)  grad: 0.0559 (0.0522)  time: 0.6156  data: 0.1740  max mem: 22446
+train: [16]  [320/400]  eta: 0:00:37  lr: 0.000032  loss: 0.1088 (0.1128)  grad: 0.0513 (0.0523)  time: 0.4612  data: 0.0036  max mem: 22446
+train: [16]  [340/400]  eta: 0:00:27  lr: 0.000031  loss: 0.1185 (0.1137)  grad: 0.0523 (0.0526)  time: 0.4501  data: 0.0042  max mem: 22446
+train: [16]  [360/400]  eta: 0:00:18  lr: 0.000031  loss: 0.1202 (0.1135)  grad: 0.0552 (0.0526)  time: 0.4381  data: 0.0042  max mem: 22446
+train: [16]  [380/400]  eta: 0:00:09  lr: 0.000030  loss: 0.1046 (0.1128)  grad: 0.0487 (0.0524)  time: 0.4465  data: 0.0042  max mem: 22446
+train: [16]  [399/400]  eta: 0:00:00  lr: 0.000029  loss: 0.1046 (0.1127)  grad: 0.0508 (0.0526)  time: 0.4543  data: 0.0045  max mem: 22446
+train: [16] Total time: 0:03:04 (0.4601 s / it)
+train: [16] Summary: lr: 0.000029  loss: 0.1046 (0.1127)  grad: 0.0508 (0.0526)
+eval (validation): [16]  [ 0/63]  eta: 0:03:07    time: 2.9746  data: 2.7499  max mem: 22446
+eval (validation): [16]  [20/63]  eta: 0:00:20    time: 0.3418  data: 0.0107  max mem: 22446
+eval (validation): [16]  [40/63]  eta: 0:00:09    time: 0.3244  data: 0.0032  max mem: 22446
+eval (validation): [16]  [60/63]  eta: 0:00:01    time: 0.3160  data: 0.0034  max mem: 22446
+eval (validation): [16]  [62/63]  eta: 0:00:00    time: 0.3132  data: 0.0035  max mem: 22446
+eval (validation): [16] Total time: 0:00:23 (0.3729 s / it)
+cv: [16]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.076  acc: 0.981  f1: 0.977
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [17]  [  0/400]  eta: 0:21:06  lr: nan  time: 3.1659  data: 2.7672  max mem: 22446
+train: [17]  [ 20/400]  eta: 0:03:41  lr: 0.000028  loss: 0.0944 (0.1025)  grad: 0.0497 (0.0507)  time: 0.4547  data: 0.0037  max mem: 22446
+train: [17]  [ 40/400]  eta: 0:03:04  lr: 0.000027  loss: 0.1017 (0.1075)  grad: 0.0508 (0.0518)  time: 0.4377  data: 0.0039  max mem: 22446
+train: [17]  [ 60/400]  eta: 0:02:45  lr: 0.000026  loss: 0.1126 (0.1121)  grad: 0.0494 (0.0514)  time: 0.4365  data: 0.0044  max mem: 22446
+train: [17]  [ 80/400]  eta: 0:02:31  lr: 0.000025  loss: 0.1218 (0.1141)  grad: 0.0494 (0.0515)  time: 0.4353  data: 0.0042  max mem: 22446
+train: [17]  [100/400]  eta: 0:02:20  lr: 0.000024  loss: 0.1168 (0.1136)  grad: 0.0496 (0.0515)  time: 0.4373  data: 0.0042  max mem: 22446
+train: [17]  [120/400]  eta: 0:02:10  lr: 0.000023  loss: 0.1079 (0.1125)  grad: 0.0473 (0.0510)  time: 0.4514  data: 0.0044  max mem: 22446
+train: [17]  [140/400]  eta: 0:02:00  lr: 0.000023  loss: 0.1070 (0.1113)  grad: 0.0463 (0.0509)  time: 0.4617  data: 0.0042  max mem: 22446
+train: [17]  [160/400]  eta: 0:01:50  lr: 0.000022  loss: 0.1182 (0.1115)  grad: 0.0520 (0.0513)  time: 0.4425  data: 0.0043  max mem: 22446
+train: [17]  [180/400]  eta: 0:01:40  lr: 0.000021  loss: 0.1049 (0.1112)  grad: 0.0520 (0.0511)  time: 0.4327  data: 0.0040  max mem: 22446
+train: [17]  [200/400]  eta: 0:01:31  lr: 0.000020  loss: 0.1024 (0.1109)  grad: 0.0486 (0.0513)  time: 0.4542  data: 0.0039  max mem: 22446
+train: [17]  [220/400]  eta: 0:01:22  lr: 0.000019  loss: 0.1027 (0.1108)  grad: 0.0486 (0.0512)  time: 0.4419  data: 0.0041  max mem: 22446
+train: [17]  [240/400]  eta: 0:01:12  lr: 0.000019  loss: 0.0999 (0.1107)  grad: 0.0486 (0.0511)  time: 0.4371  data: 0.0042  max mem: 22446
+train: [17]  [260/400]  eta: 0:01:03  lr: 0.000018  loss: 0.1095 (0.1110)  grad: 0.0507 (0.0513)  time: 0.4377  data: 0.0043  max mem: 22446
+train: [17]  [280/400]  eta: 0:00:54  lr: 0.000017  loss: 0.1039 (0.1106)  grad: 0.0496 (0.0511)  time: 0.4422  data: 0.0044  max mem: 22446
+train: [17]  [300/400]  eta: 0:00:46  lr: 0.000016  loss: 0.1039 (0.1109)  grad: 0.0499 (0.0515)  time: 0.6405  data: 0.1950  max mem: 22446
+train: [17]  [320/400]  eta: 0:00:37  lr: 0.000016  loss: 0.1134 (0.1111)  grad: 0.0509 (0.0516)  time: 0.4630  data: 0.0032  max mem: 22446
+train: [17]  [340/400]  eta: 0:00:27  lr: 0.000015  loss: 0.1067 (0.1109)  grad: 0.0479 (0.0514)  time: 0.4482  data: 0.0045  max mem: 22446
+train: [17]  [360/400]  eta: 0:00:18  lr: 0.000014  loss: 0.0955 (0.1104)  grad: 0.0461 (0.0511)  time: 0.4454  data: 0.0039  max mem: 22446
+train: [17]  [380/400]  eta: 0:00:09  lr: 0.000014  loss: 0.0992 (0.1106)  grad: 0.0475 (0.0511)  time: 0.4511  data: 0.0042  max mem: 22446
+train: [17]  [399/400]  eta: 0:00:00  lr: 0.000013  loss: 0.1074 (0.1104)  grad: 0.0487 (0.0510)  time: 0.4548  data: 0.0042  max mem: 22446
+train: [17] Total time: 0:03:05 (0.4627 s / it)
+train: [17] Summary: lr: 0.000013  loss: 0.1074 (0.1104)  grad: 0.0487 (0.0510)
+eval (validation): [17]  [ 0/63]  eta: 0:03:15    time: 3.0968  data: 2.8661  max mem: 22446
+eval (validation): [17]  [20/63]  eta: 0:00:21    time: 0.3592  data: 0.0036  max mem: 22446
+eval (validation): [17]  [40/63]  eta: 0:00:09    time: 0.3218  data: 0.0034  max mem: 22446
+eval (validation): [17]  [60/63]  eta: 0:00:01    time: 0.3155  data: 0.0031  max mem: 22446
+eval (validation): [17]  [62/63]  eta: 0:00:00    time: 0.3117  data: 0.0031  max mem: 22446
+eval (validation): [17] Total time: 0:00:23 (0.3800 s / it)
+cv: [17]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.076  acc: 0.981  f1: 0.977
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [18]  [  0/400]  eta: 0:19:52  lr: nan  time: 2.9813  data: 2.5983  max mem: 22446
+train: [18]  [ 20/400]  eta: 0:03:39  lr: 0.000012  loss: 0.1039 (0.1126)  grad: 0.0489 (0.0513)  time: 0.4565  data: 0.0042  max mem: 22446
+train: [18]  [ 40/400]  eta: 0:03:03  lr: 0.000012  loss: 0.1039 (0.1113)  grad: 0.0509 (0.0507)  time: 0.4389  data: 0.0037  max mem: 22446
+train: [18]  [ 60/400]  eta: 0:02:45  lr: 0.000011  loss: 0.1004 (0.1098)  grad: 0.0509 (0.0508)  time: 0.4366  data: 0.0041  max mem: 22446
+train: [18]  [ 80/400]  eta: 0:02:32  lr: 0.000011  loss: 0.1064 (0.1096)  grad: 0.0488 (0.0503)  time: 0.4427  data: 0.0042  max mem: 22446
+train: [18]  [100/400]  eta: 0:02:20  lr: 0.000010  loss: 0.1064 (0.1109)  grad: 0.0505 (0.0510)  time: 0.4479  data: 0.0042  max mem: 22446
+train: [18]  [120/400]  eta: 0:02:10  lr: 0.000009  loss: 0.1039 (0.1086)  grad: 0.0506 (0.0511)  time: 0.4437  data: 0.0041  max mem: 22446
+train: [18]  [140/400]  eta: 0:02:00  lr: 0.000009  loss: 0.1069 (0.1089)  grad: 0.0492 (0.0511)  time: 0.4533  data: 0.0043  max mem: 22446
+train: [18]  [160/400]  eta: 0:01:50  lr: 0.000008  loss: 0.1069 (0.1082)  grad: 0.0492 (0.0509)  time: 0.4526  data: 0.0042  max mem: 22446
+train: [18]  [180/400]  eta: 0:01:40  lr: 0.000008  loss: 0.1059 (0.1080)  grad: 0.0520 (0.0513)  time: 0.4328  data: 0.0039  max mem: 22446
+train: [18]  [200/400]  eta: 0:01:31  lr: 0.000007  loss: 0.1063 (0.1078)  grad: 0.0516 (0.0514)  time: 0.4551  data: 0.0043  max mem: 22446
+train: [18]  [220/400]  eta: 0:01:22  lr: 0.000007  loss: 0.1021 (0.1074)  grad: 0.0492 (0.0512)  time: 0.4501  data: 0.0044  max mem: 22446
+train: [18]  [240/400]  eta: 0:01:13  lr: 0.000006  loss: 0.1032 (0.1073)  grad: 0.0491 (0.0513)  time: 0.4441  data: 0.0043  max mem: 22446
+train: [18]  [260/400]  eta: 0:01:03  lr: 0.000006  loss: 0.1070 (0.1075)  grad: 0.0481 (0.0510)  time: 0.4427  data: 0.0042  max mem: 22446
+train: [18]  [280/400]  eta: 0:00:54  lr: 0.000006  loss: 0.1070 (0.1078)  grad: 0.0513 (0.0514)  time: 0.4370  data: 0.0043  max mem: 22446
+train: [18]  [300/400]  eta: 0:00:46  lr: 0.000005  loss: 0.1032 (0.1077)  grad: 0.0532 (0.0515)  time: 0.6069  data: 0.1719  max mem: 22446
+train: [18]  [320/400]  eta: 0:00:37  lr: 0.000005  loss: 0.1025 (0.1079)  grad: 0.0498 (0.0515)  time: 0.4573  data: 0.0033  max mem: 22446
+train: [18]  [340/400]  eta: 0:00:27  lr: 0.000004  loss: 0.1085 (0.1085)  grad: 0.0478 (0.0514)  time: 0.4654  data: 0.0043  max mem: 22446
+train: [18]  [360/400]  eta: 0:00:18  lr: 0.000004  loss: 0.1009 (0.1080)  grad: 0.0502 (0.0515)  time: 0.4424  data: 0.0039  max mem: 22446
+train: [18]  [380/400]  eta: 0:00:09  lr: 0.000004  loss: 0.1009 (0.1074)  grad: 0.0480 (0.0513)  time: 0.4480  data: 0.0039  max mem: 22446
+train: [18]  [399/400]  eta: 0:00:00  lr: 0.000003  loss: 0.0945 (0.1070)  grad: 0.0487 (0.0514)  time: 0.4514  data: 0.0041  max mem: 22446
+train: [18] Total time: 0:03:04 (0.4621 s / it)
+train: [18] Summary: lr: 0.000003  loss: 0.0945 (0.1070)  grad: 0.0487 (0.0514)
+eval (validation): [18]  [ 0/63]  eta: 0:03:11    time: 3.0320  data: 2.7611  max mem: 22446
+eval (validation): [18]  [20/63]  eta: 0:00:20    time: 0.3384  data: 0.0042  max mem: 22446
+eval (validation): [18]  [40/63]  eta: 0:00:09    time: 0.3172  data: 0.0030  max mem: 22446
+eval (validation): [18]  [60/63]  eta: 0:00:01    time: 0.3131  data: 0.0034  max mem: 22446
+eval (validation): [18]  [62/63]  eta: 0:00:00    time: 0.3104  data: 0.0033  max mem: 22446
+eval (validation): [18] Total time: 0:00:23 (0.3709 s / it)
+cv: [18]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.076  acc: 0.981  f1: 0.977
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+train: [19]  [  0/400]  eta: 0:20:10  lr: nan  time: 3.0266  data: 2.6546  max mem: 22446
+train: [19]  [ 20/400]  eta: 0:03:36  lr: 0.000003  loss: 0.1070 (0.1057)  grad: 0.0475 (0.0478)  time: 0.4457  data: 0.0035  max mem: 22446
+train: [19]  [ 40/400]  eta: 0:03:02  lr: 0.000003  loss: 0.1070 (0.1072)  grad: 0.0494 (0.0515)  time: 0.4432  data: 0.0041  max mem: 22446
+train: [19]  [ 60/400]  eta: 0:02:44  lr: 0.000002  loss: 0.1071 (0.1071)  grad: 0.0504 (0.0507)  time: 0.4381  data: 0.0041  max mem: 22446
+train: [19]  [ 80/400]  eta: 0:02:31  lr: 0.000002  loss: 0.1062 (0.1076)  grad: 0.0495 (0.0501)  time: 0.4406  data: 0.0043  max mem: 22446
+train: [19]  [100/400]  eta: 0:02:19  lr: 0.000002  loss: 0.1144 (0.1094)  grad: 0.0467 (0.0495)  time: 0.4365  data: 0.0043  max mem: 22446
+train: [19]  [120/400]  eta: 0:02:10  lr: 0.000002  loss: 0.1121 (0.1086)  grad: 0.0468 (0.0495)  time: 0.4543  data: 0.0043  max mem: 22446
+train: [19]  [140/400]  eta: 0:02:00  lr: 0.000001  loss: 0.0975 (0.1075)  grad: 0.0490 (0.0494)  time: 0.4508  data: 0.0042  max mem: 22446
+train: [19]  [160/400]  eta: 0:01:50  lr: 0.000001  loss: 0.1056 (0.1075)  grad: 0.0490 (0.0493)  time: 0.4465  data: 0.0042  max mem: 22446
+train: [19]  [180/400]  eta: 0:01:40  lr: 0.000001  loss: 0.1082 (0.1073)  grad: 0.0510 (0.0496)  time: 0.4398  data: 0.0041  max mem: 22446
+train: [19]  [200/400]  eta: 0:01:31  lr: 0.000001  loss: 0.1028 (0.1072)  grad: 0.0513 (0.0497)  time: 0.4423  data: 0.0043  max mem: 22446
+train: [19]  [220/400]  eta: 0:01:22  lr: 0.000001  loss: 0.1039 (0.1080)  grad: 0.0534 (0.0501)  time: 0.4532  data: 0.0042  max mem: 22446
+train: [19]  [240/400]  eta: 0:01:12  lr: 0.000001  loss: 0.1048 (0.1077)  grad: 0.0540 (0.0501)  time: 0.4422  data: 0.0043  max mem: 22446
+train: [19]  [260/400]  eta: 0:01:03  lr: 0.000000  loss: 0.1014 (0.1075)  grad: 0.0487 (0.0502)  time: 0.4350  data: 0.0041  max mem: 22446
+train: [19]  [280/400]  eta: 0:00:54  lr: 0.000000  loss: 0.1114 (0.1083)  grad: 0.0490 (0.0503)  time: 0.4337  data: 0.0044  max mem: 22446
+train: [19]  [300/400]  eta: 0:00:46  lr: 0.000000  loss: 0.1175 (0.1086)  grad: 0.0496 (0.0504)  time: 0.6138  data: 0.1636  max mem: 22446
+train: [19]  [320/400]  eta: 0:00:36  lr: 0.000000  loss: 0.1123 (0.1087)  grad: 0.0533 (0.0507)  time: 0.4509  data: 0.0031  max mem: 22446
+train: [19]  [340/400]  eta: 0:00:27  lr: 0.000000  loss: 0.1034 (0.1085)  grad: 0.0529 (0.0506)  time: 0.4672  data: 0.0045  max mem: 22446
+train: [19]  [360/400]  eta: 0:00:18  lr: 0.000000  loss: 0.1103 (0.1086)  grad: 0.0513 (0.0508)  time: 0.4527  data: 0.0043  max mem: 22446
+train: [19]  [380/400]  eta: 0:00:09  lr: 0.000000  loss: 0.1103 (0.1086)  grad: 0.0519 (0.0509)  time: 0.4395  data: 0.0039  max mem: 22446
+train: [19]  [399/400]  eta: 0:00:00  lr: 0.000000  loss: 0.1046 (0.1083)  grad: 0.0478 (0.0508)  time: 0.4772  data: 0.0043  max mem: 22446
+train: [19] Total time: 0:03:04 (0.4621 s / it)
+train: [19] Summary: lr: 0.000000  loss: 0.1046 (0.1083)  grad: 0.0478 (0.0508)
+eval (validation): [19]  [ 0/63]  eta: 0:03:10    time: 3.0314  data: 2.7559  max mem: 22446
+eval (validation): [19]  [20/63]  eta: 0:00:20    time: 0.3603  data: 0.0043  max mem: 22446
+eval (validation): [19]  [40/63]  eta: 0:00:09    time: 0.3323  data: 0.0036  max mem: 22446
+eval (validation): [19]  [60/63]  eta: 0:00:01    time: 0.3240  data: 0.0037  max mem: 22446
+eval (validation): [19]  [62/63]  eta: 0:00:00    time: 0.3246  data: 0.0037  max mem: 22446
+eval (validation): [19] Total time: 0:00:24 (0.3858 s / it)
+cv: [19]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.076  acc: 0.981  f1: 0.977
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+evaluating last checkpoint: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-last.pth
+eval model info:
+{"score": 0.9811507936507936, "hparam": [3.7, 1.0], "hparam_id": 32, "epoch": 19, "is_best": false, "best_score": 0.9818948412698413}
+eval (train): [20]  [  0/297]  eta: 0:14:03    time: 2.8402  data: 2.5720  max mem: 22446
+eval (train): [20]  [ 20/297]  eta: 0:02:08    time: 0.3459  data: 0.0030  max mem: 22446
+eval (train): [20]  [ 40/297]  eta: 0:01:47    time: 0.3677  data: 0.0036  max mem: 22446
+eval (train): [20]  [ 60/297]  eta: 0:01:32    time: 0.3317  data: 0.0036  max mem: 22446
+eval (train): [20]  [ 80/297]  eta: 0:01:21    time: 0.3426  data: 0.0039  max mem: 22446
+eval (train): [20]  [100/297]  eta: 0:01:13    time: 0.3420  data: 0.0038  max mem: 22446
+eval (train): [20]  [120/297]  eta: 0:01:04    time: 0.3289  data: 0.0035  max mem: 22446
+eval (train): [20]  [140/297]  eta: 0:00:56    time: 0.3292  data: 0.0035  max mem: 22446
+eval (train): [20]  [160/297]  eta: 0:00:48    time: 0.3368  data: 0.0037  max mem: 22446
+eval (train): [20]  [180/297]  eta: 0:00:41    time: 0.3529  data: 0.0035  max mem: 22446
+eval (train): [20]  [200/297]  eta: 0:00:34    time: 0.3621  data: 0.0034  max mem: 22446
+eval (train): [20]  [220/297]  eta: 0:00:27    time: 0.3351  data: 0.0035  max mem: 22446
+eval (train): [20]  [240/297]  eta: 0:00:20    time: 0.3337  data: 0.0035  max mem: 22446
+eval (train): [20]  [260/297]  eta: 0:00:13    time: 0.3352  data: 0.0032  max mem: 22446
+eval (train): [20]  [280/297]  eta: 0:00:05    time: 0.3446  data: 0.0033  max mem: 22446
+eval (train): [20]  [296/297]  eta: 0:00:00    time: 0.3243  data: 0.0034  max mem: 22446
+eval (train): [20] Total time: 0:01:44 (0.3511 s / it)
+eval (validation): [20]  [ 0/63]  eta: 0:02:56    time: 2.8016  data: 2.5117  max mem: 22446
+eval (validation): [20]  [20/63]  eta: 0:00:19    time: 0.3411  data: 0.0036  max mem: 22446
+eval (validation): [20]  [40/63]  eta: 0:00:09    time: 0.3377  data: 0.0033  max mem: 22446
+eval (validation): [20]  [60/63]  eta: 0:00:01    time: 0.3059  data: 0.0032  max mem: 22446
+eval (validation): [20]  [62/63]  eta: 0:00:00    time: 0.3048  data: 0.0032  max mem: 22446
+eval (validation): [20] Total time: 0:00:23 (0.3708 s / it)
+eval (test): [20]  [ 0/79]  eta: 0:03:41    time: 2.8036  data: 2.5212  max mem: 22446
+eval (test): [20]  [20/79]  eta: 0:00:28    time: 0.3695  data: 0.0035  max mem: 22446
+eval (test): [20]  [40/79]  eta: 0:00:16    time: 0.3750  data: 0.0034  max mem: 22446
+eval (test): [20]  [60/79]  eta: 0:00:07    time: 0.3590  data: 0.0038  max mem: 22446
+eval (test): [20]  [78/79]  eta: 0:00:00    time: 0.3285  data: 0.0035  max mem: 22446
+eval (test): [20] Total time: 0:00:31 (0.3931 s / it)
+evaluating best checkpoint: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/hcpya_task21__patch__attn/checkpoint-best.pth
+eval model info:
+{"score": 0.9818948412698413, "hparam": [3.7, 1.0], "hparam_id": 32, "epoch": 8, "is_best": true, "best_score": 0.9818948412698413}
+eval (train): [20]  [  0/297]  eta: 0:13:35    time: 2.7457  data: 2.4825  max mem: 22446
+eval (train): [20]  [ 20/297]  eta: 0:02:01    time: 0.3248  data: 0.0033  max mem: 22446
+eval (train): [20]  [ 40/297]  eta: 0:01:39    time: 0.3354  data: 0.0028  max mem: 22446
+eval (train): [20]  [ 60/297]  eta: 0:01:30    time: 0.3700  data: 0.0040  max mem: 22446
+eval (train): [20]  [ 80/297]  eta: 0:01:20    time: 0.3426  data: 0.0034  max mem: 22446
+eval (train): [20]  [100/297]  eta: 0:01:11    time: 0.3338  data: 0.0035  max mem: 22446
+eval (train): [20]  [120/297]  eta: 0:01:04    time: 0.3438  data: 0.0033  max mem: 22446
+eval (train): [20]  [140/297]  eta: 0:00:56    time: 0.3522  data: 0.0037  max mem: 22446
+eval (train): [20]  [160/297]  eta: 0:00:48    time: 0.3285  data: 0.0031  max mem: 22446
+eval (train): [20]  [180/297]  eta: 0:00:41    time: 0.3158  data: 0.0031  max mem: 22446
+eval (train): [20]  [200/297]  eta: 0:00:34    time: 0.3725  data: 0.0038  max mem: 22446
+eval (train): [20]  [220/297]  eta: 0:00:27    time: 0.3606  data: 0.0031  max mem: 22446
+eval (train): [20]  [240/297]  eta: 0:00:20    time: 0.3311  data: 0.0033  max mem: 22446
+eval (train): [20]  [260/297]  eta: 0:00:13    time: 0.3466  data: 0.0035  max mem: 22446
+eval (train): [20]  [280/297]  eta: 0:00:05    time: 0.3371  data: 0.0035  max mem: 22446
+eval (train): [20]  [296/297]  eta: 0:00:00    time: 0.3211  data: 0.0036  max mem: 22446
+eval (train): [20] Total time: 0:01:44 (0.3513 s / it)
+eval (validation): [20]  [ 0/63]  eta: 0:02:54    time: 2.7638  data: 2.5012  max mem: 22446
+eval (validation): [20]  [20/63]  eta: 0:00:20    time: 0.3714  data: 0.0042  max mem: 22446
+eval (validation): [20]  [40/63]  eta: 0:00:09    time: 0.3503  data: 0.0035  max mem: 22446
+eval (validation): [20]  [60/63]  eta: 0:00:01    time: 0.3227  data: 0.0034  max mem: 22446
+eval (validation): [20]  [62/63]  eta: 0:00:00    time: 0.3199  data: 0.0034  max mem: 22446
+eval (validation): [20] Total time: 0:00:24 (0.3896 s / it)
+eval (test): [20]  [ 0/79]  eta: 0:03:37    time: 2.7593  data: 2.4966  max mem: 22446
+eval (test): [20]  [20/79]  eta: 0:00:27    time: 0.3426  data: 0.0042  max mem: 22446
+eval (test): [20]  [40/79]  eta: 0:00:15    time: 0.3479  data: 0.0030  max mem: 22446
+eval (test): [20]  [60/79]  eta: 0:00:07    time: 0.3435  data: 0.0036  max mem: 22446
+eval (test): [20]  [78/79]  eta: 0:00:00    time: 0.3242  data: 0.0033  max mem: 22446
+eval (test): [20] Total time: 0:00:29 (0.3732 s / it)
+eval results:
+
+| model    | repr   | clf   | dataset      | ckpt   |   epoch |      lr |   wd |   hparam_id | hparam     | split      |       loss |     acc |   acc_std |      f1 |    f1_std |
+|:---------|:-------|:------|:-------------|:-------|--------:|--------:|-----:|------------:|:-----------|:-----------|-----------:|--------:|----------:|--------:|----------:|
+| flat_mae | patch  | attn  | hcpya_task21 | best   |       8 | 0.00111 | 0.05 |          32 | [3.7, 1.0] | train      | 0.00051376 | 1       | 0         | 1       | 0         |
+| flat_mae | patch  | attn  | hcpya_task21 | best   |       8 | 0.00111 | 0.05 |          32 | [3.7, 1.0] | validation | 0.07408    | 0.98189 | 0.002097  | 0.97814 | 0.0027741 |
+| flat_mae | patch  | attn  | hcpya_task21 | best   |       8 | 0.00111 | 0.05 |          32 | [3.7, 1.0] | test       | 0.074462   | 0.98056 | 0.0019951 | 0.97846 | 0.0024018 |
+
+
+done! total time: 1:17:30

data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/config.yaml

@@ -0,0 +1,96 @@
+output_root: experiments/data_scaling/output
+name_prefix: eval_probe
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (nsd_cococlip patch attn)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: true
+  norm: true
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn
+model: flat_mae
+representation: patch
+classifier: attn
+dataset: nsd_cococlip
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn
+remote_dir: null

data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/eval_log.json

@@ -0,0 +1 @@
+{"eval/epoch": 5, "eval/id_best": 25, "eval/lr_best": 0.00035999999999999997, "eval/wd_best": 0.05, "eval/train/loss": 2.0278518199920654, "eval/train/acc": 0.38605980515688865, "eval/train/acc_std": 0.0024390732770308, "eval/train/f1": 0.32683064710456294, "eval/train/f1_std": 0.0026072174595867434, "eval/validation/loss": 2.3861210346221924, "eval/validation/acc": 0.27593207825765964, "eval/validation/acc_std": 0.005400515837060602, "eval/validation/f1": 0.20648864936047728, "eval/validation/f1_std": 0.004769139133979471, "eval/test/loss": 2.3202176094055176, "eval/test/acc": 0.299443413729128, "eval/test/acc_std": 0.005292847465158215, "eval/test/f1": 0.234522776971707, "eval/test/f1_std": 0.005352633686304655, "eval/testid/loss": 2.275526285171509, "eval/testid/acc": 0.30364372469635625, "eval/testid/acc_std": 0.005979370856113823, "eval/testid/f1": 0.24622010956570986, "eval/testid/f1_std": 0.0057295414352910475}

data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/eval_log_best.json

@@ -0,0 +1 @@
+{"eval/best/epoch": 5, "eval/best/id_best": 25, "eval/best/lr_best": 0.00035999999999999997, "eval/best/wd_best": 0.05, "eval/best/train/loss": 2.0278518199920654, "eval/best/train/acc": 0.38605980515688865, "eval/best/train/acc_std": 0.0024390732770308, "eval/best/train/f1": 0.32683064710456294, "eval/best/train/f1_std": 0.0026072174595867434, "eval/best/validation/loss": 2.3861210346221924, "eval/best/validation/acc": 0.27593207825765964, "eval/best/validation/acc_std": 0.005400515837060602, "eval/best/validation/f1": 0.20648864936047728, "eval/best/validation/f1_std": 0.004769139133979471, "eval/best/test/loss": 2.3202176094055176, "eval/best/test/acc": 0.299443413729128, "eval/best/test/acc_std": 0.005292847465158215, "eval/best/test/f1": 0.234522776971707, "eval/best/test/f1_std": 0.005352633686304655, "eval/best/testid/loss": 2.275526285171509, "eval/best/testid/acc": 0.30364372469635625, "eval/best/testid/acc_std": 0.005979370856113823, "eval/best/testid/f1": 0.24622010956570986, "eval/best/testid/f1_std": 0.0057295414352910475}

data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/eval_log_last.json

@@ -0,0 +1 @@
+{"eval/last/epoch": 19, "eval/last/id_best": 19, "eval/last/lr_best": 0.00013199999999999998, "eval/last/wd_best": 0.05, "eval/last/train/loss": 1.8982449769973755, "eval/last/train/acc": 0.4274870155813024, "eval/last/train/acc_std": 0.0024958864613864123, "eval/last/train/f1": 0.37622806674509074, "eval/last/train/f1_std": 0.0027841325931396185, "eval/last/validation/loss": 2.42641019821167, "eval/last/validation/acc": 0.26891842008121075, "eval/last/validation/acc_std": 0.005440777827485976, "eval/last/validation/f1": 0.21265296750433957, "eval/last/validation/f1_std": 0.005197201381780099, "eval/last/test/loss": 2.379150390625, "eval/last/test/acc": 0.29573283858998145, "eval/last/test/acc_std": 0.005501312904378919, "eval/last/test/f1": 0.23091473888547545, "eval/last/test/f1_std": 0.005716840638132547, "eval/last/testid/loss": 2.197214126586914, "eval/last/testid/acc": 0.3227299016772701, "eval/last/testid/acc_std": 0.006143241739032288, "eval/last/testid/f1": 0.2743443587301291, "eval/last/testid/f1_std": 0.006057829224491564}

data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/eval_table.csv

@@ -0,0 +1,5 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,attn,nsd_cococlip,best,5,0.00035999999999999997,0.05,25,"[1.2, 1.0]",train,2.0278518199920654,0.38605980515688865,0.0024390732770308,0.32683064710456294,0.0026072174595867434
+flat_mae,patch,attn,nsd_cococlip,best,5,0.00035999999999999997,0.05,25,"[1.2, 1.0]",validation,2.3861210346221924,0.27593207825765964,0.005400515837060602,0.20648864936047728,0.004769139133979471
+flat_mae,patch,attn,nsd_cococlip,best,5,0.00035999999999999997,0.05,25,"[1.2, 1.0]",test,2.3202176094055176,0.299443413729128,0.005292847465158215,0.234522776971707,0.005352633686304655
+flat_mae,patch,attn,nsd_cococlip,best,5,0.00035999999999999997,0.05,25,"[1.2, 1.0]",testid,2.275526285171509,0.30364372469635625,0.005979370856113823,0.24622010956570986,0.0057295414352910475

data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/eval_table_best.csv

@@ -0,0 +1,5 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,attn,nsd_cococlip,best,5,0.00035999999999999997,0.05,25,"[1.2, 1.0]",train,2.0278518199920654,0.38605980515688865,0.0024390732770308,0.32683064710456294,0.0026072174595867434
+flat_mae,patch,attn,nsd_cococlip,best,5,0.00035999999999999997,0.05,25,"[1.2, 1.0]",validation,2.3861210346221924,0.27593207825765964,0.005400515837060602,0.20648864936047728,0.004769139133979471
+flat_mae,patch,attn,nsd_cococlip,best,5,0.00035999999999999997,0.05,25,"[1.2, 1.0]",test,2.3202176094055176,0.299443413729128,0.005292847465158215,0.234522776971707,0.005352633686304655
+flat_mae,patch,attn,nsd_cococlip,best,5,0.00035999999999999997,0.05,25,"[1.2, 1.0]",testid,2.275526285171509,0.30364372469635625,0.005979370856113823,0.24622010956570986,0.0057295414352910475

data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/eval_table_last.csv

@@ -0,0 +1,5 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,attn,nsd_cococlip,last,19,0.00013199999999999998,0.05,19,"[0.44, 1.0]",train,1.8982449769973755,0.4274870155813024,0.0024958864613864123,0.37622806674509074,0.0027841325931396185
+flat_mae,patch,attn,nsd_cococlip,last,19,0.00013199999999999998,0.05,19,"[0.44, 1.0]",validation,2.42641019821167,0.26891842008121075,0.005440777827485976,0.21265296750433957,0.005197201381780099
+flat_mae,patch,attn,nsd_cococlip,last,19,0.00013199999999999998,0.05,19,"[0.44, 1.0]",test,2.379150390625,0.29573283858998145,0.005501312904378919,0.23091473888547545,0.005716840638132547
+flat_mae,patch,attn,nsd_cococlip,last,19,0.00013199999999999998,0.05,19,"[0.44, 1.0]",testid,2.197214126586914,0.3227299016772701,0.006143241739032288,0.2743443587301291,0.006057829224491564

data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/log.txt

@@ -0,0 +1,960 @@
+fMRI foundation model probe eval
+version: 0.1.dev65+g4003a1397
+sha: 6c01b606db98add5848cecd23e5d599250c0bf86, status: clean, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-02-24 19:39:50
+config:
+output_root: experiments/data_scaling/output
+name_prefix: eval_probe
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (nsd_cococlip patch attn)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: true
+  norm: true
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn
+model: flat_mae
+representation: patch
+classifier: attn
+dataset: nsd_cococlip
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=True, reg_tokens=0, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: nsd_cococlip (flat)
+train (n=32539):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'run', 'trial_id', 'nsd_id', 'category_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 32539
+}),
+    labels=[ 3  4  5  6 10 11 17 18 19 20 22 23 25 30 31 33 36 37 38 53 55 59 61 74],
+    counts=[1286 1180 1639 1868  834  824 1026 1042  913 1853 1503 2092 1001 1410
+  794 1241 1904 1872 2267 1428  889  904 1447 1322]
+)
+
+validation (n=5418):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'run', 'trial_id', 'nsd_id', 'category_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 5418
+}),
+    labels=[ 3  4  5  6 10 11 17 18 19 20 22 23 25 30 31 33 36 37 38 53 55 59 61 74],
+    counts=[197 161 276 345 126 142 143 185 112 295 285 387 169 250 159 193 316 334
+ 343 215 172 141 226 246]
+)
+
+test (n=5390):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'run', 'trial_id', 'nsd_id', 'category_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 5390
+}),
+    labels=[ 3  4  5  6 10 11 17 18 19 20 22 23 25 30 31 33 36 37 38 53 55 59 61 74],
+    counts=[202 172 274 298 144 180 134 182 186 293 218 343 165 185 140 177 346 333
+ 345 271 165 140 251 246]
+)
+
+testid (n=5187):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'run', 'trial_id', 'nsd_id', 'category_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 5187
+}),
+    labels=[ 3  4  5  6 10 11 17 18 19 20 22 23 25 30 31 33 36 37 38 53 55 59 61 74],
+    counts=[197 159 267 273 123 153 175 184 139 310 215 386 153 230 118 192 330 306
+ 349 223 143 127 249 186]
+)
+
+running backbone on example batch to get embedding dim
+embedding feature dim (patch): 768
+initializing sweep of classifier heads
+classifiers:
+ModuleList(
+  (0-48): 49 x AttnPoolClassifier(
+    (kv): Linear(in_features=768, out_features=1536, bias=True)
+    (linear): Linear(in_features=768, out_features=24, bias=True)
+  )
+)
+classifier params (train): 58.8M (58.8M)
+setting up optimizer
+total batch size: 128 = 64 bs per gpu x 2 accum
+lr: 3.00e-04
+full schedule: epochs = 20 (steps = 4000) (decay = True)
+warmup: epochs = 5 (steps = 1000)
+start training for 20 epochs
+train: [0]  [  0/400]  eta: 0:22:14  lr: nan  time: 3.3366  data: 2.8073  max mem: 21740
+train: [0]  [ 20/400]  eta: 0:03:38  lr: 0.000003  loss: 3.1883 (3.1901)  grad: 0.1806 (0.1810)  time: 0.4379  data: 0.0029  max mem: 22448
+train: [0]  [ 40/400]  eta: 0:03:03  lr: 0.000006  loss: 3.1840 (3.1820)  grad: 0.1762 (0.1801)  time: 0.4427  data: 0.0048  max mem: 22448
+train: [0]  [ 60/400]  eta: 0:02:47  lr: 0.000009  loss: 3.1737 (3.1817)  grad: 0.1732 (0.1793)  time: 0.4558  data: 0.0048  max mem: 22448
+train: [0]  [ 80/400]  eta: 0:02:33  lr: 0.000012  loss: 3.1650 (3.1763)  grad: 0.1702 (0.1774)  time: 0.4372  data: 0.0047  max mem: 22448
+train: [0]  [100/400]  eta: 0:02:22  lr: 0.000015  loss: 3.1619 (3.1741)  grad: 0.1662 (0.1756)  time: 0.4530  data: 0.0050  max mem: 22448
+train: [0]  [120/400]  eta: 0:02:12  lr: 0.000018  loss: 3.1607 (3.1716)  grad: 0.1584 (0.1729)  time: 0.4633  data: 0.0050  max mem: 22448
+train: [0]  [140/400]  eta: 0:02:02  lr: 0.000021  loss: 3.1513 (3.1683)  grad: 0.1596 (0.1720)  time: 0.4565  data: 0.0049  max mem: 22448
+train: [0]  [160/400]  eta: 0:01:52  lr: 0.000024  loss: 3.1438 (3.1649)  grad: 0.1735 (0.1733)  time: 0.4510  data: 0.0049  max mem: 22448
+train: [0]  [180/400]  eta: 0:01:42  lr: 0.000027  loss: 3.1336 (3.1619)  grad: 0.1774 (0.1727)  time: 0.4319  data: 0.0046  max mem: 22448
+train: [0]  [200/400]  eta: 0:01:32  lr: 0.000030  loss: 3.1289 (3.1600)  grad: 0.1588 (0.1712)  time: 0.4750  data: 0.0051  max mem: 22448
+train: [0]  [220/400]  eta: 0:01:23  lr: 0.000033  loss: 3.1487 (3.1584)  grad: 0.1604 (0.1706)  time: 0.4548  data: 0.0049  max mem: 22448
+train: [0]  [240/400]  eta: 0:01:13  lr: 0.000036  loss: 3.1316 (3.1554)  grad: 0.1639 (0.1699)  time: 0.4328  data: 0.0048  max mem: 22448
+train: [0]  [260/400]  eta: 0:01:04  lr: 0.000039  loss: 3.1196 (3.1525)  grad: 0.1601 (0.1690)  time: 0.4596  data: 0.0049  max mem: 22448
+train: [0]  [280/400]  eta: 0:00:55  lr: 0.000042  loss: 3.0968 (3.1483)  grad: 0.1583 (0.1685)  time: 0.4584  data: 0.0050  max mem: 22448
+train: [0]  [300/400]  eta: 0:00:46  lr: 0.000045  loss: 3.0760 (3.1425)  grad: 0.1613 (0.1682)  time: 0.4466  data: 0.0050  max mem: 22448
+train: [0]  [320/400]  eta: 0:00:36  lr: 0.000048  loss: 3.0616 (3.1385)  grad: 0.1700 (0.1688)  time: 0.4513  data: 0.0048  max mem: 22448
+train: [0]  [340/400]  eta: 0:00:27  lr: 0.000051  loss: 3.0707 (3.1349)  grad: 0.1717 (0.1690)  time: 0.4746  data: 0.0054  max mem: 22448
+train: [0]  [360/400]  eta: 0:00:18  lr: 0.000054  loss: 3.0616 (3.1304)  grad: 0.1717 (0.1697)  time: 0.4507  data: 0.0050  max mem: 22448
+train: [0]  [380/400]  eta: 0:00:09  lr: 0.000057  loss: 3.0527 (3.1265)  grad: 0.1858 (0.1704)  time: 0.4695  data: 0.0050  max mem: 22448
+train: [0]  [399/400]  eta: 0:00:00  lr: 0.000060  loss: 3.0484 (3.1233)  grad: 0.1860 (0.1710)  time: 0.4558  data: 0.0048  max mem: 22448
+train: [0] Total time: 0:03:04 (0.4606 s / it)
+train: [0] Summary: lr: 0.000060  loss: 3.0484 (3.1233)  grad: 0.1860 (0.1710)
+eval (validation): [0]  [ 0/85]  eta: 0:05:35    time: 3.9480  data: 3.6305  max mem: 22448
+eval (validation): [0]  [20/85]  eta: 0:00:37    time: 0.4150  data: 0.0047  max mem: 22448
+eval (validation): [0]  [40/85]  eta: 0:00:21    time: 0.3580  data: 0.0042  max mem: 22448
+eval (validation): [0]  [60/85]  eta: 0:00:10    time: 0.3478  data: 0.0038  max mem: 22448
+eval (validation): [0]  [80/85]  eta: 0:00:02    time: 0.3310  data: 0.0042  max mem: 22448
+eval (validation): [0]  [84/85]  eta: 0:00:00    time: 0.3196  data: 0.0040  max mem: 22448
+eval (validation): [0] Total time: 0:00:34 (0.4056 s / it)
+cv: [0]  best hparam: (31, 1.0) (045) ('045_lr3.1e+01_wd1.0e+00')  loss: 2.589  acc: 0.225  f1: 0.151
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-best.pth
+train: [1]  [  0/400]  eta: 0:22:52  lr: nan  time: 3.4300  data: 3.0773  max mem: 22448
+train: [1]  [ 20/400]  eta: 0:03:43  lr: 0.000063  loss: 2.9996 (3.0046)  grad: 0.1796 (0.1771)  time: 0.4456  data: 0.0045  max mem: 22448
+train: [1]  [ 40/400]  eta: 0:03:08  lr: 0.000066  loss: 3.0096 (3.0107)  grad: 0.1719 (0.1736)  time: 0.4536  data: 0.0045  max mem: 22448
+train: [1]  [ 60/400]  eta: 0:02:49  lr: 0.000069  loss: 2.9983 (2.9962)  grad: 0.1719 (0.1742)  time: 0.4512  data: 0.0047  max mem: 22448
+train: [1]  [ 80/400]  eta: 0:02:35  lr: 0.000072  loss: 2.9770 (2.9964)  grad: 0.1748 (0.1765)  time: 0.4475  data: 0.0046  max mem: 22448
+train: [1]  [100/400]  eta: 0:02:24  lr: 0.000075  loss: 2.9759 (2.9889)  grad: 0.1750 (0.1776)  time: 0.4596  data: 0.0048  max mem: 22448
+train: [1]  [120/400]  eta: 0:02:13  lr: 0.000078  loss: 2.9624 (2.9863)  grad: 0.1860 (0.1791)  time: 0.4513  data: 0.0049  max mem: 22448
+train: [1]  [140/400]  eta: 0:02:03  lr: 0.000081  loss: 2.9684 (2.9834)  grad: 0.1860 (0.1807)  time: 0.4654  data: 0.0048  max mem: 22448
+train: [1]  [160/400]  eta: 0:01:52  lr: 0.000084  loss: 2.9707 (2.9837)  grad: 0.1858 (0.1813)  time: 0.4429  data: 0.0046  max mem: 22448
+train: [1]  [180/400]  eta: 0:01:43  lr: 0.000087  loss: 2.9864 (2.9836)  grad: 0.1849 (0.1820)  time: 0.4501  data: 0.0048  max mem: 22448
+train: [1]  [200/400]  eta: 0:01:33  lr: 0.000090  loss: 2.9720 (2.9809)  grad: 0.1850 (0.1829)  time: 0.4489  data: 0.0046  max mem: 22448
+train: [1]  [220/400]  eta: 0:01:23  lr: 0.000093  loss: 2.9133 (2.9745)  grad: 0.1954 (0.1849)  time: 0.4529  data: 0.0047  max mem: 22448
+train: [1]  [240/400]  eta: 0:01:14  lr: 0.000096  loss: 2.9033 (2.9708)  grad: 0.1957 (0.1852)  time: 0.4613  data: 0.0049  max mem: 22448
+train: [1]  [260/400]  eta: 0:01:05  lr: 0.000099  loss: 2.9244 (2.9687)  grad: 0.1857 (0.1860)  time: 0.4586  data: 0.0048  max mem: 22448
+train: [1]  [280/400]  eta: 0:00:55  lr: 0.000102  loss: 2.9143 (2.9651)  grad: 0.1856 (0.1861)  time: 0.4465  data: 0.0048  max mem: 22448
+train: [1]  [300/400]  eta: 0:00:46  lr: 0.000105  loss: 2.9062 (2.9625)  grad: 0.1893 (0.1867)  time: 0.4318  data: 0.0047  max mem: 22448
+train: [1]  [320/400]  eta: 0:00:36  lr: 0.000108  loss: 2.9003 (2.9587)  grad: 0.1992 (0.1876)  time: 0.4469  data: 0.0048  max mem: 22448
+train: [1]  [340/400]  eta: 0:00:27  lr: 0.000111  loss: 2.8780 (2.9536)  grad: 0.1962 (0.1880)  time: 0.4417  data: 0.0047  max mem: 22448
+train: [1]  [360/400]  eta: 0:00:18  lr: 0.000114  loss: 2.8985 (2.9519)  grad: 0.1931 (0.1883)  time: 0.4402  data: 0.0047  max mem: 22448
+train: [1]  [380/400]  eta: 0:00:09  lr: 0.000117  loss: 2.8985 (2.9486)  grad: 0.1931 (0.1890)  time: 0.4456  data: 0.0047  max mem: 22448
+train: [1]  [399/400]  eta: 0:00:00  lr: 0.000120  loss: 2.8838 (2.9467)  grad: 0.2055 (0.1901)  time: 0.4575  data: 0.0048  max mem: 22448
+train: [1] Total time: 0:03:03 (0.4579 s / it)
+train: [1] Summary: lr: 0.000120  loss: 2.8838 (2.9467)  grad: 0.2055 (0.1901)
+eval (validation): [1]  [ 0/85]  eta: 0:04:16    time: 3.0171  data: 2.7833  max mem: 22448
+eval (validation): [1]  [20/85]  eta: 0:00:29    time: 0.3303  data: 0.0041  max mem: 22448
+eval (validation): [1]  [40/85]  eta: 0:00:17    time: 0.3278  data: 0.0039  max mem: 22448
+eval (validation): [1]  [60/85]  eta: 0:00:09    time: 0.3310  data: 0.0036  max mem: 22448
+eval (validation): [1]  [80/85]  eta: 0:00:01    time: 0.3287  data: 0.0038  max mem: 22448
+eval (validation): [1]  [84/85]  eta: 0:00:00    time: 0.3261  data: 0.0039  max mem: 22448
+eval (validation): [1] Total time: 0:00:31 (0.3648 s / it)
+cv: [1]  best hparam: (16, 1.0) (041) ('041_lr1.6e+01_wd1.0e+00')  loss: 2.508  acc: 0.243  f1: 0.170
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-best.pth
+train: [2]  [  0/400]  eta: 0:22:00  lr: nan  time: 3.3005  data: 2.9118  max mem: 22448
+train: [2]  [ 20/400]  eta: 0:03:36  lr: 0.000123  loss: 2.8776 (2.8744)  grad: 0.2191 (0.2246)  time: 0.4340  data: 0.0036  max mem: 22448
+train: [2]  [ 40/400]  eta: 0:03:02  lr: 0.000126  loss: 2.8848 (2.8794)  grad: 0.2180 (0.2209)  time: 0.4428  data: 0.0045  max mem: 22448
+train: [2]  [ 60/400]  eta: 0:02:46  lr: 0.000129  loss: 2.8834 (2.8752)  grad: 0.2117 (0.2159)  time: 0.4532  data: 0.0048  max mem: 22448
+train: [2]  [ 80/400]  eta: 0:02:34  lr: 0.000132  loss: 2.8684 (2.8699)  grad: 0.2023 (0.2157)  time: 0.4555  data: 0.0049  max mem: 22448
+train: [2]  [100/400]  eta: 0:02:22  lr: 0.000135  loss: 2.8385 (2.8633)  grad: 0.2126 (0.2160)  time: 0.4484  data: 0.0049  max mem: 22448
+train: [2]  [120/400]  eta: 0:02:11  lr: 0.000138  loss: 2.8265 (2.8557)  grad: 0.2188 (0.2168)  time: 0.4501  data: 0.0051  max mem: 22448
+train: [2]  [140/400]  eta: 0:02:01  lr: 0.000141  loss: 2.8332 (2.8569)  grad: 0.2201 (0.2182)  time: 0.4467  data: 0.0048  max mem: 22448
+train: [2]  [160/400]  eta: 0:01:51  lr: 0.000144  loss: 2.8271 (2.8515)  grad: 0.2235 (0.2195)  time: 0.4508  data: 0.0050  max mem: 22448
+train: [2]  [180/400]  eta: 0:01:41  lr: 0.000147  loss: 2.8043 (2.8455)  grad: 0.2207 (0.2196)  time: 0.4491  data: 0.0051  max mem: 22448
+train: [2]  [200/400]  eta: 0:01:32  lr: 0.000150  loss: 2.8043 (2.8418)  grad: 0.2235 (0.2209)  time: 0.4410  data: 0.0047  max mem: 22448
+train: [2]  [220/400]  eta: 0:01:23  lr: 0.000153  loss: 2.8318 (2.8429)  grad: 0.2400 (0.2249)  time: 0.4632  data: 0.0048  max mem: 22448
+train: [2]  [240/400]  eta: 0:01:13  lr: 0.000156  loss: 2.9222 (2.8690)  grad: 0.3131 (0.2714)  time: 0.4531  data: 0.0051  max mem: 22448
+WARNING: classifier 48 (50, 1.0) diverged (loss=96.71 > 63.56) at step 524. Freezing.
+train: [2]  [260/400]  eta: 0:01:04  lr: 0.000159  loss: 2.9378 (2.8968)  grad: 0.5387 (0.3075)  time: 0.4488  data: 0.0047  max mem: 22448
+train: [2]  [280/400]  eta: 0:00:55  lr: 0.000162  loss: 2.8132 (2.8907)  grad: 0.2317 (0.3026)  time: 0.4429  data: 0.0046  max mem: 22448
+train: [2]  [300/400]  eta: 0:00:45  lr: 0.000165  loss: 2.7971 (2.8842)  grad: 0.2318 (0.2979)  time: 0.4577  data: 0.0049  max mem: 22448
+train: [2]  [320/400]  eta: 0:00:36  lr: 0.000168  loss: 2.7953 (2.8786)  grad: 0.2285 (0.2932)  time: 0.4508  data: 0.0049  max mem: 22448
+train: [2]  [340/400]  eta: 0:00:27  lr: 0.000171  loss: 2.7861 (2.8743)  grad: 0.2230 (0.2895)  time: 0.4461  data: 0.0048  max mem: 22448
+train: [2]  [360/400]  eta: 0:00:18  lr: 0.000174  loss: 2.7806 (2.8715)  grad: 0.2266 (0.2862)  time: 0.4447  data: 0.0049  max mem: 22448
+train: [2]  [380/400]  eta: 0:00:09  lr: 0.000177  loss: 2.8250 (2.8700)  grad: 0.2425 (0.2842)  time: 0.4726  data: 0.0052  max mem: 22448
+train: [2]  [399/400]  eta: 0:00:00  lr: 0.000180  loss: 2.8488 (2.8662)  grad: 0.2459 (0.2817)  time: 0.4746  data: 0.0052  max mem: 22448
+train: [2] Total time: 0:03:03 (0.4589 s / it)
+train: [2] Summary: lr: 0.000180  loss: 2.8488 (2.8662)  grad: 0.2459 (0.2817)
+eval (validation): [2]  [ 0/85]  eta: 0:04:28    time: 3.1538  data: 2.8819  max mem: 22448
+eval (validation): [2]  [20/85]  eta: 0:00:31    time: 0.3508  data: 0.0162  max mem: 22448
+eval (validation): [2]  [40/85]  eta: 0:00:18    time: 0.3280  data: 0.0033  max mem: 22448
+eval (validation): [2]  [60/85]  eta: 0:00:09    time: 0.3503  data: 0.0040  max mem: 22448
+eval (validation): [2]  [80/85]  eta: 0:00:01    time: 0.3308  data: 0.0037  max mem: 22448
+eval (validation): [2]  [84/85]  eta: 0:00:00    time: 0.3175  data: 0.0037  max mem: 22448
+eval (validation): [2] Total time: 0:00:31 (0.3742 s / it)
+cv: [2]  best hparam: (2.7, 1.0) (030) ('030_lr2.7e+00_wd1.0e+00')  loss: 2.426  acc: 0.269  f1: 0.192
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-best.pth
+train: [3]  [  0/400]  eta: 0:20:19  lr: nan  time: 3.0489  data: 2.6761  max mem: 22448
+train: [3]  [ 20/400]  eta: 0:03:28  lr: 0.000183  loss: 2.7121 (2.7546)  grad: 0.2356 (0.2464)  time: 0.4235  data: 0.0033  max mem: 22448
+train: [3]  [ 40/400]  eta: 0:02:56  lr: 0.000186  loss: 2.8201 (2.8429)  grad: 0.2805 (0.3646)  time: 0.4317  data: 0.0042  max mem: 22448
+WARNING: classifier 47 (43, 1.0) diverged (loss=69.06 > 63.56) at step 627. Freezing.
+train: [3]  [ 60/400]  eta: 0:02:40  lr: 0.000189  loss: 2.8856 (3.0187)  grad: 0.4626 (0.5799)  time: 0.4357  data: 0.0045  max mem: 22448
+train: [3]  [ 80/400]  eta: 0:02:28  lr: 0.000192  loss: 2.8063 (2.9573)  grad: 0.2326 (0.4905)  time: 0.4332  data: 0.0046  max mem: 22448
+train: [3]  [100/400]  eta: 0:02:17  lr: 0.000195  loss: 2.7426 (2.9163)  grad: 0.2192 (0.4379)  time: 0.4428  data: 0.0048  max mem: 22448
+train: [3]  [120/400]  eta: 0:02:07  lr: 0.000198  loss: 2.7420 (2.8890)  grad: 0.2282 (0.4043)  time: 0.4398  data: 0.0048  max mem: 22448
+train: [3]  [140/400]  eta: 0:01:57  lr: 0.000201  loss: 2.7503 (2.8728)  grad: 0.2531 (0.3832)  time: 0.4397  data: 0.0047  max mem: 22448
+train: [3]  [160/400]  eta: 0:01:48  lr: 0.000204  loss: 2.7934 (2.8644)  grad: 0.2556 (0.3664)  time: 0.4364  data: 0.0047  max mem: 22448
+train: [3]  [180/400]  eta: 0:01:39  lr: 0.000207  loss: 2.7679 (2.8514)  grad: 0.2399 (0.3519)  time: 0.4466  data: 0.0049  max mem: 22448
+train: [3]  [200/400]  eta: 0:01:30  lr: 0.000210  loss: 2.7585 (2.8439)  grad: 0.2398 (0.3410)  time: 0.4405  data: 0.0048  max mem: 22448
+train: [3]  [220/400]  eta: 0:01:20  lr: 0.000213  loss: 2.7672 (2.8375)  grad: 0.2468 (0.3324)  time: 0.4391  data: 0.0048  max mem: 22448
+train: [3]  [240/400]  eta: 0:01:11  lr: 0.000216  loss: 2.7500 (2.8302)  grad: 0.2493 (0.3260)  time: 0.4378  data: 0.0047  max mem: 22448
+train: [3]  [260/400]  eta: 0:01:02  lr: 0.000219  loss: 2.7447 (2.8245)  grad: 0.2519 (0.3206)  time: 0.4362  data: 0.0050  max mem: 22448
+train: [3]  [280/400]  eta: 0:00:53  lr: 0.000222  loss: 2.7426 (2.8182)  grad: 0.2466 (0.3159)  time: 0.4394  data: 0.0049  max mem: 22448
+train: [3]  [300/400]  eta: 0:00:44  lr: 0.000225  loss: 2.8179 (2.8213)  grad: 0.2786 (0.3224)  time: 0.4455  data: 0.0047  max mem: 22448
+WARNING: classifier 46 (36, 1.0) diverged (loss=80.14 > 63.56) at step 759. Freezing.
+train: [3]  [320/400]  eta: 0:00:35  lr: 0.000228  loss: 2.9803 (2.8536)  grad: 0.5460 (0.3745)  time: 0.4416  data: 0.0048  max mem: 22448
+train: [3]  [340/400]  eta: 0:00:26  lr: 0.000231  loss: 2.8070 (2.8494)  grad: 0.2724 (0.3671)  time: 0.4427  data: 0.0050  max mem: 22448
+train: [3]  [360/400]  eta: 0:00:17  lr: 0.000234  loss: 2.7943 (2.8477)  grad: 0.2528 (0.3609)  time: 0.4492  data: 0.0049  max mem: 22448
+train: [3]  [380/400]  eta: 0:00:08  lr: 0.000237  loss: 2.7834 (2.8432)  grad: 0.2603 (0.3556)  time: 0.4722  data: 0.0049  max mem: 22448
+train: [3]  [399/400]  eta: 0:00:00  lr: 0.000240  loss: 2.7351 (2.8381)  grad: 0.2533 (0.3500)  time: 0.4545  data: 0.0050  max mem: 22448
+train: [3] Total time: 0:02:59 (0.4484 s / it)
+train: [3] Summary: lr: 0.000240  loss: 2.7351 (2.8381)  grad: 0.2533 (0.3500)
+eval (validation): [3]  [ 0/85]  eta: 0:04:32    time: 3.2020  data: 2.9605  max mem: 22448
+eval (validation): [3]  [20/85]  eta: 0:00:30    time: 0.3325  data: 0.0047  max mem: 22448
+eval (validation): [3]  [40/85]  eta: 0:00:18    time: 0.3399  data: 0.0035  max mem: 22448
+eval (validation): [3]  [60/85]  eta: 0:00:09    time: 0.3374  data: 0.0043  max mem: 22448
+eval (validation): [3]  [80/85]  eta: 0:00:01    time: 0.3303  data: 0.0041  max mem: 22448
+eval (validation): [3]  [84/85]  eta: 0:00:00    time: 0.3187  data: 0.0040  max mem: 22448
+eval (validation): [3] Total time: 0:00:31 (0.3704 s / it)
+cv: [3]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 2.433  acc: 0.267  f1: 0.197
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [4]  [  0/400]  eta: 0:21:37  lr: nan  time: 3.2443  data: 2.8683  max mem: 22448
+train: [4]  [ 20/400]  eta: 0:03:35  lr: 0.000243  loss: 2.6687 (2.6921)  grad: 0.2634 (0.2632)  time: 0.4337  data: 0.0039  max mem: 22448
+train: [4]  [ 40/400]  eta: 0:03:03  lr: 0.000246  loss: 2.6992 (2.7144)  grad: 0.2798 (0.2969)  time: 0.4508  data: 0.0049  max mem: 22448
+train: [4]  [ 60/400]  eta: 0:02:45  lr: 0.000249  loss: 2.8284 (2.8595)  grad: 0.4322 (0.5396)  time: 0.4408  data: 0.0049  max mem: 22448
+WARNING: classifier 45 (31, 1.0) diverged (loss=106.04 > 63.56) at step 833. Freezing.
+train: [4]  [ 80/400]  eta: 0:02:32  lr: 0.000252  loss: 2.8935 (2.9191)  grad: 0.7091 (0.5799)  time: 0.4421  data: 0.0048  max mem: 22448
+train: [4]  [100/400]  eta: 0:02:20  lr: 0.000255  loss: 2.7288 (2.8832)  grad: 0.2369 (0.5116)  time: 0.4385  data: 0.0051  max mem: 22448
+train: [4]  [120/400]  eta: 0:02:10  lr: 0.000258  loss: 2.6861 (2.8500)  grad: 0.2369 (0.4668)  time: 0.4502  data: 0.0050  max mem: 22448
+train: [4]  [140/400]  eta: 0:02:00  lr: 0.000261  loss: 2.7057 (2.8290)  grad: 0.2467 (0.4371)  time: 0.4449  data: 0.0049  max mem: 22448
+train: [4]  [160/400]  eta: 0:01:50  lr: 0.000264  loss: 2.7103 (2.8158)  grad: 0.2581 (0.4150)  time: 0.4370  data: 0.0050  max mem: 22448
+train: [4]  [180/400]  eta: 0:01:40  lr: 0.000267  loss: 2.7202 (2.8088)  grad: 0.2605 (0.3980)  time: 0.4368  data: 0.0049  max mem: 22448
+train: [4]  [200/400]  eta: 0:01:31  lr: 0.000270  loss: 2.7040 (2.7969)  grad: 0.2534 (0.3828)  time: 0.4520  data: 0.0051  max mem: 22448
+train: [4]  [220/400]  eta: 0:01:21  lr: 0.000273  loss: 2.7144 (2.7939)  grad: 0.2534 (0.3719)  time: 0.4377  data: 0.0049  max mem: 22448
+train: [4]  [240/400]  eta: 0:01:12  lr: 0.000276  loss: 2.7497 (2.7880)  grad: 0.2572 (0.3624)  time: 0.4378  data: 0.0048  max mem: 22448
+train: [4]  [260/400]  eta: 0:01:03  lr: 0.000279  loss: 2.7410 (2.7833)  grad: 0.2572 (0.3545)  time: 0.4401  data: 0.0049  max mem: 22448
+train: [4]  [280/400]  eta: 0:00:54  lr: 0.000282  loss: 2.7432 (2.7794)  grad: 0.2673 (0.3502)  time: 0.4411  data: 0.0050  max mem: 22448
+train: [4]  [300/400]  eta: 0:00:45  lr: 0.000285  loss: 2.7884 (2.7925)  grad: 0.3252 (0.3780)  time: 0.4499  data: 0.0049  max mem: 22448
+WARNING: classifier 44 (26, 1.0) diverged (loss=74.89 > 63.56) at step 954. Freezing.
+train: [4]  [320/400]  eta: 0:00:36  lr: 0.000288  loss: 2.8622 (2.8110)  grad: 0.5872 (0.4058)  time: 0.4411  data: 0.0048  max mem: 22448
+train: [4]  [340/400]  eta: 0:00:27  lr: 0.000291  loss: 2.6996 (2.8043)  grad: 0.2199 (0.3951)  time: 0.4389  data: 0.0049  max mem: 22448
+train: [4]  [360/400]  eta: 0:00:18  lr: 0.000294  loss: 2.6987 (2.8004)  grad: 0.2269 (0.3862)  time: 0.4485  data: 0.0050  max mem: 22448
+train: [4]  [380/400]  eta: 0:00:08  lr: 0.000297  loss: 2.7153 (2.7957)  grad: 0.2354 (0.3782)  time: 0.4428  data: 0.0049  max mem: 22448
+train: [4]  [399/400]  eta: 0:00:00  lr: 0.000300  loss: 2.6949 (2.7900)  grad: 0.2365 (0.3711)  time: 0.4613  data: 0.0050  max mem: 22448
+train: [4] Total time: 0:03:00 (0.4509 s / it)
+train: [4] Summary: lr: 0.000300  loss: 2.6949 (2.7900)  grad: 0.2365 (0.3711)
+eval (validation): [4]  [ 0/85]  eta: 0:04:15    time: 3.0084  data: 2.7376  max mem: 22448
+eval (validation): [4]  [20/85]  eta: 0:00:32    time: 0.3706  data: 0.0044  max mem: 22448
+eval (validation): [4]  [40/85]  eta: 0:00:19    time: 0.3540  data: 0.0042  max mem: 22448
+eval (validation): [4]  [60/85]  eta: 0:00:09    time: 0.3336  data: 0.0039  max mem: 22448
+eval (validation): [4]  [80/85]  eta: 0:00:01    time: 0.3314  data: 0.0040  max mem: 22448
+eval (validation): [4]  [84/85]  eta: 0:00:00    time: 0.3306  data: 0.0041  max mem: 22448
+eval (validation): [4] Total time: 0:00:32 (0.3816 s / it)
+cv: [4]  best hparam: (1.2, 1.0) (025) ('025_lr1.2e+00_wd1.0e+00')  loss: 2.436  acc: 0.263  f1: 0.194
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [5]  [  0/400]  eta: 0:21:57  lr: nan  time: 3.2934  data: 2.9125  max mem: 22448
+train: [5]  [ 20/400]  eta: 0:03:38  lr: 0.000300  loss: 2.6006 (2.5989)  grad: 0.2409 (0.2432)  time: 0.4402  data: 0.0031  max mem: 22448
+train: [5]  [ 40/400]  eta: 0:03:02  lr: 0.000300  loss: 2.6222 (2.6491)  grad: 0.2436 (0.2474)  time: 0.4330  data: 0.0049  max mem: 22448
+train: [5]  [ 60/400]  eta: 0:02:44  lr: 0.000300  loss: 2.6544 (2.6555)  grad: 0.2533 (0.2505)  time: 0.4359  data: 0.0049  max mem: 22448
+train: [5]  [ 80/400]  eta: 0:02:30  lr: 0.000300  loss: 2.6612 (2.6536)  grad: 0.2473 (0.2502)  time: 0.4345  data: 0.0049  max mem: 22448
+train: [5]  [100/400]  eta: 0:02:19  lr: 0.000300  loss: 2.6631 (2.6555)  grad: 0.2470 (0.2521)  time: 0.4358  data: 0.0047  max mem: 22448
+train: [5]  [120/400]  eta: 0:02:08  lr: 0.000300  loss: 2.6091 (2.6482)  grad: 0.2472 (0.2512)  time: 0.4375  data: 0.0047  max mem: 22448
+train: [5]  [140/400]  eta: 0:01:59  lr: 0.000300  loss: 2.6011 (2.6383)  grad: 0.2404 (0.2490)  time: 0.4573  data: 0.0050  max mem: 22448
+train: [5]  [160/400]  eta: 0:01:49  lr: 0.000299  loss: 2.6106 (2.6379)  grad: 0.2364 (0.2484)  time: 0.4383  data: 0.0050  max mem: 22448
+train: [5]  [180/400]  eta: 0:01:40  lr: 0.000299  loss: 2.6571 (2.6423)  grad: 0.2419 (0.2481)  time: 0.4378  data: 0.0049  max mem: 22448
+train: [5]  [200/400]  eta: 0:01:30  lr: 0.000299  loss: 2.6571 (2.6409)  grad: 0.2527 (0.2485)  time: 0.4326  data: 0.0050  max mem: 22448
+train: [5]  [220/400]  eta: 0:01:21  lr: 0.000299  loss: 2.5830 (2.6407)  grad: 0.2527 (0.2480)  time: 0.4515  data: 0.0049  max mem: 22448
+train: [5]  [240/400]  eta: 0:01:12  lr: 0.000299  loss: 2.6052 (2.6409)  grad: 0.2485 (0.2484)  time: 0.4387  data: 0.0047  max mem: 22448
+train: [5]  [260/400]  eta: 0:01:02  lr: 0.000299  loss: 2.6141 (2.6374)  grad: 0.2480 (0.2480)  time: 0.4292  data: 0.0047  max mem: 22448
+train: [5]  [280/400]  eta: 0:00:53  lr: 0.000298  loss: 2.6191 (2.6391)  grad: 0.2481 (0.2488)  time: 0.4399  data: 0.0048  max mem: 22448
+train: [5]  [300/400]  eta: 0:00:44  lr: 0.000298  loss: 2.6191 (2.6359)  grad: 0.2566 (0.2492)  time: 0.4349  data: 0.0048  max mem: 22448
+train: [5]  [320/400]  eta: 0:00:35  lr: 0.000298  loss: 2.6067 (2.6379)  grad: 0.2570 (0.2499)  time: 0.4329  data: 0.0049  max mem: 22448
+train: [5]  [340/400]  eta: 0:00:26  lr: 0.000298  loss: 2.6491 (2.6365)  grad: 0.2548 (0.2499)  time: 0.4390  data: 0.0050  max mem: 22448
+train: [5]  [360/400]  eta: 0:00:17  lr: 0.000297  loss: 2.6181 (2.6352)  grad: 0.2571 (0.2504)  time: 0.4488  data: 0.0050  max mem: 22448
+train: [5]  [380/400]  eta: 0:00:08  lr: 0.000297  loss: 2.5821 (2.6352)  grad: 0.2578 (0.2510)  time: 0.4489  data: 0.0049  max mem: 22448
+train: [5]  [399/400]  eta: 0:00:00  lr: 0.000297  loss: 2.5683 (2.6318)  grad: 0.2510 (0.2506)  time: 0.4476  data: 0.0049  max mem: 22448
+train: [5] Total time: 0:02:58 (0.4474 s / it)
+train: [5] Summary: lr: 0.000297  loss: 2.5683 (2.6318)  grad: 0.2510 (0.2506)
+eval (validation): [5]  [ 0/85]  eta: 0:04:18    time: 3.0387  data: 2.8027  max mem: 22448
+eval (validation): [5]  [20/85]  eta: 0:00:30    time: 0.3347  data: 0.0030  max mem: 22448
+eval (validation): [5]  [40/85]  eta: 0:00:18    time: 0.3578  data: 0.0035  max mem: 22448
+eval (validation): [5]  [60/85]  eta: 0:00:09    time: 0.3313  data: 0.0040  max mem: 22448
+eval (validation): [5]  [80/85]  eta: 0:00:01    time: 0.3217  data: 0.0036  max mem: 22448
+eval (validation): [5]  [84/85]  eta: 0:00:00    time: 0.3186  data: 0.0036  max mem: 22448
+eval (validation): [5] Total time: 0:00:31 (0.3698 s / it)
+cv: [5]  best hparam: (1.2, 1.0) (025) ('025_lr1.2e+00_wd1.0e+00')  loss: 2.386  acc: 0.276  f1: 0.206
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+saving best checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-best.pth
+train: [6]  [  0/400]  eta: 0:21:07  lr: nan  time: 3.1680  data: 2.8003  max mem: 22448
+train: [6]  [ 20/400]  eta: 0:03:43  lr: 0.000296  loss: 2.5579 (2.5481)  grad: 0.2468 (0.2477)  time: 0.4602  data: 0.0039  max mem: 22448
+train: [6]  [ 40/400]  eta: 0:03:04  lr: 0.000296  loss: 2.5618 (2.5676)  grad: 0.2447 (0.2484)  time: 0.4331  data: 0.0049  max mem: 22448
+train: [6]  [ 60/400]  eta: 0:02:45  lr: 0.000296  loss: 2.5551 (2.5657)  grad: 0.2427 (0.2492)  time: 0.4288  data: 0.0049  max mem: 22448
+train: [6]  [ 80/400]  eta: 0:02:31  lr: 0.000295  loss: 2.5221 (2.5518)  grad: 0.2536 (0.2502)  time: 0.4405  data: 0.0050  max mem: 22448
+train: [6]  [100/400]  eta: 0:02:19  lr: 0.000295  loss: 2.5518 (2.5525)  grad: 0.2562 (0.2518)  time: 0.4339  data: 0.0049  max mem: 22448
+train: [6]  [120/400]  eta: 0:02:09  lr: 0.000295  loss: 2.5527 (2.5559)  grad: 0.2566 (0.2536)  time: 0.4373  data: 0.0048  max mem: 22448
+train: [6]  [140/400]  eta: 0:01:59  lr: 0.000294  loss: 2.5970 (2.5645)  grad: 0.2568 (0.2547)  time: 0.4401  data: 0.0048  max mem: 22448
+train: [6]  [160/400]  eta: 0:01:49  lr: 0.000294  loss: 2.6202 (2.5716)  grad: 0.2585 (0.2548)  time: 0.4520  data: 0.0051  max mem: 22448
+train: [6]  [180/400]  eta: 0:01:40  lr: 0.000293  loss: 2.5789 (2.5696)  grad: 0.2589 (0.2557)  time: 0.4441  data: 0.0050  max mem: 22448
+train: [6]  [200/400]  eta: 0:01:30  lr: 0.000293  loss: 2.5758 (2.5717)  grad: 0.2627 (0.2566)  time: 0.4400  data: 0.0049  max mem: 22448
+train: [6]  [220/400]  eta: 0:01:21  lr: 0.000292  loss: 2.5831 (2.5704)  grad: 0.2657 (0.2579)  time: 0.4338  data: 0.0047  max mem: 22448
+train: [6]  [240/400]  eta: 0:01:12  lr: 0.000292  loss: 2.5930 (2.5715)  grad: 0.2624 (0.2581)  time: 0.4528  data: 0.0051  max mem: 22448
+train: [6]  [260/400]  eta: 0:01:03  lr: 0.000291  loss: 2.5588 (2.5681)  grad: 0.2539 (0.2577)  time: 0.4356  data: 0.0049  max mem: 22448
+train: [6]  [280/400]  eta: 0:00:54  lr: 0.000291  loss: 2.5701 (2.5691)  grad: 0.2500 (0.2577)  time: 0.4351  data: 0.0048  max mem: 22448
+train: [6]  [300/400]  eta: 0:00:44  lr: 0.000290  loss: 2.5914 (2.5700)  grad: 0.2570 (0.2578)  time: 0.4341  data: 0.0048  max mem: 22448
+train: [6]  [320/400]  eta: 0:00:35  lr: 0.000290  loss: 2.5899 (2.5704)  grad: 0.2577 (0.2585)  time: 0.4329  data: 0.0049  max mem: 22448
+train: [6]  [340/400]  eta: 0:00:26  lr: 0.000289  loss: 2.5807 (2.5716)  grad: 0.2608 (0.2584)  time: 0.4309  data: 0.0047  max mem: 22448
+train: [6]  [360/400]  eta: 0:00:17  lr: 0.000288  loss: 2.5439 (2.5688)  grad: 0.2580 (0.2586)  time: 0.4502  data: 0.0047  max mem: 22448
+train: [6]  [380/400]  eta: 0:00:08  lr: 0.000288  loss: 2.5142 (2.5702)  grad: 0.2611 (0.2587)  time: 0.4458  data: 0.0049  max mem: 22448
+train: [6]  [399/400]  eta: 0:00:00  lr: 0.000287  loss: 2.5671 (2.5697)  grad: 0.2524 (0.2582)  time: 0.4473  data: 0.0050  max mem: 22448
+train: [6] Total time: 0:02:59 (0.4477 s / it)
+train: [6] Summary: lr: 0.000287  loss: 2.5671 (2.5697)  grad: 0.2524 (0.2582)
+eval (validation): [6]  [ 0/85]  eta: 0:04:18    time: 3.0385  data: 2.7919  max mem: 22448
+eval (validation): [6]  [20/85]  eta: 0:00:30    time: 0.3459  data: 0.0036  max mem: 22448
+eval (validation): [6]  [40/85]  eta: 0:00:19    time: 0.3683  data: 0.0043  max mem: 22448
+eval (validation): [6]  [60/85]  eta: 0:00:09    time: 0.3336  data: 0.0041  max mem: 22448
+eval (validation): [6]  [80/85]  eta: 0:00:01    time: 0.3240  data: 0.0038  max mem: 22448
+eval (validation): [6]  [84/85]  eta: 0:00:00    time: 0.3197  data: 0.0038  max mem: 22448
+eval (validation): [6] Total time: 0:00:31 (0.3762 s / it)
+cv: [6]  best hparam: (1, 1.0) (024) ('024_lr1.0e+00_wd1.0e+00')  loss: 2.404  acc: 0.273  f1: 0.216
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [7]  [  0/400]  eta: 0:21:01  lr: nan  time: 3.1526  data: 2.8314  max mem: 22448
+train: [7]  [ 20/400]  eta: 0:03:35  lr: 0.000286  loss: 2.4271 (2.4561)  grad: 0.2414 (0.2549)  time: 0.4373  data: 0.0039  max mem: 22448
+train: [7]  [ 40/400]  eta: 0:03:00  lr: 0.000286  loss: 2.4430 (2.4596)  grad: 0.2530 (0.2565)  time: 0.4352  data: 0.0047  max mem: 22448
+train: [7]  [ 60/400]  eta: 0:02:43  lr: 0.000285  loss: 2.4622 (2.4539)  grad: 0.2655 (0.2623)  time: 0.4398  data: 0.0048  max mem: 22448
+train: [7]  [ 80/400]  eta: 0:02:30  lr: 0.000284  loss: 2.4693 (2.4719)  grad: 0.2640 (0.2608)  time: 0.4319  data: 0.0046  max mem: 22448
+train: [7]  [100/400]  eta: 0:02:18  lr: 0.000284  loss: 2.4693 (2.4678)  grad: 0.2605 (0.2611)  time: 0.4322  data: 0.0050  max mem: 22448
+train: [7]  [120/400]  eta: 0:02:08  lr: 0.000283  loss: 2.4484 (2.4688)  grad: 0.2670 (0.2623)  time: 0.4434  data: 0.0051  max mem: 22448
+train: [7]  [140/400]  eta: 0:01:58  lr: 0.000282  loss: 2.4591 (2.4736)  grad: 0.2591 (0.2617)  time: 0.4316  data: 0.0051  max mem: 22448
+train: [7]  [160/400]  eta: 0:01:49  lr: 0.000282  loss: 2.4815 (2.4740)  grad: 0.2578 (0.2614)  time: 0.4587  data: 0.0048  max mem: 22448
+train: [7]  [180/400]  eta: 0:01:39  lr: 0.000281  loss: 2.5295 (2.4829)  grad: 0.2668 (0.2628)  time: 0.4453  data: 0.0050  max mem: 22448
+train: [7]  [200/400]  eta: 0:01:30  lr: 0.000280  loss: 2.5272 (2.4825)  grad: 0.2668 (0.2628)  time: 0.4510  data: 0.0050  max mem: 22448
+train: [7]  [220/400]  eta: 0:01:21  lr: 0.000279  loss: 2.4707 (2.4810)  grad: 0.2631 (0.2629)  time: 0.4277  data: 0.0047  max mem: 22448
+train: [7]  [240/400]  eta: 0:01:12  lr: 0.000278  loss: 2.4890 (2.4849)  grad: 0.2618 (0.2631)  time: 0.4482  data: 0.0049  max mem: 22448
+train: [7]  [260/400]  eta: 0:01:03  lr: 0.000278  loss: 2.5051 (2.4837)  grad: 0.2560 (0.2623)  time: 0.4425  data: 0.0049  max mem: 22448
+train: [7]  [280/400]  eta: 0:00:53  lr: 0.000277  loss: 2.4723 (2.4821)  grad: 0.2543 (0.2622)  time: 0.4311  data: 0.0046  max mem: 22448
+train: [7]  [300/400]  eta: 0:00:44  lr: 0.000276  loss: 2.4171 (2.4803)  grad: 0.2563 (0.2624)  time: 0.4318  data: 0.0050  max mem: 22448
+train: [7]  [320/400]  eta: 0:00:35  lr: 0.000275  loss: 2.4499 (2.4799)  grad: 0.2595 (0.2621)  time: 0.4296  data: 0.0049  max mem: 22448
+train: [7]  [340/400]  eta: 0:00:26  lr: 0.000274  loss: 2.4552 (2.4783)  grad: 0.2553 (0.2617)  time: 0.4369  data: 0.0050  max mem: 22448
+train: [7]  [360/400]  eta: 0:00:17  lr: 0.000273  loss: 2.4580 (2.4794)  grad: 0.2615 (0.2621)  time: 0.4490  data: 0.0050  max mem: 22448
+train: [7]  [380/400]  eta: 0:00:08  lr: 0.000272  loss: 2.4838 (2.4803)  grad: 0.2659 (0.2626)  time: 0.4435  data: 0.0048  max mem: 22448
+train: [7]  [399/400]  eta: 0:00:00  lr: 0.000271  loss: 2.5126 (2.4819)  grad: 0.2683 (0.2630)  time: 0.4378  data: 0.0049  max mem: 22448
+train: [7] Total time: 0:02:58 (0.4466 s / it)
+train: [7] Summary: lr: 0.000271  loss: 2.5126 (2.4819)  grad: 0.2683 (0.2630)
+eval (validation): [7]  [ 0/85]  eta: 0:04:17    time: 3.0308  data: 2.7923  max mem: 22448
+eval (validation): [7]  [20/85]  eta: 0:00:31    time: 0.3510  data: 0.0049  max mem: 22448
+eval (validation): [7]  [40/85]  eta: 0:00:19    time: 0.3743  data: 0.0040  max mem: 22448
+eval (validation): [7]  [60/85]  eta: 0:00:10    time: 0.3561  data: 0.0049  max mem: 22448
+eval (validation): [7]  [80/85]  eta: 0:00:01    time: 0.3192  data: 0.0043  max mem: 22448
+eval (validation): [7]  [84/85]  eta: 0:00:00    time: 0.3147  data: 0.0041  max mem: 22448
+eval (validation): [7] Total time: 0:00:32 (0.3829 s / it)
+cv: [7]  best hparam: (0.72, 1.0) (022) ('022_lr7.2e-01_wd1.0e+00')  loss: 2.403  acc: 0.272  f1: 0.215
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [8]  [  0/400]  eta: 0:22:00  lr: nan  time: 3.3019  data: 2.9128  max mem: 22448
+train: [8]  [ 20/400]  eta: 0:03:40  lr: 0.000270  loss: 2.3179 (2.3559)  grad: 0.2456 (0.2490)  time: 0.4438  data: 0.0041  max mem: 22448
+train: [8]  [ 40/400]  eta: 0:03:05  lr: 0.000270  loss: 2.3870 (2.3784)  grad: 0.2503 (0.2534)  time: 0.4485  data: 0.0045  max mem: 22448
+train: [8]  [ 60/400]  eta: 0:02:47  lr: 0.000269  loss: 2.4014 (2.3946)  grad: 0.2549 (0.2543)  time: 0.4422  data: 0.0046  max mem: 22448
+train: [8]  [ 80/400]  eta: 0:02:32  lr: 0.000268  loss: 2.4159 (2.4018)  grad: 0.2597 (0.2597)  time: 0.4327  data: 0.0048  max mem: 22448
+train: [8]  [100/400]  eta: 0:02:20  lr: 0.000267  loss: 2.4269 (2.4028)  grad: 0.2741 (0.2637)  time: 0.4364  data: 0.0049  max mem: 22448
+train: [8]  [120/400]  eta: 0:02:09  lr: 0.000266  loss: 2.4311 (2.4039)  grad: 0.2802 (0.2670)  time: 0.4312  data: 0.0050  max mem: 22448
+train: [8]  [140/400]  eta: 0:01:59  lr: 0.000265  loss: 2.4029 (2.4080)  grad: 0.2738 (0.2682)  time: 0.4394  data: 0.0050  max mem: 22448
+train: [8]  [160/400]  eta: 0:01:49  lr: 0.000264  loss: 2.4224 (2.4097)  grad: 0.2775 (0.2709)  time: 0.4466  data: 0.0050  max mem: 22448
+train: [8]  [180/400]  eta: 0:01:40  lr: 0.000263  loss: 2.4134 (2.4055)  grad: 0.2715 (0.2701)  time: 0.4524  data: 0.0051  max mem: 22448
+train: [8]  [200/400]  eta: 0:01:31  lr: 0.000262  loss: 2.4092 (2.4081)  grad: 0.2693 (0.2703)  time: 0.4378  data: 0.0049  max mem: 22448
+train: [8]  [220/400]  eta: 0:01:21  lr: 0.000260  loss: 2.4294 (2.4096)  grad: 0.2742 (0.2704)  time: 0.4342  data: 0.0047  max mem: 22448
+train: [8]  [240/400]  eta: 0:01:12  lr: 0.000259  loss: 2.4237 (2.4101)  grad: 0.2707 (0.2708)  time: 0.4345  data: 0.0048  max mem: 22448
+train: [8]  [260/400]  eta: 0:01:03  lr: 0.000258  loss: 2.4287 (2.4124)  grad: 0.2693 (0.2711)  time: 0.4546  data: 0.0050  max mem: 22448
+train: [8]  [280/400]  eta: 0:00:54  lr: 0.000257  loss: 2.4287 (2.4114)  grad: 0.2651 (0.2715)  time: 0.4336  data: 0.0050  max mem: 22448
+train: [8]  [300/400]  eta: 0:00:44  lr: 0.000256  loss: 2.4074 (2.4118)  grad: 0.2643 (0.2715)  time: 0.4325  data: 0.0049  max mem: 22448
+train: [8]  [320/400]  eta: 0:00:35  lr: 0.000255  loss: 2.4108 (2.4125)  grad: 0.2631 (0.2710)  time: 0.4348  data: 0.0048  max mem: 22448
+train: [8]  [340/400]  eta: 0:00:26  lr: 0.000254  loss: 2.4125 (2.4133)  grad: 0.2683 (0.2715)  time: 0.4377  data: 0.0046  max mem: 22448
+train: [8]  [360/400]  eta: 0:00:17  lr: 0.000253  loss: 2.4258 (2.4142)  grad: 0.2683 (0.2711)  time: 0.4528  data: 0.0052  max mem: 22448
+train: [8]  [380/400]  eta: 0:00:08  lr: 0.000252  loss: 2.4237 (2.4149)  grad: 0.2651 (0.2708)  time: 0.4460  data: 0.0049  max mem: 22448
+train: [8]  [399/400]  eta: 0:00:00  lr: 0.000250  loss: 2.4103 (2.4155)  grad: 0.2659 (0.2714)  time: 0.4344  data: 0.0047  max mem: 22448
+train: [8] Total time: 0:02:59 (0.4480 s / it)
+train: [8] Summary: lr: 0.000250  loss: 2.4103 (2.4155)  grad: 0.2659 (0.2714)
+eval (validation): [8]  [ 0/85]  eta: 0:04:22    time: 3.0901  data: 2.8057  max mem: 22448
+eval (validation): [8]  [20/85]  eta: 0:00:32    time: 0.3647  data: 0.0052  max mem: 22448
+eval (validation): [8]  [40/85]  eta: 0:00:19    time: 0.3603  data: 0.0042  max mem: 22448
+eval (validation): [8]  [60/85]  eta: 0:00:10    time: 0.3487  data: 0.0044  max mem: 22448
+eval (validation): [8]  [80/85]  eta: 0:00:01    time: 0.3213  data: 0.0039  max mem: 22448
+eval (validation): [8]  [84/85]  eta: 0:00:00    time: 0.3108  data: 0.0038  max mem: 22448
+eval (validation): [8] Total time: 0:00:32 (0.3824 s / it)
+cv: [8]  best hparam: (0.52, 1.0) (020) ('020_lr5.2e-01_wd1.0e+00')  loss: 2.422  acc: 0.268  f1: 0.209
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [9]  [  0/400]  eta: 0:21:49  lr: nan  time: 3.2736  data: 2.8890  max mem: 22448
+train: [9]  [ 20/400]  eta: 0:03:46  lr: 0.000249  loss: 2.3444 (2.3466)  grad: 0.2667 (0.2748)  time: 0.4615  data: 0.0042  max mem: 22448
+train: [9]  [ 40/400]  eta: 0:03:05  lr: 0.000248  loss: 2.3572 (2.3735)  grad: 0.2667 (0.2701)  time: 0.4309  data: 0.0047  max mem: 22448
+train: [9]  [ 60/400]  eta: 0:02:45  lr: 0.000247  loss: 2.3737 (2.3635)  grad: 0.2640 (0.2672)  time: 0.4302  data: 0.0048  max mem: 22448
+train: [9]  [ 80/400]  eta: 0:02:32  lr: 0.000246  loss: 2.3585 (2.3714)  grad: 0.2717 (0.2697)  time: 0.4392  data: 0.0048  max mem: 22448
+train: [9]  [100/400]  eta: 0:02:20  lr: 0.000244  loss: 2.3585 (2.3664)  grad: 0.2721 (0.2699)  time: 0.4354  data: 0.0047  max mem: 22448
+train: [9]  [120/400]  eta: 0:02:09  lr: 0.000243  loss: 2.3480 (2.3645)  grad: 0.2674 (0.2698)  time: 0.4315  data: 0.0047  max mem: 22448
+train: [9]  [140/400]  eta: 0:01:59  lr: 0.000242  loss: 2.3480 (2.3652)  grad: 0.2674 (0.2703)  time: 0.4362  data: 0.0050  max mem: 22448
+train: [9]  [160/400]  eta: 0:01:49  lr: 0.000241  loss: 2.3403 (2.3600)  grad: 0.2649 (0.2705)  time: 0.4368  data: 0.0046  max mem: 22448
+train: [9]  [180/400]  eta: 0:01:40  lr: 0.000240  loss: 2.3317 (2.3622)  grad: 0.2681 (0.2718)  time: 0.4610  data: 0.0052  max mem: 22448
+train: [9]  [200/400]  eta: 0:01:31  lr: 0.000238  loss: 2.3298 (2.3616)  grad: 0.2825 (0.2735)  time: 0.4485  data: 0.0052  max mem: 22448
+train: [9]  [220/400]  eta: 0:01:21  lr: 0.000237  loss: 2.3300 (2.3601)  grad: 0.2798 (0.2744)  time: 0.4447  data: 0.0049  max mem: 22448
+train: [9]  [240/400]  eta: 0:01:12  lr: 0.000236  loss: 2.3655 (2.3634)  grad: 0.2795 (0.2741)  time: 0.4258  data: 0.0048  max mem: 22448
+train: [9]  [260/400]  eta: 0:01:03  lr: 0.000234  loss: 2.3988 (2.3620)  grad: 0.2687 (0.2732)  time: 0.4423  data: 0.0050  max mem: 22448
+train: [9]  [280/400]  eta: 0:00:53  lr: 0.000233  loss: 2.3909 (2.3634)  grad: 0.2693 (0.2737)  time: 0.4338  data: 0.0048  max mem: 22448
+train: [9]  [300/400]  eta: 0:00:44  lr: 0.000232  loss: 2.3844 (2.3651)  grad: 0.2772 (0.2744)  time: 0.4342  data: 0.0051  max mem: 22448
+train: [9]  [320/400]  eta: 0:00:35  lr: 0.000230  loss: 2.3341 (2.3664)  grad: 0.2748 (0.2740)  time: 0.4369  data: 0.0049  max mem: 22448
+train: [9]  [340/400]  eta: 0:00:26  lr: 0.000229  loss: 2.3309 (2.3640)  grad: 0.2715 (0.2743)  time: 0.4418  data: 0.0049  max mem: 22448
+train: [9]  [360/400]  eta: 0:00:17  lr: 0.000228  loss: 2.3253 (2.3638)  grad: 0.2741 (0.2748)  time: 0.4482  data: 0.0052  max mem: 22448
+train: [9]  [380/400]  eta: 0:00:08  lr: 0.000226  loss: 2.3256 (2.3630)  grad: 0.2716 (0.2748)  time: 0.4500  data: 0.0049  max mem: 22448
+train: [9]  [399/400]  eta: 0:00:00  lr: 0.000225  loss: 2.3569 (2.3638)  grad: 0.2714 (0.2752)  time: 0.4281  data: 0.0047  max mem: 22448
+train: [9] Total time: 0:02:58 (0.4474 s / it)
+train: [9] Summary: lr: 0.000225  loss: 2.3569 (2.3638)  grad: 0.2714 (0.2752)
+eval (validation): [9]  [ 0/85]  eta: 0:04:19    time: 3.0482  data: 2.8088  max mem: 22448
+eval (validation): [9]  [20/85]  eta: 0:00:31    time: 0.3516  data: 0.0039  max mem: 22448
+eval (validation): [9]  [40/85]  eta: 0:00:18    time: 0.3331  data: 0.0040  max mem: 22448
+eval (validation): [9]  [60/85]  eta: 0:00:09    time: 0.3437  data: 0.0040  max mem: 22448
+eval (validation): [9]  [80/85]  eta: 0:00:01    time: 0.3260  data: 0.0040  max mem: 22448
+eval (validation): [9]  [84/85]  eta: 0:00:00    time: 0.3192  data: 0.0038  max mem: 22448
+eval (validation): [9] Total time: 0:00:31 (0.3723 s / it)
+cv: [9]  best hparam: (0.52, 1.0) (020) ('020_lr5.2e-01_wd1.0e+00')  loss: 2.406  acc: 0.269  f1: 0.206
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [10]  [  0/400]  eta: 0:21:49  lr: nan  time: 3.2740  data: 2.8836  max mem: 22448
+train: [10]  [ 20/400]  eta: 0:03:40  lr: 0.000224  loss: 2.2392 (2.2700)  grad: 0.2676 (0.2772)  time: 0.4456  data: 0.0039  max mem: 22448
+train: [10]  [ 40/400]  eta: 0:03:03  lr: 0.000222  loss: 2.2827 (2.2696)  grad: 0.2682 (0.2733)  time: 0.4337  data: 0.0041  max mem: 22448
+train: [10]  [ 60/400]  eta: 0:02:46  lr: 0.000221  loss: 2.2895 (2.2879)  grad: 0.2668 (0.2703)  time: 0.4545  data: 0.0051  max mem: 22448
+train: [10]  [ 80/400]  eta: 0:02:33  lr: 0.000220  loss: 2.2900 (2.2790)  grad: 0.2651 (0.2699)  time: 0.4473  data: 0.0048  max mem: 22448
+train: [10]  [100/400]  eta: 0:02:21  lr: 0.000218  loss: 2.2879 (2.2818)  grad: 0.2724 (0.2710)  time: 0.4335  data: 0.0049  max mem: 22448
+train: [10]  [120/400]  eta: 0:02:10  lr: 0.000217  loss: 2.2931 (2.2852)  grad: 0.2714 (0.2722)  time: 0.4432  data: 0.0049  max mem: 22448
+train: [10]  [140/400]  eta: 0:02:00  lr: 0.000215  loss: 2.3151 (2.2902)  grad: 0.2714 (0.2725)  time: 0.4432  data: 0.0050  max mem: 22448
+train: [10]  [160/400]  eta: 0:01:50  lr: 0.000214  loss: 2.3157 (2.2933)  grad: 0.2711 (0.2729)  time: 0.4355  data: 0.0050  max mem: 22448
+train: [10]  [180/400]  eta: 0:01:40  lr: 0.000213  loss: 2.2938 (2.2959)  grad: 0.2704 (0.2732)  time: 0.4495  data: 0.0047  max mem: 22448
+train: [10]  [200/400]  eta: 0:01:31  lr: 0.000211  loss: 2.2938 (2.2980)  grad: 0.2745 (0.2739)  time: 0.4557  data: 0.0052  max mem: 22448
+train: [10]  [220/400]  eta: 0:01:22  lr: 0.000210  loss: 2.3438 (2.2989)  grad: 0.2777 (0.2742)  time: 0.4417  data: 0.0047  max mem: 22448
+train: [10]  [240/400]  eta: 0:01:12  lr: 0.000208  loss: 2.2846 (2.2988)  grad: 0.2712 (0.2738)  time: 0.4458  data: 0.0050  max mem: 22448
+train: [10]  [260/400]  eta: 0:01:03  lr: 0.000207  loss: 2.2971 (2.2988)  grad: 0.2682 (0.2739)  time: 0.4431  data: 0.0049  max mem: 22448
+train: [10]  [280/400]  eta: 0:00:54  lr: 0.000205  loss: 2.3215 (2.2996)  grad: 0.2692 (0.2742)  time: 0.4612  data: 0.0051  max mem: 22448
+train: [10]  [300/400]  eta: 0:00:45  lr: 0.000204  loss: 2.2468 (2.2966)  grad: 0.2701 (0.2739)  time: 0.4553  data: 0.0051  max mem: 22448
+train: [10]  [320/400]  eta: 0:00:36  lr: 0.000202  loss: 2.2446 (2.2948)  grad: 0.2737 (0.2744)  time: 0.4340  data: 0.0047  max mem: 22448
+train: [10]  [340/400]  eta: 0:00:27  lr: 0.000201  loss: 2.2695 (2.2944)  grad: 0.2762 (0.2747)  time: 0.4365  data: 0.0047  max mem: 22448
+train: [10]  [360/400]  eta: 0:00:18  lr: 0.000199  loss: 2.2728 (2.2936)  grad: 0.2714 (0.2745)  time: 0.4534  data: 0.0051  max mem: 22448
+train: [10]  [380/400]  eta: 0:00:09  lr: 0.000198  loss: 2.2511 (2.2928)  grad: 0.2706 (0.2744)  time: 0.4402  data: 0.0051  max mem: 22448
+train: [10]  [399/400]  eta: 0:00:00  lr: 0.000196  loss: 2.2735 (2.2929)  grad: 0.2751 (0.2744)  time: 0.4383  data: 0.0051  max mem: 22448
+train: [10] Total time: 0:03:00 (0.4523 s / it)
+train: [10] Summary: lr: 0.000196  loss: 2.2735 (2.2929)  grad: 0.2751 (0.2744)
+eval (validation): [10]  [ 0/85]  eta: 0:04:25    time: 3.1211  data: 2.8230  max mem: 22448
+eval (validation): [10]  [20/85]  eta: 0:00:33    time: 0.3803  data: 0.0059  max mem: 22448
+eval (validation): [10]  [40/85]  eta: 0:00:19    time: 0.3505  data: 0.0039  max mem: 22448
+eval (validation): [10]  [60/85]  eta: 0:00:10    time: 0.3584  data: 0.0047  max mem: 22448
+eval (validation): [10]  [80/85]  eta: 0:00:01    time: 0.3245  data: 0.0040  max mem: 22448
+eval (validation): [10]  [84/85]  eta: 0:00:00    time: 0.3118  data: 0.0040  max mem: 22448
+eval (validation): [10] Total time: 0:00:32 (0.3856 s / it)
+cv: [10]  best hparam: (0.52, 1.0) (020) ('020_lr5.2e-01_wd1.0e+00')  loss: 2.407  acc: 0.269  f1: 0.210
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [11]  [  0/400]  eta: 0:22:03  lr: nan  time: 3.3089  data: 2.9617  max mem: 22448
+train: [11]  [ 20/400]  eta: 0:03:48  lr: 0.000195  loss: 2.2191 (2.2002)  grad: 0.2641 (0.2681)  time: 0.4657  data: 0.0047  max mem: 22448
+train: [11]  [ 40/400]  eta: 0:03:07  lr: 0.000193  loss: 2.2346 (2.2175)  grad: 0.2667 (0.2681)  time: 0.4360  data: 0.0047  max mem: 22448
+train: [11]  [ 60/400]  eta: 0:02:46  lr: 0.000192  loss: 2.2207 (2.2084)  grad: 0.2704 (0.2708)  time: 0.4248  data: 0.0049  max mem: 22448
+train: [11]  [ 80/400]  eta: 0:02:32  lr: 0.000190  loss: 2.2314 (2.2233)  grad: 0.2723 (0.2730)  time: 0.4387  data: 0.0049  max mem: 22448
+train: [11]  [100/400]  eta: 0:02:20  lr: 0.000189  loss: 2.2454 (2.2222)  grad: 0.2712 (0.2733)  time: 0.4380  data: 0.0050  max mem: 22448
+train: [11]  [120/400]  eta: 0:02:09  lr: 0.000187  loss: 2.1711 (2.2108)  grad: 0.2699 (0.2742)  time: 0.4353  data: 0.0049  max mem: 22448
+train: [11]  [140/400]  eta: 0:01:59  lr: 0.000186  loss: 2.2103 (2.2171)  grad: 0.2756 (0.2752)  time: 0.4446  data: 0.0048  max mem: 22448
+train: [11]  [160/400]  eta: 0:01:50  lr: 0.000184  loss: 2.2350 (2.2241)  grad: 0.2770 (0.2762)  time: 0.4440  data: 0.0049  max mem: 22448
+train: [11]  [180/400]  eta: 0:01:40  lr: 0.000183  loss: 2.2314 (2.2243)  grad: 0.2887 (0.2778)  time: 0.4422  data: 0.0050  max mem: 22448
+train: [11]  [200/400]  eta: 0:01:31  lr: 0.000181  loss: 2.2570 (2.2275)  grad: 0.2851 (0.2781)  time: 0.4607  data: 0.0051  max mem: 22448
+train: [11]  [220/400]  eta: 0:01:22  lr: 0.000180  loss: 2.2820 (2.2349)  grad: 0.2755 (0.2781)  time: 0.4469  data: 0.0051  max mem: 22448
+train: [11]  [240/400]  eta: 0:01:12  lr: 0.000178  loss: 2.2934 (2.2363)  grad: 0.2699 (0.2779)  time: 0.4431  data: 0.0051  max mem: 22448
+train: [11]  [260/400]  eta: 0:01:03  lr: 0.000177  loss: 2.2590 (2.2374)  grad: 0.2735 (0.2783)  time: 0.4317  data: 0.0055  max mem: 22448
+train: [11]  [280/400]  eta: 0:00:54  lr: 0.000175  loss: 2.2539 (2.2392)  grad: 0.2854 (0.2788)  time: 0.4485  data: 0.0044  max mem: 22448
+train: [11]  [300/400]  eta: 0:00:45  lr: 0.000174  loss: 2.2570 (2.2428)  grad: 0.2892 (0.2794)  time: 0.4420  data: 0.0049  max mem: 22448
+train: [11]  [320/400]  eta: 0:00:36  lr: 0.000172  loss: 2.2630 (2.2448)  grad: 0.2834 (0.2799)  time: 0.4474  data: 0.0049  max mem: 22448
+train: [11]  [340/400]  eta: 0:00:27  lr: 0.000170  loss: 2.2130 (2.2456)  grad: 0.2830 (0.2805)  time: 0.4432  data: 0.0048  max mem: 22448
+train: [11]  [360/400]  eta: 0:00:18  lr: 0.000169  loss: 2.2286 (2.2454)  grad: 0.2831 (0.2810)  time: 0.4473  data: 0.0049  max mem: 22448
+train: [11]  [380/400]  eta: 0:00:09  lr: 0.000167  loss: 2.2087 (2.2431)  grad: 0.2772 (0.2808)  time: 0.4431  data: 0.0050  max mem: 22448
+train: [11]  [399/400]  eta: 0:00:00  lr: 0.000166  loss: 2.2610 (2.2469)  grad: 0.2707 (0.2806)  time: 0.4421  data: 0.0048  max mem: 22448
+train: [11] Total time: 0:03:00 (0.4510 s / it)
+train: [11] Summary: lr: 0.000166  loss: 2.2610 (2.2469)  grad: 0.2707 (0.2806)
+eval (validation): [11]  [ 0/85]  eta: 0:04:43    time: 3.3309  data: 3.0284  max mem: 22448
+eval (validation): [11]  [20/85]  eta: 0:00:34    time: 0.3930  data: 0.0036  max mem: 22448
+eval (validation): [11]  [40/85]  eta: 0:00:20    time: 0.3574  data: 0.0041  max mem: 22448
+eval (validation): [11]  [60/85]  eta: 0:00:10    time: 0.3404  data: 0.0041  max mem: 22448
+eval (validation): [11]  [80/85]  eta: 0:00:01    time: 0.3168  data: 0.0038  max mem: 22448
+eval (validation): [11]  [84/85]  eta: 0:00:00    time: 0.3120  data: 0.0038  max mem: 22448
+eval (validation): [11] Total time: 0:00:32 (0.3878 s / it)
+cv: [11]  best hparam: (0.52, 1.0) (020) ('020_lr5.2e-01_wd1.0e+00')  loss: 2.420  acc: 0.274  f1: 0.218
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [12]  [  0/400]  eta: 0:23:31  lr: nan  time: 3.5279  data: 3.1221  max mem: 22448
+train: [12]  [ 20/400]  eta: 0:03:44  lr: 0.000164  loss: 2.1168 (2.1431)  grad: 0.2573 (0.2578)  time: 0.4443  data: 0.0044  max mem: 22448
+train: [12]  [ 40/400]  eta: 0:03:08  lr: 0.000163  loss: 2.1661 (2.1539)  grad: 0.2631 (0.2665)  time: 0.4517  data: 0.0040  max mem: 22448
+train: [12]  [ 60/400]  eta: 0:02:47  lr: 0.000161  loss: 2.1481 (2.1541)  grad: 0.2714 (0.2672)  time: 0.4337  data: 0.0046  max mem: 22448
+train: [12]  [ 80/400]  eta: 0:02:34  lr: 0.000160  loss: 2.1901 (2.1603)  grad: 0.2656 (0.2666)  time: 0.4530  data: 0.0047  max mem: 22448
+train: [12]  [100/400]  eta: 0:02:22  lr: 0.000158  loss: 2.1901 (2.1628)  grad: 0.2658 (0.2678)  time: 0.4464  data: 0.0050  max mem: 22448
+train: [12]  [120/400]  eta: 0:02:11  lr: 0.000156  loss: 2.1636 (2.1676)  grad: 0.2658 (0.2665)  time: 0.4356  data: 0.0050  max mem: 22448
+train: [12]  [140/400]  eta: 0:02:00  lr: 0.000155  loss: 2.1708 (2.1683)  grad: 0.2671 (0.2683)  time: 0.4307  data: 0.0048  max mem: 22448
+train: [12]  [160/400]  eta: 0:01:52  lr: 0.000153  loss: 2.1824 (2.1713)  grad: 0.2826 (0.2709)  time: 0.4898  data: 0.0053  max mem: 22448
+train: [12]  [180/400]  eta: 0:01:42  lr: 0.000152  loss: 2.1872 (2.1714)  grad: 0.2823 (0.2715)  time: 0.4472  data: 0.0047  max mem: 22448
+train: [12]  [200/400]  eta: 0:01:32  lr: 0.000150  loss: 2.2198 (2.1779)  grad: 0.2741 (0.2718)  time: 0.4534  data: 0.0048  max mem: 22448
+train: [12]  [220/400]  eta: 0:01:22  lr: 0.000149  loss: 2.2276 (2.1816)  grad: 0.2696 (0.2717)  time: 0.4284  data: 0.0049  max mem: 22448
+train: [12]  [240/400]  eta: 0:01:13  lr: 0.000147  loss: 2.1629 (2.1789)  grad: 0.2732 (0.2731)  time: 0.4328  data: 0.0048  max mem: 22448
+train: [12]  [260/400]  eta: 0:01:03  lr: 0.000145  loss: 2.1471 (2.1795)  grad: 0.2819 (0.2734)  time: 0.4301  data: 0.0045  max mem: 22448
+train: [12]  [280/400]  eta: 0:00:54  lr: 0.000144  loss: 2.1566 (2.1765)  grad: 0.2752 (0.2734)  time: 0.4347  data: 0.0048  max mem: 22448
+train: [12]  [300/400]  eta: 0:00:45  lr: 0.000142  loss: 2.1710 (2.1798)  grad: 0.2775 (0.2747)  time: 0.4343  data: 0.0047  max mem: 22448
+train: [12]  [320/400]  eta: 0:00:36  lr: 0.000141  loss: 2.2002 (2.1805)  grad: 0.2783 (0.2749)  time: 0.4324  data: 0.0047  max mem: 22448
+train: [12]  [340/400]  eta: 0:00:27  lr: 0.000139  loss: 2.1817 (2.1798)  grad: 0.2715 (0.2747)  time: 0.4312  data: 0.0048  max mem: 22448
+train: [12]  [360/400]  eta: 0:00:17  lr: 0.000138  loss: 2.1784 (2.1804)  grad: 0.2714 (0.2746)  time: 0.4323  data: 0.0047  max mem: 22448
+train: [12]  [380/400]  eta: 0:00:08  lr: 0.000136  loss: 2.1826 (2.1816)  grad: 0.2745 (0.2747)  time: 0.4397  data: 0.0049  max mem: 22448
+train: [12]  [399/400]  eta: 0:00:00  lr: 0.000134  loss: 2.1634 (2.1805)  grad: 0.2798 (0.2751)  time: 0.4297  data: 0.0048  max mem: 22448
+train: [12] Total time: 0:02:59 (0.4488 s / it)
+train: [12] Summary: lr: 0.000134  loss: 2.1634 (2.1805)  grad: 0.2798 (0.2751)
+eval (validation): [12]  [ 0/85]  eta: 0:04:12    time: 2.9718  data: 2.7343  max mem: 22448
+eval (validation): [12]  [20/85]  eta: 0:00:30    time: 0.3454  data: 0.0037  max mem: 22448
+eval (validation): [12]  [40/85]  eta: 0:00:17    time: 0.3186  data: 0.0034  max mem: 22448
+eval (validation): [12]  [60/85]  eta: 0:00:09    time: 0.3301  data: 0.0038  max mem: 22448
+eval (validation): [12]  [80/85]  eta: 0:00:01    time: 0.3144  data: 0.0039  max mem: 22448
+eval (validation): [12]  [84/85]  eta: 0:00:00    time: 0.3061  data: 0.0039  max mem: 22448
+eval (validation): [12] Total time: 0:00:30 (0.3594 s / it)
+cv: [12]  best hparam: (0.44, 1.0) (019) ('019_lr4.4e-01_wd1.0e+00')  loss: 2.445  acc: 0.264  f1: 0.205
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [13]  [  0/400]  eta: 0:21:02  lr: nan  time: 3.1551  data: 2.7447  max mem: 22448
+train: [13]  [ 20/400]  eta: 0:03:45  lr: 0.000133  loss: 2.1346 (2.1317)  grad: 0.2673 (0.2729)  time: 0.4651  data: 0.0043  max mem: 22448
+train: [13]  [ 40/400]  eta: 0:03:05  lr: 0.000131  loss: 2.1415 (2.1277)  grad: 0.2673 (0.2709)  time: 0.4335  data: 0.0051  max mem: 22448
+train: [13]  [ 60/400]  eta: 0:02:45  lr: 0.000130  loss: 2.1291 (2.1210)  grad: 0.2716 (0.2711)  time: 0.4317  data: 0.0046  max mem: 22448
+train: [13]  [ 80/400]  eta: 0:02:31  lr: 0.000128  loss: 2.1003 (2.1191)  grad: 0.2716 (0.2725)  time: 0.4291  data: 0.0048  max mem: 22448
+train: [13]  [100/400]  eta: 0:02:19  lr: 0.000127  loss: 2.0937 (2.1187)  grad: 0.2680 (0.2722)  time: 0.4269  data: 0.0048  max mem: 22448
+train: [13]  [120/400]  eta: 0:02:08  lr: 0.000125  loss: 2.1101 (2.1169)  grad: 0.2731 (0.2733)  time: 0.4271  data: 0.0049  max mem: 22448
+train: [13]  [140/400]  eta: 0:01:58  lr: 0.000124  loss: 2.1407 (2.1283)  grad: 0.2810 (0.2750)  time: 0.4287  data: 0.0050  max mem: 22448
+train: [13]  [160/400]  eta: 0:01:48  lr: 0.000122  loss: 2.1497 (2.1280)  grad: 0.2797 (0.2753)  time: 0.4272  data: 0.0049  max mem: 22448
+train: [13]  [180/400]  eta: 0:01:38  lr: 0.000120  loss: 2.1278 (2.1342)  grad: 0.2794 (0.2759)  time: 0.4300  data: 0.0050  max mem: 22448
+train: [13]  [200/400]  eta: 0:01:29  lr: 0.000119  loss: 2.1071 (2.1302)  grad: 0.2794 (0.2760)  time: 0.4448  data: 0.0050  max mem: 22448
+train: [13]  [220/400]  eta: 0:01:20  lr: 0.000117  loss: 2.1043 (2.1298)  grad: 0.2744 (0.2767)  time: 0.4375  data: 0.0049  max mem: 22448
+train: [13]  [240/400]  eta: 0:01:11  lr: 0.000116  loss: 2.1391 (2.1302)  grad: 0.2765 (0.2766)  time: 0.4402  data: 0.0048  max mem: 22448
+train: [13]  [260/400]  eta: 0:01:02  lr: 0.000114  loss: 2.1396 (2.1332)  grad: 0.2709 (0.2761)  time: 0.4306  data: 0.0048  max mem: 22448
+train: [13]  [280/400]  eta: 0:00:53  lr: 0.000113  loss: 2.1007 (2.1301)  grad: 0.2615 (0.2751)  time: 0.4491  data: 0.0049  max mem: 22448
+train: [13]  [300/400]  eta: 0:00:44  lr: 0.000111  loss: 2.1007 (2.1306)  grad: 0.2590 (0.2743)  time: 0.4379  data: 0.0050  max mem: 22448
+train: [13]  [320/400]  eta: 0:00:35  lr: 0.000110  loss: 2.1613 (2.1330)  grad: 0.2708 (0.2748)  time: 0.4274  data: 0.0048  max mem: 22448
+train: [13]  [340/400]  eta: 0:00:26  lr: 0.000108  loss: 2.1197 (2.1329)  grad: 0.2772 (0.2745)  time: 0.4350  data: 0.0048  max mem: 22448
+train: [13]  [360/400]  eta: 0:00:17  lr: 0.000107  loss: 2.1137 (2.1337)  grad: 0.2828 (0.2753)  time: 0.4346  data: 0.0050  max mem: 22448
+train: [13]  [380/400]  eta: 0:00:08  lr: 0.000105  loss: 2.1150 (2.1329)  grad: 0.2855 (0.2756)  time: 0.4286  data: 0.0048  max mem: 22448
+train: [13]  [399/400]  eta: 0:00:00  lr: 0.000104  loss: 2.1101 (2.1328)  grad: 0.2800 (0.2756)  time: 0.4304  data: 0.0047  max mem: 22448
+train: [13] Total time: 0:02:56 (0.4422 s / it)
+train: [13] Summary: lr: 0.000104  loss: 2.1101 (2.1328)  grad: 0.2800 (0.2756)
+eval (validation): [13]  [ 0/85]  eta: 0:04:23    time: 3.1002  data: 2.8426  max mem: 22448
+eval (validation): [13]  [20/85]  eta: 0:00:33    time: 0.3808  data: 0.0188  max mem: 22448
+eval (validation): [13]  [40/85]  eta: 0:00:20    time: 0.3774  data: 0.0044  max mem: 22448
+eval (validation): [13]  [60/85]  eta: 0:00:10    time: 0.3473  data: 0.0043  max mem: 22448
+eval (validation): [13]  [80/85]  eta: 0:00:01    time: 0.3407  data: 0.0043  max mem: 22448
+eval (validation): [13]  [84/85]  eta: 0:00:00    time: 0.3295  data: 0.0042  max mem: 22448
+eval (validation): [13] Total time: 0:00:33 (0.3944 s / it)
+cv: [13]  best hparam: (0.38, 1.0) (018) ('018_lr3.8e-01_wd1.0e+00')  loss: 2.417  acc: 0.268  f1: 0.206
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [14]  [  0/400]  eta: 0:21:19  lr: nan  time: 3.1986  data: 2.8025  max mem: 22448
+train: [14]  [ 20/400]  eta: 0:03:49  lr: 0.000102  loss: 2.0469 (2.0456)  grad: 0.2593 (0.2639)  time: 0.4731  data: 0.0049  max mem: 22448
+train: [14]  [ 40/400]  eta: 0:03:11  lr: 0.000101  loss: 2.0524 (2.0583)  grad: 0.2562 (0.2643)  time: 0.4599  data: 0.0050  max mem: 22448
+train: [14]  [ 60/400]  eta: 0:02:49  lr: 0.000099  loss: 2.0403 (2.0605)  grad: 0.2635 (0.2661)  time: 0.4282  data: 0.0048  max mem: 22448
+train: [14]  [ 80/400]  eta: 0:02:33  lr: 0.000098  loss: 2.0786 (2.0684)  grad: 0.2623 (0.2649)  time: 0.4259  data: 0.0049  max mem: 22448
+train: [14]  [100/400]  eta: 0:02:21  lr: 0.000096  loss: 2.0871 (2.0771)  grad: 0.2627 (0.2662)  time: 0.4302  data: 0.0048  max mem: 22448
+train: [14]  [120/400]  eta: 0:02:09  lr: 0.000095  loss: 2.0505 (2.0698)  grad: 0.2720 (0.2669)  time: 0.4249  data: 0.0048  max mem: 22448
+train: [14]  [140/400]  eta: 0:01:58  lr: 0.000093  loss: 2.0147 (2.0719)  grad: 0.2756 (0.2682)  time: 0.4211  data: 0.0048  max mem: 22448
+train: [14]  [160/400]  eta: 0:01:48  lr: 0.000092  loss: 2.0597 (2.0708)  grad: 0.2704 (0.2684)  time: 0.4227  data: 0.0050  max mem: 22448
+train: [14]  [180/400]  eta: 0:01:39  lr: 0.000090  loss: 2.0648 (2.0707)  grad: 0.2662 (0.2686)  time: 0.4269  data: 0.0049  max mem: 22448
+train: [14]  [200/400]  eta: 0:01:30  lr: 0.000089  loss: 2.0648 (2.0716)  grad: 0.2699 (0.2695)  time: 0.4501  data: 0.0050  max mem: 22448
+train: [14]  [220/400]  eta: 0:01:20  lr: 0.000088  loss: 2.0504 (2.0706)  grad: 0.2717 (0.2700)  time: 0.4388  data: 0.0051  max mem: 22448
+train: [14]  [240/400]  eta: 0:01:11  lr: 0.000086  loss: 2.0514 (2.0741)  grad: 0.2694 (0.2705)  time: 0.4383  data: 0.0050  max mem: 22448
+train: [14]  [260/400]  eta: 0:01:02  lr: 0.000085  loss: 2.0771 (2.0744)  grad: 0.2761 (0.2709)  time: 0.4234  data: 0.0048  max mem: 22448
+train: [14]  [280/400]  eta: 0:00:53  lr: 0.000083  loss: 2.0759 (2.0735)  grad: 0.2711 (0.2704)  time: 0.4478  data: 0.0050  max mem: 22448
+train: [14]  [300/400]  eta: 0:00:44  lr: 0.000082  loss: 2.0955 (2.0796)  grad: 0.2711 (0.2710)  time: 0.4288  data: 0.0048  max mem: 22448
+train: [14]  [320/400]  eta: 0:00:35  lr: 0.000081  loss: 2.1209 (2.0808)  grad: 0.2794 (0.2718)  time: 0.4209  data: 0.0046  max mem: 22448
+train: [14]  [340/400]  eta: 0:00:26  lr: 0.000079  loss: 2.1099 (2.0813)  grad: 0.2847 (0.2727)  time: 0.4293  data: 0.0050  max mem: 22448
+train: [14]  [360/400]  eta: 0:00:17  lr: 0.000078  loss: 2.0864 (2.0818)  grad: 0.2794 (0.2726)  time: 0.4258  data: 0.0049  max mem: 22448
+train: [14]  [380/400]  eta: 0:00:08  lr: 0.000076  loss: 2.0676 (2.0800)  grad: 0.2750 (0.2727)  time: 0.4242  data: 0.0047  max mem: 22448
+train: [14]  [399/400]  eta: 0:00:00  lr: 0.000075  loss: 2.0603 (2.0808)  grad: 0.2782 (0.2733)  time: 0.4269  data: 0.0047  max mem: 22448
+train: [14] Total time: 0:02:56 (0.4409 s / it)
+train: [14] Summary: lr: 0.000075  loss: 2.0603 (2.0808)  grad: 0.2782 (0.2733)
+eval (validation): [14]  [ 0/85]  eta: 0:04:35    time: 3.2438  data: 2.9466  max mem: 22448
+eval (validation): [14]  [20/85]  eta: 0:00:32    time: 0.3566  data: 0.0048  max mem: 22448
+eval (validation): [14]  [40/85]  eta: 0:00:18    time: 0.3380  data: 0.0037  max mem: 22448
+eval (validation): [14]  [60/85]  eta: 0:00:09    time: 0.3615  data: 0.0045  max mem: 22448
+eval (validation): [14]  [80/85]  eta: 0:00:01    time: 0.3183  data: 0.0041  max mem: 22448
+eval (validation): [14]  [84/85]  eta: 0:00:00    time: 0.3051  data: 0.0039  max mem: 22448
+eval (validation): [14] Total time: 0:00:32 (0.3789 s / it)
+cv: [14]  best hparam: (0.44, 1.0) (019) ('019_lr4.4e-01_wd1.0e+00')  loss: 2.422  acc: 0.268  f1: 0.210
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [15]  [  0/400]  eta: 0:22:36  lr: nan  time: 3.3919  data: 3.0004  max mem: 22448
+train: [15]  [ 20/400]  eta: 0:03:38  lr: 0.000074  loss: 2.0024 (2.0253)  grad: 0.2572 (0.2602)  time: 0.4352  data: 0.0041  max mem: 22448
+train: [15]  [ 40/400]  eta: 0:03:02  lr: 0.000072  loss: 2.0086 (2.0286)  grad: 0.2572 (0.2605)  time: 0.4325  data: 0.0048  max mem: 22448
+train: [15]  [ 60/400]  eta: 0:02:44  lr: 0.000071  loss: 1.9867 (2.0150)  grad: 0.2688 (0.2628)  time: 0.4376  data: 0.0048  max mem: 22448
+train: [15]  [ 80/400]  eta: 0:02:30  lr: 0.000070  loss: 1.9787 (2.0078)  grad: 0.2697 (0.2625)  time: 0.4292  data: 0.0049  max mem: 22448
+train: [15]  [100/400]  eta: 0:02:18  lr: 0.000068  loss: 1.9713 (2.0080)  grad: 0.2639 (0.2640)  time: 0.4293  data: 0.0047  max mem: 22448
+train: [15]  [120/400]  eta: 0:02:07  lr: 0.000067  loss: 2.0030 (2.0111)  grad: 0.2676 (0.2654)  time: 0.4217  data: 0.0045  max mem: 22448
+train: [15]  [140/400]  eta: 0:01:57  lr: 0.000066  loss: 2.0329 (2.0181)  grad: 0.2744 (0.2678)  time: 0.4389  data: 0.0050  max mem: 22448
+train: [15]  [160/400]  eta: 0:01:48  lr: 0.000064  loss: 2.0179 (2.0183)  grad: 0.2741 (0.2682)  time: 0.4290  data: 0.0049  max mem: 22448
+train: [15]  [180/400]  eta: 0:01:38  lr: 0.000063  loss: 2.0551 (2.0275)  grad: 0.2733 (0.2694)  time: 0.4292  data: 0.0048  max mem: 22448
+train: [15]  [200/400]  eta: 0:01:29  lr: 0.000062  loss: 2.0614 (2.0280)  grad: 0.2708 (0.2690)  time: 0.4499  data: 0.0048  max mem: 22448
+train: [15]  [220/400]  eta: 0:01:20  lr: 0.000061  loss: 2.0257 (2.0289)  grad: 0.2639 (0.2690)  time: 0.4401  data: 0.0050  max mem: 22448
+train: [15]  [240/400]  eta: 0:01:11  lr: 0.000059  loss: 2.0237 (2.0289)  grad: 0.2681 (0.2691)  time: 0.4411  data: 0.0049  max mem: 22448
+train: [15]  [260/400]  eta: 0:01:02  lr: 0.000058  loss: 2.0629 (2.0334)  grad: 0.2696 (0.2690)  time: 0.4203  data: 0.0046  max mem: 22448
+train: [15]  [280/400]  eta: 0:00:53  lr: 0.000057  loss: 2.0471 (2.0330)  grad: 0.2582 (0.2683)  time: 0.4400  data: 0.0051  max mem: 22448
+train: [15]  [300/400]  eta: 0:00:44  lr: 0.000056  loss: 2.0211 (2.0333)  grad: 0.2563 (0.2685)  time: 0.4281  data: 0.0049  max mem: 22448
+train: [15]  [320/400]  eta: 0:00:35  lr: 0.000054  loss: 2.0497 (2.0337)  grad: 0.2651 (0.2684)  time: 0.4236  data: 0.0050  max mem: 22448
+train: [15]  [340/400]  eta: 0:00:26  lr: 0.000053  loss: 2.0497 (2.0343)  grad: 0.2692 (0.2685)  time: 0.4366  data: 0.0050  max mem: 22448
+train: [15]  [360/400]  eta: 0:00:17  lr: 0.000052  loss: 2.0427 (2.0358)  grad: 0.2731 (0.2690)  time: 0.4293  data: 0.0053  max mem: 22448
+train: [15]  [380/400]  eta: 0:00:08  lr: 0.000051  loss: 2.0180 (2.0339)  grad: 0.2656 (0.2686)  time: 0.4376  data: 0.0048  max mem: 22448
+train: [15]  [399/400]  eta: 0:00:00  lr: 0.000050  loss: 2.0202 (2.0349)  grad: 0.2617 (0.2684)  time: 0.4394  data: 0.0048  max mem: 22448
+train: [15] Total time: 0:02:56 (0.4412 s / it)
+train: [15] Summary: lr: 0.000050  loss: 2.0202 (2.0349)  grad: 0.2617 (0.2684)
+eval (validation): [15]  [ 0/85]  eta: 0:04:27    time: 3.1495  data: 2.9155  max mem: 22448
+eval (validation): [15]  [20/85]  eta: 0:00:30    time: 0.3423  data: 0.0057  max mem: 22448
+eval (validation): [15]  [40/85]  eta: 0:00:18    time: 0.3356  data: 0.0036  max mem: 22448
+eval (validation): [15]  [60/85]  eta: 0:00:09    time: 0.3276  data: 0.0041  max mem: 22448
+eval (validation): [15]  [80/85]  eta: 0:00:01    time: 0.3412  data: 0.0043  max mem: 22448
+eval (validation): [15]  [84/85]  eta: 0:00:00    time: 0.3287  data: 0.0043  max mem: 22448
+eval (validation): [15] Total time: 0:00:31 (0.3719 s / it)
+cv: [15]  best hparam: (0.44, 1.0) (019) ('019_lr4.4e-01_wd1.0e+00')  loss: 2.427  acc: 0.267  f1: 0.209
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [16]  [  0/400]  eta: 0:27:54  lr: nan  time: 4.1861  data: 3.7956  max mem: 22448
+train: [16]  [ 20/400]  eta: 0:03:55  lr: 0.000048  loss: 1.9586 (1.9752)  grad: 0.2406 (0.2464)  time: 0.4403  data: 0.0033  max mem: 22448
+train: [16]  [ 40/400]  eta: 0:03:07  lr: 0.000047  loss: 1.9870 (1.9810)  grad: 0.2439 (0.2476)  time: 0.4198  data: 0.0048  max mem: 22448
+train: [16]  [ 60/400]  eta: 0:02:48  lr: 0.000046  loss: 1.9870 (1.9785)  grad: 0.2483 (0.2504)  time: 0.4399  data: 0.0049  max mem: 22448
+train: [16]  [ 80/400]  eta: 0:02:34  lr: 0.000045  loss: 1.9929 (1.9928)  grad: 0.2560 (0.2541)  time: 0.4509  data: 0.0048  max mem: 22448
+train: [16]  [100/400]  eta: 0:02:22  lr: 0.000044  loss: 1.9939 (1.9918)  grad: 0.2637 (0.2559)  time: 0.4307  data: 0.0049  max mem: 22448
+train: [16]  [120/400]  eta: 0:02:10  lr: 0.000043  loss: 1.9880 (1.9931)  grad: 0.2633 (0.2565)  time: 0.4272  data: 0.0049  max mem: 22448
+train: [16]  [140/400]  eta: 0:01:59  lr: 0.000042  loss: 1.9912 (1.9916)  grad: 0.2541 (0.2564)  time: 0.4247  data: 0.0050  max mem: 22448
+train: [16]  [160/400]  eta: 0:01:49  lr: 0.000041  loss: 1.9914 (1.9958)  grad: 0.2591 (0.2577)  time: 0.4285  data: 0.0051  max mem: 22448
+train: [16]  [180/400]  eta: 0:01:39  lr: 0.000040  loss: 1.9905 (1.9955)  grad: 0.2626 (0.2581)  time: 0.4333  data: 0.0050  max mem: 22448
+train: [16]  [200/400]  eta: 0:01:30  lr: 0.000039  loss: 1.9805 (1.9922)  grad: 0.2568 (0.2572)  time: 0.4491  data: 0.0051  max mem: 22448
+train: [16]  [220/400]  eta: 0:01:21  lr: 0.000038  loss: 1.9645 (1.9928)  grad: 0.2468 (0.2570)  time: 0.4324  data: 0.0047  max mem: 22448
+train: [16]  [240/400]  eta: 0:01:12  lr: 0.000036  loss: 1.9866 (1.9935)  grad: 0.2537 (0.2576)  time: 0.4389  data: 0.0048  max mem: 22448
+train: [16]  [260/400]  eta: 0:01:02  lr: 0.000035  loss: 1.9897 (1.9963)  grad: 0.2701 (0.2586)  time: 0.4235  data: 0.0045  max mem: 22448
+train: [16]  [280/400]  eta: 0:00:53  lr: 0.000034  loss: 2.0060 (1.9973)  grad: 0.2675 (0.2593)  time: 0.4425  data: 0.0049  max mem: 22448
+train: [16]  [300/400]  eta: 0:00:44  lr: 0.000033  loss: 2.0218 (1.9990)  grad: 0.2646 (0.2597)  time: 0.4387  data: 0.0049  max mem: 22448
+train: [16]  [320/400]  eta: 0:00:35  lr: 0.000032  loss: 2.0314 (2.0019)  grad: 0.2643 (0.2600)  time: 0.4264  data: 0.0049  max mem: 22448
+train: [16]  [340/400]  eta: 0:00:26  lr: 0.000031  loss: 1.9874 (2.0009)  grad: 0.2605 (0.2603)  time: 0.4289  data: 0.0049  max mem: 22448
+train: [16]  [360/400]  eta: 0:00:17  lr: 0.000031  loss: 1.9717 (2.0017)  grad: 0.2577 (0.2605)  time: 0.4331  data: 0.0051  max mem: 22448
+train: [16]  [380/400]  eta: 0:00:08  lr: 0.000030  loss: 2.0059 (2.0021)  grad: 0.2638 (0.2610)  time: 0.4266  data: 0.0050  max mem: 22448
+train: [16]  [399/400]  eta: 0:00:00  lr: 0.000029  loss: 2.0059 (2.0028)  grad: 0.2692 (0.2617)  time: 0.4253  data: 0.0049  max mem: 22448
+train: [16] Total time: 0:02:57 (0.4431 s / it)
+train: [16] Summary: lr: 0.000029  loss: 2.0059 (2.0028)  grad: 0.2692 (0.2617)
+eval (validation): [16]  [ 0/85]  eta: 0:04:22    time: 3.0857  data: 2.8421  max mem: 22448
+eval (validation): [16]  [20/85]  eta: 0:00:32    time: 0.3718  data: 0.0050  max mem: 22448
+eval (validation): [16]  [40/85]  eta: 0:00:19    time: 0.3418  data: 0.0040  max mem: 22448
+eval (validation): [16]  [60/85]  eta: 0:00:09    time: 0.3411  data: 0.0039  max mem: 22448
+eval (validation): [16]  [80/85]  eta: 0:00:01    time: 0.3272  data: 0.0041  max mem: 22448
+eval (validation): [16]  [84/85]  eta: 0:00:00    time: 0.3201  data: 0.0040  max mem: 22448
+eval (validation): [16] Total time: 0:00:32 (0.3797 s / it)
+cv: [16]  best hparam: (0.44, 1.0) (019) ('019_lr4.4e-01_wd1.0e+00')  loss: 2.427  acc: 0.268  f1: 0.212
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [17]  [  0/400]  eta: 0:20:55  lr: nan  time: 3.1399  data: 2.8109  max mem: 22448
+train: [17]  [ 20/400]  eta: 0:03:35  lr: 0.000028  loss: 1.9744 (1.9514)  grad: 0.2434 (0.2507)  time: 0.4391  data: 0.0042  max mem: 22448
+train: [17]  [ 40/400]  eta: 0:03:02  lr: 0.000027  loss: 1.9755 (1.9613)  grad: 0.2562 (0.2538)  time: 0.4450  data: 0.0049  max mem: 22448
+train: [17]  [ 60/400]  eta: 0:02:43  lr: 0.000026  loss: 1.9811 (1.9723)  grad: 0.2491 (0.2518)  time: 0.4254  data: 0.0048  max mem: 22448
+train: [17]  [ 80/400]  eta: 0:02:31  lr: 0.000025  loss: 1.9585 (1.9596)  grad: 0.2435 (0.2516)  time: 0.4560  data: 0.0048  max mem: 22448
+train: [17]  [100/400]  eta: 0:02:19  lr: 0.000024  loss: 1.9585 (1.9664)  grad: 0.2472 (0.2515)  time: 0.4297  data: 0.0049  max mem: 22448
+train: [17]  [120/400]  eta: 0:02:08  lr: 0.000023  loss: 1.9852 (1.9687)  grad: 0.2508 (0.2523)  time: 0.4285  data: 0.0050  max mem: 22448
+train: [17]  [140/400]  eta: 0:01:58  lr: 0.000023  loss: 1.9834 (1.9702)  grad: 0.2508 (0.2524)  time: 0.4386  data: 0.0049  max mem: 22448
+train: [17]  [160/400]  eta: 0:01:48  lr: 0.000022  loss: 1.9756 (1.9680)  grad: 0.2537 (0.2532)  time: 0.4267  data: 0.0049  max mem: 22448
+train: [17]  [180/400]  eta: 0:01:39  lr: 0.000021  loss: 1.9334 (1.9657)  grad: 0.2505 (0.2526)  time: 0.4304  data: 0.0047  max mem: 22448
+train: [17]  [200/400]  eta: 0:01:29  lr: 0.000020  loss: 1.9334 (1.9640)  grad: 0.2505 (0.2523)  time: 0.4416  data: 0.0050  max mem: 22448
+train: [17]  [220/400]  eta: 0:01:20  lr: 0.000019  loss: 1.9749 (1.9661)  grad: 0.2550 (0.2534)  time: 0.4362  data: 0.0047  max mem: 22448
+train: [17]  [240/400]  eta: 0:01:11  lr: 0.000019  loss: 1.9841 (1.9657)  grad: 0.2550 (0.2539)  time: 0.4372  data: 0.0051  max mem: 22448
+train: [17]  [260/400]  eta: 0:01:02  lr: 0.000018  loss: 1.9301 (1.9657)  grad: 0.2599 (0.2543)  time: 0.4306  data: 0.0048  max mem: 22448
+train: [17]  [280/400]  eta: 0:00:53  lr: 0.000017  loss: 1.9509 (1.9665)  grad: 0.2561 (0.2542)  time: 0.4458  data: 0.0050  max mem: 22448
+train: [17]  [300/400]  eta: 0:00:44  lr: 0.000016  loss: 1.9695 (1.9684)  grad: 0.2533 (0.2542)  time: 0.4358  data: 0.0050  max mem: 22448
+train: [17]  [320/400]  eta: 0:00:35  lr: 0.000016  loss: 1.9641 (1.9675)  grad: 0.2467 (0.2534)  time: 0.4277  data: 0.0048  max mem: 22448
+train: [17]  [340/400]  eta: 0:00:26  lr: 0.000015  loss: 1.9469 (1.9676)  grad: 0.2431 (0.2536)  time: 0.4358  data: 0.0050  max mem: 22448
+train: [17]  [360/400]  eta: 0:00:17  lr: 0.000014  loss: 1.9768 (1.9684)  grad: 0.2523 (0.2537)  time: 0.4299  data: 0.0051  max mem: 22448
+train: [17]  [380/400]  eta: 0:00:08  lr: 0.000014  loss: 1.9605 (1.9682)  grad: 0.2552 (0.2545)  time: 0.4266  data: 0.0051  max mem: 22448
+train: [17]  [399/400]  eta: 0:00:00  lr: 0.000013  loss: 1.9591 (1.9674)  grad: 0.2532 (0.2543)  time: 0.4305  data: 0.0049  max mem: 22448
+train: [17] Total time: 0:02:56 (0.4422 s / it)
+train: [17] Summary: lr: 0.000013  loss: 1.9591 (1.9674)  grad: 0.2532 (0.2543)
+eval (validation): [17]  [ 0/85]  eta: 0:04:32    time: 3.2004  data: 2.9211  max mem: 22448
+eval (validation): [17]  [20/85]  eta: 0:00:30    time: 0.3405  data: 0.0036  max mem: 22448
+eval (validation): [17]  [40/85]  eta: 0:00:18    time: 0.3475  data: 0.0043  max mem: 22448
+eval (validation): [17]  [60/85]  eta: 0:00:09    time: 0.3501  data: 0.0033  max mem: 22448
+eval (validation): [17]  [80/85]  eta: 0:00:01    time: 0.3302  data: 0.0041  max mem: 22448
+eval (validation): [17]  [84/85]  eta: 0:00:00    time: 0.3210  data: 0.0038  max mem: 22448
+eval (validation): [17] Total time: 0:00:32 (0.3768 s / it)
+cv: [17]  best hparam: (0.44, 1.0) (019) ('019_lr4.4e-01_wd1.0e+00')  loss: 2.421  acc: 0.269  f1: 0.213
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [18]  [  0/400]  eta: 0:21:09  lr: nan  time: 3.1748  data: 2.8438  max mem: 22448
+train: [18]  [ 20/400]  eta: 0:03:42  lr: 0.000012  loss: 1.9885 (1.9912)  grad: 0.2503 (0.2546)  time: 0.4547  data: 0.0044  max mem: 22448
+train: [18]  [ 40/400]  eta: 0:03:04  lr: 0.000012  loss: 1.9594 (1.9567)  grad: 0.2458 (0.2505)  time: 0.4359  data: 0.0045  max mem: 22448
+train: [18]  [ 60/400]  eta: 0:02:44  lr: 0.000011  loss: 1.9228 (1.9496)  grad: 0.2559 (0.2531)  time: 0.4297  data: 0.0050  max mem: 22448
+train: [18]  [ 80/400]  eta: 0:02:31  lr: 0.000011  loss: 1.9334 (1.9497)  grad: 0.2510 (0.2519)  time: 0.4439  data: 0.0050  max mem: 22448
+train: [18]  [100/400]  eta: 0:02:19  lr: 0.000010  loss: 1.9536 (1.9518)  grad: 0.2440 (0.2515)  time: 0.4321  data: 0.0049  max mem: 22448
+train: [18]  [120/400]  eta: 0:02:08  lr: 0.000009  loss: 1.9410 (1.9466)  grad: 0.2495 (0.2512)  time: 0.4266  data: 0.0050  max mem: 22448
+train: [18]  [140/400]  eta: 0:01:58  lr: 0.000009  loss: 1.9410 (1.9532)  grad: 0.2495 (0.2516)  time: 0.4281  data: 0.0050  max mem: 22448
+train: [18]  [160/400]  eta: 0:01:48  lr: 0.000008  loss: 1.9430 (1.9525)  grad: 0.2465 (0.2509)  time: 0.4313  data: 0.0049  max mem: 22448
+train: [18]  [180/400]  eta: 0:01:38  lr: 0.000008  loss: 1.9423 (1.9504)  grad: 0.2499 (0.2516)  time: 0.4297  data: 0.0048  max mem: 22448
+train: [18]  [200/400]  eta: 0:01:30  lr: 0.000007  loss: 1.9535 (1.9524)  grad: 0.2514 (0.2519)  time: 0.4532  data: 0.0049  max mem: 22448
+train: [18]  [220/400]  eta: 0:01:20  lr: 0.000007  loss: 1.9713 (1.9526)  grad: 0.2496 (0.2517)  time: 0.4427  data: 0.0051  max mem: 22448
+train: [18]  [240/400]  eta: 0:01:11  lr: 0.000006  loss: 1.9613 (1.9556)  grad: 0.2496 (0.2517)  time: 0.4390  data: 0.0050  max mem: 22448
+train: [18]  [260/400]  eta: 0:01:02  lr: 0.000006  loss: 1.9468 (1.9546)  grad: 0.2536 (0.2516)  time: 0.4251  data: 0.0047  max mem: 22448
+train: [18]  [280/400]  eta: 0:00:53  lr: 0.000006  loss: 1.9504 (1.9546)  grad: 0.2531 (0.2517)  time: 0.4405  data: 0.0048  max mem: 22448
+train: [18]  [300/400]  eta: 0:00:44  lr: 0.000005  loss: 1.9430 (1.9528)  grad: 0.2483 (0.2513)  time: 0.4443  data: 0.0048  max mem: 22448
+train: [18]  [320/400]  eta: 0:00:35  lr: 0.000005  loss: 1.9472 (1.9546)  grad: 0.2465 (0.2513)  time: 0.4287  data: 0.0051  max mem: 22448
+train: [18]  [340/400]  eta: 0:00:26  lr: 0.000004  loss: 1.9569 (1.9519)  grad: 0.2495 (0.2510)  time: 0.4342  data: 0.0050  max mem: 22448
+train: [18]  [360/400]  eta: 0:00:17  lr: 0.000004  loss: 1.9140 (1.9523)  grad: 0.2495 (0.2508)  time: 0.4356  data: 0.0050  max mem: 22448
+train: [18]  [380/400]  eta: 0:00:08  lr: 0.000004  loss: 1.9558 (1.9517)  grad: 0.2446 (0.2506)  time: 0.4352  data: 0.0049  max mem: 22448
+train: [18]  [399/400]  eta: 0:00:00  lr: 0.000003  loss: 1.9195 (1.9498)  grad: 0.2468 (0.2508)  time: 0.4390  data: 0.0049  max mem: 22448
+train: [18] Total time: 0:02:57 (0.4439 s / it)
+train: [18] Summary: lr: 0.000003  loss: 1.9195 (1.9498)  grad: 0.2468 (0.2508)
+eval (validation): [18]  [ 0/85]  eta: 0:04:27    time: 3.1435  data: 2.8590  max mem: 22448
+eval (validation): [18]  [20/85]  eta: 0:00:32    time: 0.3647  data: 0.0045  max mem: 22448
+eval (validation): [18]  [40/85]  eta: 0:00:18    time: 0.3376  data: 0.0043  max mem: 22448
+eval (validation): [18]  [60/85]  eta: 0:00:09    time: 0.3438  data: 0.0044  max mem: 22448
+eval (validation): [18]  [80/85]  eta: 0:00:01    time: 0.3270  data: 0.0039  max mem: 22448
+eval (validation): [18]  [84/85]  eta: 0:00:00    time: 0.3138  data: 0.0037  max mem: 22448
+eval (validation): [18] Total time: 0:00:32 (0.3772 s / it)
+cv: [18]  best hparam: (0.44, 1.0) (019) ('019_lr4.4e-01_wd1.0e+00')  loss: 2.427  acc: 0.269  f1: 0.213
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+train: [19]  [  0/400]  eta: 0:21:24  lr: nan  time: 3.2111  data: 2.8342  max mem: 22448
+train: [19]  [ 20/400]  eta: 0:03:37  lr: 0.000003  loss: 1.9521 (1.9621)  grad: 0.2390 (0.2447)  time: 0.4412  data: 0.0032  max mem: 22448
+train: [19]  [ 40/400]  eta: 0:03:01  lr: 0.000003  loss: 1.9458 (1.9302)  grad: 0.2442 (0.2467)  time: 0.4290  data: 0.0046  max mem: 22448
+train: [19]  [ 60/400]  eta: 0:02:41  lr: 0.000002  loss: 1.9474 (1.9474)  grad: 0.2455 (0.2464)  time: 0.4189  data: 0.0048  max mem: 22448
+train: [19]  [ 80/400]  eta: 0:02:29  lr: 0.000002  loss: 1.9741 (1.9484)  grad: 0.2408 (0.2452)  time: 0.4388  data: 0.0049  max mem: 22448
+train: [19]  [100/400]  eta: 0:02:18  lr: 0.000002  loss: 1.9455 (1.9441)  grad: 0.2416 (0.2463)  time: 0.4455  data: 0.0051  max mem: 22448
+train: [19]  [120/400]  eta: 0:02:08  lr: 0.000002  loss: 1.9455 (1.9496)  grad: 0.2490 (0.2477)  time: 0.4333  data: 0.0050  max mem: 22448
+train: [19]  [140/400]  eta: 0:01:58  lr: 0.000001  loss: 1.9682 (1.9478)  grad: 0.2444 (0.2467)  time: 0.4342  data: 0.0047  max mem: 22448
+train: [19]  [160/400]  eta: 0:01:48  lr: 0.000001  loss: 1.9453 (1.9426)  grad: 0.2387 (0.2458)  time: 0.4284  data: 0.0048  max mem: 22448
+train: [19]  [180/400]  eta: 0:01:38  lr: 0.000001  loss: 1.9092 (1.9403)  grad: 0.2445 (0.2471)  time: 0.4292  data: 0.0049  max mem: 22448
+train: [19]  [200/400]  eta: 0:01:29  lr: 0.000001  loss: 1.9296 (1.9403)  grad: 0.2503 (0.2468)  time: 0.4381  data: 0.0051  max mem: 22448
+train: [19]  [220/400]  eta: 0:01:20  lr: 0.000001  loss: 1.9318 (1.9397)  grad: 0.2451 (0.2465)  time: 0.4573  data: 0.0052  max mem: 22448
+train: [19]  [240/400]  eta: 0:01:11  lr: 0.000001  loss: 1.9282 (1.9375)  grad: 0.2459 (0.2466)  time: 0.4449  data: 0.0050  max mem: 22448
+train: [19]  [260/400]  eta: 0:01:02  lr: 0.000000  loss: 1.9013 (1.9359)  grad: 0.2455 (0.2466)  time: 0.4321  data: 0.0048  max mem: 22448
+train: [19]  [280/400]  eta: 0:00:53  lr: 0.000000  loss: 1.9291 (1.9376)  grad: 0.2447 (0.2466)  time: 0.4329  data: 0.0048  max mem: 22448
+train: [19]  [300/400]  eta: 0:00:44  lr: 0.000000  loss: 1.9407 (1.9372)  grad: 0.2481 (0.2465)  time: 0.4473  data: 0.0051  max mem: 22448
+train: [19]  [320/400]  eta: 0:00:35  lr: 0.000000  loss: 1.9500 (1.9398)  grad: 0.2483 (0.2469)  time: 0.4368  data: 0.0049  max mem: 22448
+train: [19]  [340/400]  eta: 0:00:26  lr: 0.000000  loss: 1.9930 (1.9433)  grad: 0.2454 (0.2467)  time: 0.4314  data: 0.0050  max mem: 22448
+train: [19]  [360/400]  eta: 0:00:17  lr: 0.000000  loss: 1.9676 (1.9438)  grad: 0.2433 (0.2465)  time: 0.4312  data: 0.0050  max mem: 22448
+train: [19]  [380/400]  eta: 0:00:08  lr: 0.000000  loss: 1.9368 (1.9444)  grad: 0.2435 (0.2464)  time: 0.4280  data: 0.0050  max mem: 22448
+train: [19]  [399/400]  eta: 0:00:00  lr: 0.000000  loss: 1.9541 (1.9469)  grad: 0.2468 (0.2466)  time: 0.4274  data: 0.0050  max mem: 22448
+train: [19] Total time: 0:02:57 (0.4428 s / it)
+train: [19] Summary: lr: 0.000000  loss: 1.9541 (1.9469)  grad: 0.2468 (0.2466)
+eval (validation): [19]  [ 0/85]  eta: 0:04:53    time: 3.4475  data: 3.1624  max mem: 22448
+eval (validation): [19]  [20/85]  eta: 0:00:35    time: 0.3951  data: 0.0035  max mem: 22448
+eval (validation): [19]  [40/85]  eta: 0:00:19    time: 0.3319  data: 0.0041  max mem: 22448
+eval (validation): [19]  [60/85]  eta: 0:00:10    time: 0.3416  data: 0.0042  max mem: 22448
+eval (validation): [19]  [80/85]  eta: 0:00:01    time: 0.3229  data: 0.0042  max mem: 22448
+eval (validation): [19]  [84/85]  eta: 0:00:00    time: 0.3161  data: 0.0041  max mem: 22448
+eval (validation): [19] Total time: 0:00:32 (0.3857 s / it)
+cv: [19]  best hparam: (0.44, 1.0) (019) ('019_lr4.4e-01_wd1.0e+00')  loss: 2.426  acc: 0.269  f1: 0.213
+saving checkpoint experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+evaluating last checkpoint: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-last.pth
+eval model info:
+{"score": 0.26891842008121075, "hparam": [0.44, 1.0], "hparam_id": 19, "epoch": 19, "is_best": false, "best_score": 0.27593207825765964}
+eval (train): [20]  [  0/509]  eta: 0:25:55    time: 3.0559  data: 2.7559  max mem: 22448
+eval (train): [20]  [ 20/509]  eta: 0:03:54    time: 0.3511  data: 0.0038  max mem: 22448
+eval (train): [20]  [ 40/509]  eta: 0:03:14    time: 0.3459  data: 0.0042  max mem: 22448
+eval (train): [20]  [ 60/509]  eta: 0:02:54    time: 0.3334  data: 0.0042  max mem: 22448
+eval (train): [20]  [ 80/509]  eta: 0:02:40    time: 0.3298  data: 0.0042  max mem: 22448
+eval (train): [20]  [100/509]  eta: 0:02:29    time: 0.3366  data: 0.0040  max mem: 22448
+eval (train): [20]  [120/509]  eta: 0:02:21    time: 0.3469  data: 0.0042  max mem: 22448
+eval (train): [20]  [140/509]  eta: 0:02:12    time: 0.3273  data: 0.0039  max mem: 22448
+eval (train): [20]  [160/509]  eta: 0:02:04    time: 0.3401  data: 0.0044  max mem: 22448
+eval (train): [20]  [180/509]  eta: 0:01:55    time: 0.3248  data: 0.0039  max mem: 22448
+eval (train): [20]  [200/509]  eta: 0:01:48    time: 0.3303  data: 0.0040  max mem: 22448
+eval (train): [20]  [220/509]  eta: 0:01:40    time: 0.3337  data: 0.0040  max mem: 22448
+eval (train): [20]  [240/509]  eta: 0:01:33    time: 0.3291  data: 0.0040  max mem: 22448
+eval (train): [20]  [260/509]  eta: 0:01:26    time: 0.3361  data: 0.0039  max mem: 22448
+eval (train): [20]  [280/509]  eta: 0:01:19    time: 0.3551  data: 0.0040  max mem: 22448
+eval (train): [20]  [300/509]  eta: 0:01:12    time: 0.3491  data: 0.0043  max mem: 22448
+eval (train): [20]  [320/509]  eta: 0:01:05    time: 0.3465  data: 0.0042  max mem: 22448
+eval (train): [20]  [340/509]  eta: 0:00:58    time: 0.3358  data: 0.0041  max mem: 22448
+eval (train): [20]  [360/509]  eta: 0:00:51    time: 0.3209  data: 0.0040  max mem: 22448
+eval (train): [20]  [380/509]  eta: 0:00:44    time: 0.3465  data: 0.0041  max mem: 22448
+eval (train): [20]  [400/509]  eta: 0:00:37    time: 0.3381  data: 0.0039  max mem: 22448
+eval (train): [20]  [420/509]  eta: 0:00:30    time: 0.3353  data: 0.0040  max mem: 22448
+eval (train): [20]  [440/509]  eta: 0:00:23    time: 0.3346  data: 0.0040  max mem: 22448
+eval (train): [20]  [460/509]  eta: 0:00:16    time: 0.3440  data: 0.0041  max mem: 22448
+eval (train): [20]  [480/509]  eta: 0:00:09    time: 0.3370  data: 0.0041  max mem: 22448
+eval (train): [20]  [500/509]  eta: 0:00:03    time: 0.3214  data: 0.0037  max mem: 22448
+eval (train): [20]  [508/509]  eta: 0:00:00    time: 0.3116  data: 0.0037  max mem: 22448
+eval (train): [20] Total time: 0:02:54 (0.3435 s / it)
+eval (validation): [20]  [ 0/85]  eta: 0:04:05    time: 2.8839  data: 2.6023  max mem: 22448
+eval (validation): [20]  [20/85]  eta: 0:00:32    time: 0.3778  data: 0.0043  max mem: 22448
+eval (validation): [20]  [40/85]  eta: 0:00:18    time: 0.3433  data: 0.0042  max mem: 22448
+eval (validation): [20]  [60/85]  eta: 0:00:09    time: 0.3443  data: 0.0045  max mem: 22448
+eval (validation): [20]  [80/85]  eta: 0:00:01    time: 0.3363  data: 0.0041  max mem: 22448
+eval (validation): [20]  [84/85]  eta: 0:00:00    time: 0.3225  data: 0.0039  max mem: 22448
+eval (validation): [20] Total time: 0:00:32 (0.3813 s / it)
+eval (test): [20]  [ 0/85]  eta: 0:04:03    time: 2.8681  data: 2.5820  max mem: 22448
+eval (test): [20]  [20/85]  eta: 0:00:30    time: 0.3488  data: 0.0061  max mem: 22448
+eval (test): [20]  [40/85]  eta: 0:00:18    time: 0.3481  data: 0.0037  max mem: 22448
+eval (test): [20]  [60/85]  eta: 0:00:09    time: 0.3379  data: 0.0044  max mem: 22448
+eval (test): [20]  [80/85]  eta: 0:00:01    time: 0.3268  data: 0.0041  max mem: 22448
+eval (test): [20]  [84/85]  eta: 0:00:00    time: 0.3078  data: 0.0038  max mem: 22448
+eval (test): [20] Total time: 0:00:31 (0.3705 s / it)
+eval (testid): [20]  [ 0/82]  eta: 0:03:55    time: 2.8780  data: 2.6308  max mem: 22448
+eval (testid): [20]  [20/82]  eta: 0:00:30    time: 0.3785  data: 0.0056  max mem: 22448
+eval (testid): [20]  [40/82]  eta: 0:00:17    time: 0.3409  data: 0.0038  max mem: 22448
+eval (testid): [20]  [60/82]  eta: 0:00:08    time: 0.3311  data: 0.0041  max mem: 22448
+eval (testid): [20]  [80/82]  eta: 0:00:00    time: 0.3194  data: 0.0045  max mem: 22448
+eval (testid): [20]  [81/82]  eta: 0:00:00    time: 0.3070  data: 0.0043  max mem: 22448
+eval (testid): [20] Total time: 0:00:30 (0.3742 s / it)
+evaluating best checkpoint: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/nsd_cococlip__patch__attn/checkpoint-best.pth
+eval model info:
+{"score": 0.27593207825765964, "hparam": [1.2, 1.0], "hparam_id": 25, "epoch": 5, "is_best": true, "best_score": 0.27593207825765964}
+eval (train): [20]  [  0/509]  eta: 0:24:27    time: 2.8829  data: 2.6076  max mem: 22448
+eval (train): [20]  [ 20/509]  eta: 0:03:36    time: 0.3199  data: 0.0041  max mem: 22448
+eval (train): [20]  [ 40/509]  eta: 0:03:07    time: 0.3533  data: 0.0039  max mem: 22448
+eval (train): [20]  [ 60/509]  eta: 0:02:50    time: 0.3388  data: 0.0040  max mem: 22448
+eval (train): [20]  [ 80/509]  eta: 0:02:37    time: 0.3346  data: 0.0040  max mem: 22448
+eval (train): [20]  [100/509]  eta: 0:02:30    time: 0.3614  data: 0.0045  max mem: 22448
+eval (train): [20]  [120/509]  eta: 0:02:22    time: 0.3636  data: 0.0040  max mem: 22448
+eval (train): [20]  [140/509]  eta: 0:02:14    time: 0.3478  data: 0.0043  max mem: 22448
+eval (train): [20]  [160/509]  eta: 0:02:05    time: 0.3374  data: 0.0042  max mem: 22448
+eval (train): [20]  [180/509]  eta: 0:01:57    time: 0.3357  data: 0.0038  max mem: 22448
+eval (train): [20]  [200/509]  eta: 0:01:51    time: 0.3797  data: 0.0045  max mem: 22448
+eval (train): [20]  [220/509]  eta: 0:01:43    time: 0.3514  data: 0.0043  max mem: 22448
+eval (train): [20]  [240/509]  eta: 0:01:36    time: 0.3405  data: 0.0042  max mem: 22448
+eval (train): [20]  [260/509]  eta: 0:01:28    time: 0.3309  data: 0.0039  max mem: 22448
+eval (train): [20]  [280/509]  eta: 0:01:21    time: 0.3479  data: 0.0041  max mem: 22448
+eval (train): [20]  [300/509]  eta: 0:01:13    time: 0.3356  data: 0.0039  max mem: 22448
+eval (train): [20]  [320/509]  eta: 0:01:06    time: 0.3403  data: 0.0041  max mem: 22448
+eval (train): [20]  [340/509]  eta: 0:00:59    time: 0.3350  data: 0.0041  max mem: 22448
+eval (train): [20]  [360/509]  eta: 0:00:52    time: 0.3574  data: 0.0043  max mem: 22448
+eval (train): [20]  [380/509]  eta: 0:00:45    time: 0.3488  data: 0.0042  max mem: 22448
+eval (train): [20]  [400/509]  eta: 0:00:38    time: 0.3595  data: 0.0042  max mem: 22448
+eval (train): [20]  [420/509]  eta: 0:00:31    time: 0.3550  data: 0.0042  max mem: 22448
+eval (train): [20]  [440/509]  eta: 0:00:24    time: 0.3504  data: 0.0039  max mem: 22448
+eval (train): [20]  [460/509]  eta: 0:00:17    time: 0.3659  data: 0.0043  max mem: 22448
+eval (train): [20]  [480/509]  eta: 0:00:10    time: 0.3611  data: 0.0046  max mem: 22448
+eval (train): [20]  [500/509]  eta: 0:00:03    time: 0.3493  data: 0.0041  max mem: 22448
+eval (train): [20]  [508/509]  eta: 0:00:00    time: 0.3269  data: 0.0040  max mem: 22448
+eval (train): [20] Total time: 0:03:00 (0.3540 s / it)
+eval (validation): [20]  [ 0/85]  eta: 0:04:20    time: 3.0655  data: 2.7646  max mem: 22448
+eval (validation): [20]  [20/85]  eta: 0:00:33    time: 0.3859  data: 0.0039  max mem: 22448
+eval (validation): [20]  [40/85]  eta: 0:00:19    time: 0.3524  data: 0.0041  max mem: 22448
+eval (validation): [20]  [60/85]  eta: 0:00:10    time: 0.3441  data: 0.0043  max mem: 22448
+eval (validation): [20]  [80/85]  eta: 0:00:01    time: 0.3343  data: 0.0044  max mem: 22448
+eval (validation): [20]  [84/85]  eta: 0:00:00    time: 0.3299  data: 0.0043  max mem: 22448
+eval (validation): [20] Total time: 0:00:32 (0.3878 s / it)
+eval (test): [20]  [ 0/85]  eta: 0:04:16    time: 3.0215  data: 2.7859  max mem: 22448
+eval (test): [20]  [20/85]  eta: 0:00:32    time: 0.3714  data: 0.0041  max mem: 22448
+eval (test): [20]  [40/85]  eta: 0:00:19    time: 0.3583  data: 0.0036  max mem: 22448
+eval (test): [20]  [60/85]  eta: 0:00:10    time: 0.3543  data: 0.0039  max mem: 22448
+eval (test): [20]  [80/85]  eta: 0:00:01    time: 0.3334  data: 0.0044  max mem: 22448
+eval (test): [20]  [84/85]  eta: 0:00:00    time: 0.3256  data: 0.0041  max mem: 22448
+eval (test): [20] Total time: 0:00:32 (0.3877 s / it)
+eval (testid): [20]  [ 0/82]  eta: 0:04:09    time: 3.0447  data: 2.7476  max mem: 22448
+eval (testid): [20]  [20/82]  eta: 0:00:32    time: 0.4034  data: 0.0063  max mem: 22448
+eval (testid): [20]  [40/82]  eta: 0:00:18    time: 0.3416  data: 0.0035  max mem: 22448
+eval (testid): [20]  [60/82]  eta: 0:00:08    time: 0.3420  data: 0.0042  max mem: 22448
+eval (testid): [20]  [80/82]  eta: 0:00:00    time: 0.3271  data: 0.0044  max mem: 22448
+eval (testid): [20]  [81/82]  eta: 0:00:00    time: 0.3174  data: 0.0042  max mem: 22448
+eval (testid): [20] Total time: 0:00:31 (0.3871 s / it)
+eval results:
+
+| model    | repr   | clf   | dataset      | ckpt   |   epoch |      lr |   wd |   hparam_id | hparam     | split      |   loss |     acc |   acc_std |      f1 |    f1_std |
+|:---------|:-------|:------|:-------------|:-------|--------:|--------:|-----:|------------:|:-----------|:-----------|-------:|--------:|----------:|--------:|----------:|
+| flat_mae | patch  | attn  | nsd_cococlip | best   |       5 | 0.00036 | 0.05 |          25 | [1.2, 1.0] | train      | 2.0279 | 0.38606 | 0.0024391 | 0.32683 | 0.0026072 |
+| flat_mae | patch  | attn  | nsd_cococlip | best   |       5 | 0.00036 | 0.05 |          25 | [1.2, 1.0] | validation | 2.3861 | 0.27593 | 0.0054005 | 0.20649 | 0.0047691 |
+| flat_mae | patch  | attn  | nsd_cococlip | best   |       5 | 0.00036 | 0.05 |          25 | [1.2, 1.0] | test       | 2.3202 | 0.29944 | 0.0052928 | 0.23452 | 0.0053526 |
+| flat_mae | patch  | attn  | nsd_cococlip | best   |       5 | 0.00036 | 0.05 |          25 | [1.2, 1.0] | testid     | 2.2755 | 0.30364 | 0.0059794 | 0.24622 | 0.0057295 |
+
+
+done! total time: 1:22:23

data_scaling/n400_1/eval_v2/ppmi_dx__patch__logistic/config.yaml

@@ -0,0 +1,30 @@
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (ppmi_dx patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_1/eval_v2/ppmi_dx__patch__logistic
+model: flat_mae
+representation: patch
+dataset: ppmi_dx
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/ppmi_dx__patch__logistic
+remote_dir: null

data_scaling/n400_1/eval_v2/ppmi_dx__patch__logistic/eval_table.csv

@@ -0,0 +1,203 @@
+model,repr,clf,dataset,trial,C,split,acc,acc_std,f1,f1_std,bacc,bacc_std
+flat_mae,patch,logistic,ppmi_dx,,0.3593813663804626,train,0.9430604982206405,0.009718841663704219,0.9394070080862533,0.010425699278079526,0.9356708742402993,0.011094847172679138
+flat_mae,patch,logistic,ppmi_dx,,0.3593813663804626,test,0.63,0.04524961436299762,0.5906626839252129,0.04844622125624765,0.5892320892320893,0.04684232537859296
+flat_mae,patch,logistic,ppmi_dx,1,0.046415888336127774,train,0.8327402135231317,0.015080358703522843,0.8150072837292694,0.017395623217074076,0.80415061014772,0.017437234468122915
+flat_mae,patch,logistic,ppmi_dx,1,0.046415888336127774,test,0.62,0.04371370036956378,0.5766488413547237,0.04965517929109973,0.5764006791171477,0.046718817119045165
+flat_mae,patch,logistic,ppmi_dx,2,0.3593813663804626,train,0.9501779359430605,0.009144294829924085,0.9466771487048808,0.009939576492254692,0.9412732819524727,0.010996322012108933
+flat_mae,patch,logistic,ppmi_dx,2,0.3593813663804626,test,0.59,0.04961217189359885,0.5626666666666666,0.05281989014396253,0.5623938879456706,0.05229097998959533
+flat_mae,patch,logistic,ppmi_dx,3,0.046415888336127774,train,0.8185053380782918,0.015093188743586058,0.7961798839458414,0.018017797429398127,0.7838926354099764,0.01776826758149463
+flat_mae,patch,logistic,ppmi_dx,3,0.046415888336127774,test,0.69,0.04351703574463683,0.6615351020853806,0.048463542592696605,0.6583191850594228,0.047252174901520914
+flat_mae,patch,logistic,ppmi_dx,4,0.005994842503189409,train,0.7348754448398577,0.016959778826558202,0.6894209977783465,0.02213521818632936,0.6829238921001927,0.019835126964382666
+flat_mae,patch,logistic,ppmi_dx,4,0.005994842503189409,test,0.7,0.04106226978626486,0.66078697421981,0.047734870379201884,0.6561969439728353,0.044782648390475896
+flat_mae,patch,logistic,ppmi_dx,5,0.046415888336127774,train,0.8291814946619217,0.015311398136063293,0.81197724991636,0.01746619835113448,0.8021301648469279,0.017490164245246728
+flat_mae,patch,logistic,ppmi_dx,5,0.046415888336127774,test,0.56,0.050015721528335465,0.5280995280995281,0.05238886009304962,0.5280135823429541,0.05173854732883097
+flat_mae,patch,logistic,ppmi_dx,6,0.046415888336127774,train,0.8327402135231317,0.01501652859356584,0.8145527051125434,0.017436666838212458,0.8032808820381074,0.01746765581275586
+flat_mae,patch,logistic,ppmi_dx,6,0.046415888336127774,test,0.58,0.04470615170197497,0.525101763907734,0.051937224805165036,0.5288624787775891,0.04761150502647497
+flat_mae,patch,logistic,ppmi_dx,7,0.046415888336127774,train,0.8291814946619217,0.01576078240050652,0.8137016574585636,0.017659008075572125,0.8056090772853779,0.017861534345425395
+flat_mae,patch,logistic,ppmi_dx,7,0.046415888336127774,test,0.55,0.051013896146050244,0.5366079703429101,0.05184591645257948,0.5403225806451613,0.05342905332313932
+flat_mae,patch,logistic,ppmi_dx,8,0.005994842503189409,train,0.7188612099644128,0.01666144094736625,0.6732365756005653,0.020869094192100066,0.6681786555341469,0.018831212355544823
+flat_mae,patch,logistic,ppmi_dx,8,0.005994842503189409,test,0.62,0.038295294750138695,0.5386109762020399,0.048785906087796614,0.5509337860780985,0.04092566604131759
+flat_mae,patch,logistic,ppmi_dx,9,0.005994842503189409,train,0.7277580071174378,0.015475975553370755,0.681083306443537,0.019941341366071848,0.6754040890601585,0.017925887506499892
+flat_mae,patch,logistic,ppmi_dx,9,0.005994842503189409,test,0.71,0.03829867360627519,0.6579785352046232,0.04956219876842803,0.6540747028862479,0.0439527155611839
+flat_mae,patch,logistic,ppmi_dx,10,0.005994842503189409,train,0.7295373665480427,0.01697580181587017,0.6846659283868586,0.021828051478280285,0.6785886319845857,0.01969538551503112
+flat_mae,patch,logistic,ppmi_dx,10,0.005994842503189409,test,0.6,0.04425717568937268,0.5604395604395604,0.04849572941282711,0.5602716468590832,0.04660202522938627
+flat_mae,patch,logistic,ppmi_dx,11,2.782559402207126,train,0.99644128113879,0.0024616330018922554,0.9962334964144495,0.0026125582400898266,0.9953703703703703,0.0032024021922764947
+flat_mae,patch,logistic,ppmi_dx,11,2.782559402207126,test,0.7,0.044095732219796506,0.6782496782496783,0.047327395321723906,0.6765704584040747,0.04682570102858115
+flat_mae,patch,logistic,ppmi_dx,12,0.3593813663804626,train,0.9323843416370107,0.010628278765627554,0.9280458221024259,0.011423263255464645,0.9250829586812246,0.012079225946649973
+flat_mae,patch,logistic,ppmi_dx,12,0.3593813663804626,test,0.65,0.044211034821637005,0.6072270227808326,0.050228756052448094,0.6056876061120543,0.04736082615996879
+flat_mae,patch,logistic,ppmi_dx,13,0.3593813663804626,train,0.9644128113879004,0.007974016992655047,0.9622676979267375,0.008485012194154571,0.9606615285806037,0.009025187819360538
+flat_mae,patch,logistic,ppmi_dx,13,0.3593813663804626,test,0.61,0.04650550074991129,0.568536342515765,0.05122102941892769,0.5683361629881154,0.04890203025188589
+flat_mae,patch,logistic,ppmi_dx,14,0.046415888336127774,train,0.8291814946619217,0.014763273030379194,0.8110712684894665,0.017003237641987922,0.8003907086277029,0.01702512325496237
+flat_mae,patch,logistic,ppmi_dx,14,0.046415888336127774,test,0.57,0.04531796994570696,0.5305164319248826,0.04858880543837976,0.5309847198641766,0.0469430625270563
+flat_mae,patch,logistic,ppmi_dx,15,0.005994842503189409,train,0.7384341637010676,0.0156139498736111,0.6926047887481164,0.020080357747320064,0.6858140655105973,0.018013400814026113
+flat_mae,patch,logistic,ppmi_dx,15,0.005994842503189409,test,0.66,0.04194617503420306,0.609375,0.0507321062318044,0.6086587436332768,0.04615700467813121
+flat_mae,patch,logistic,ppmi_dx,16,0.046415888336127774,train,0.8309608540925267,0.014421983812988896,0.810417620970176,0.017133157851359847,0.7974871547848427,0.0170838961895997
+flat_mae,patch,logistic,ppmi_dx,16,0.046415888336127774,test,0.65,0.045189817437117405,0.6224786970121885,0.0490666338031973,0.6209677419354839,0.04831710061727428
+flat_mae,patch,logistic,ppmi_dx,17,0.046415888336127774,train,0.8131672597864769,0.015307182513498819,0.793106349857478,0.017802407520403216,0.7830362877328195,0.017739950841352273
+flat_mae,patch,logistic,ppmi_dx,17,0.046415888336127774,test,0.59,0.04571877513669849,0.5464100011063171,0.050084490057855875,0.5471137521222411,0.047829283307919995
+flat_mae,patch,logistic,ppmi_dx,18,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,18,1291.5496650148827,test,0.57,0.049609978834907795,0.557203171661003,0.05066544682878853,0.5615449915110357,0.05224853446061583
+flat_mae,patch,logistic,ppmi_dx,19,0.046415888336127774,train,0.8167259786476868,0.015128670572605219,0.7960345732779428,0.017554172514366553,0.7850567330336116,0.01738250181476914
+flat_mae,patch,logistic,ppmi_dx,19,0.046415888336127774,test,0.71,0.044014888390179974,0.6938021328265231,0.04620432585274497,0.6948217317487266,0.046390377377879745
+flat_mae,patch,logistic,ppmi_dx,20,0.046415888336127774,train,0.8131672597864769,0.015680755179614324,0.7936099889827396,0.01814515283690319,0.783906015842432,0.018076972012768755
+flat_mae,patch,logistic,ppmi_dx,20,0.046415888336127774,test,0.67,0.041664848493664294,0.6176572818908586,0.050292728202088126,0.616723259762309,0.04547365133574492
+flat_mae,patch,logistic,ppmi_dx,21,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,21,1291.5496650148827,test,0.61,0.04743608752837865,0.5882166613873931,0.049491386541767775,0.5887096774193548,0.049504562984715186
+flat_mae,patch,logistic,ppmi_dx,22,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,22,21.54434690031882,test,0.55,0.04869856260712424,0.5248653785239151,0.050756742363550975,0.5250424448217317,0.051113161301739526
+flat_mae,patch,logistic,ppmi_dx,23,0.005994842503189409,train,0.7348754448398577,0.016159836316410485,0.6913715387195336,0.020499061851053477,0.6846633483194177,0.018597363317695573
+flat_mae,patch,logistic,ppmi_dx,23,0.005994842503189409,test,0.66,0.04496213518061614,0.609375,0.054040822047925936,0.6086587436332768,0.04955820404901081
+flat_mae,patch,logistic,ppmi_dx,24,0.005994842503189409,train,0.7295373665480427,0.01610486179300103,0.683671051072402,0.020812465992023643,0.6777189038749732,0.018659874099308298
+flat_mae,patch,logistic,ppmi_dx,24,0.005994842503189409,test,0.63,0.04088236294540716,0.5636277862955537,0.050203991210757234,0.5691850594227504,0.04412904201521217
+flat_mae,patch,logistic,ppmi_dx,25,0.3593813663804626,train,0.9430604982206405,0.009744572206034944,0.9392932852648492,0.01050075122046031,0.9354929351316634,0.011278970954382815
+flat_mae,patch,logistic,ppmi_dx,25,0.3593813663804626,test,0.64,0.03998951862676019,0.5792426367461431,0.05036360647994039,0.5823429541595926,0.044420179846896654
+flat_mae,patch,logistic,ppmi_dx,26,0.3593813663804626,train,0.9377224199288257,0.009933716699858937,0.9339087894254063,0.010572397379765579,0.932027403125669,0.01095488964532246
+flat_mae,patch,logistic,ppmi_dx,26,0.3593813663804626,test,0.64,0.04565597879796249,0.6179966044142615,0.048775255708017544,0.6179966044142615,0.04896193072503621
+flat_mae,patch,logistic,ppmi_dx,27,0.005994842503189409,train,0.7206405693950177,0.017014390443075404,0.6706248623466214,0.022021648460014594,0.6661448298008992,0.019547318952198203
+flat_mae,patch,logistic,ppmi_dx,27,0.005994842503189409,test,0.62,0.041886183879651775,0.5476190476190476,0.052077775167443024,0.5560271646859083,0.04543871301389427
+flat_mae,patch,logistic,ppmi_dx,28,0.3593813663804626,train,0.9466192170818505,0.009420392644202937,0.9431940700808625,0.010110280696271377,0.9401225647612932,0.010838644573089671
+flat_mae,patch,logistic,ppmi_dx,28,0.3593813663804626,test,0.62,0.04606073816169255,0.5924495924495925,0.04968790442111471,0.5916808149405772,0.04891071030375556
+flat_mae,patch,logistic,ppmi_dx,29,0.046415888336127774,train,0.8256227758007118,0.01509909349376749,0.8076014811709635,0.017420886140189537,0.7975005352172982,0.01750987381034677
+flat_mae,patch,logistic,ppmi_dx,29,0.046415888336127774,test,0.63,0.047497974693664576,0.5906626839252129,0.05315077608236612,0.5895585738539898,0.0506647830186329
+flat_mae,patch,logistic,ppmi_dx,30,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,30,21.54434690031882,test,0.61,0.04777488461524528,0.584,0.051627396178459475,0.583616298811545,0.05144063503619579
+flat_mae,patch,logistic,ppmi_dx,31,0.046415888336127774,train,0.8345195729537367,0.015635418323509544,0.8162979104792366,0.01830563009451841,0.8047259687433097,0.0183997827499793
+flat_mae,patch,logistic,ppmi_dx,31,0.046415888336127774,test,0.66,0.04606840131804012,0.6353496353496353,0.04849142693462424,0.634125636672326,0.04791073543150732
+flat_mae,patch,logistic,ppmi_dx,32,0.046415888336127774,train,0.8167259786476868,0.015312667310758022,0.7980272362815463,0.017608889213476066,0.7885356454720616,0.01764679882154443
+flat_mae,patch,logistic,ppmi_dx,32,0.046415888336127774,test,0.67,0.04428216345211692,0.6440513428972063,0.048367590788555063,0.6421901528013583,0.04776257510752226
+flat_mae,patch,logistic,ppmi_dx,33,0.3593813663804626,train,0.9483985765124555,0.008946409711688709,0.9448264188628785,0.009691096253634918,0.9398281952472705,0.010571857307610172
+flat_mae,patch,logistic,ppmi_dx,33,0.3593813663804626,test,0.66,0.04723928449923855,0.6353496353496353,0.05120044521088683,0.634125636672326,0.05076863277495403
+flat_mae,patch,logistic,ppmi_dx,34,0.046415888336127774,train,0.8202846975088968,0.015024385053951933,0.7989693825149191,0.01771803619553632,0.7870771783344037,0.01750322001006957
+flat_mae,patch,logistic,ppmi_dx,34,0.046415888336127774,test,0.67,0.044037125246773313,0.6296711929076422,0.0515727101953694,0.6269100169779287,0.04865137855127997
+flat_mae,patch,logistic,ppmi_dx,35,0.3593813663804626,train,0.9412811387900356,0.010075971217844312,0.9376854300296689,0.010750475389341107,0.9357873046456862,0.011284628334793772
+flat_mae,patch,logistic,ppmi_dx,35,0.3593813663804626,test,0.6,0.049603528100327705,0.5659722222222222,0.053085290658552796,0.565365025466893,0.05195229570282048
+flat_mae,patch,logistic,ppmi_dx,36,0.005994842503189409,train,0.7241992882562278,0.016256035720051614,0.6818701404942681,0.020013772081335056,0.6759928280882038,0.01829930510340109
+flat_mae,patch,logistic,ppmi_dx,36,0.005994842503189409,test,0.64,0.04114984811636612,0.5792426367461431,0.04936410027607919,0.5823429541595926,0.04417171930517304
+flat_mae,patch,logistic,ppmi_dx,37,0.3593813663804626,train,0.9448398576512456,0.009581040783662042,0.9410213443016977,0.010406432563138118,0.9360682937272533,0.011432763489264777
+flat_mae,patch,logistic,ppmi_dx,37,0.3593813663804626,test,0.65,0.04500335987456937,0.630450849963045,0.04682204692793159,0.6311544991511036,0.04707293583689922
+flat_mae,patch,logistic,ppmi_dx,38,0.046415888336127774,train,0.802491103202847,0.016025263106609147,0.7807426673461857,0.01865338032541245,0.7708868550631557,0.018439236378183056
+flat_mae,patch,logistic,ppmi_dx,38,0.046415888336127774,test,0.7,0.04530611879205721,0.6847414880201765,0.04735395224066736,0.6867572156196944,0.04779090071698159
+flat_mae,patch,logistic,ppmi_dx,39,0.046415888336127774,train,0.8238434163701067,0.014379240812025287,0.8034569366581532,0.017168263000953813,0.7917068079640335,0.01727227235581512
+flat_mae,patch,logistic,ppmi_dx,39,0.046415888336127774,test,0.67,0.0459393121411281,0.6396986570586308,0.05064134614606536,0.6370967741935484,0.04919403747059671
+flat_mae,patch,logistic,ppmi_dx,40,0.000774263682681127,train,0.6690391459074733,0.014710404241846966,0.5771359223300971,0.02179916640790788,0.592927103403982,0.01678211201662067
+flat_mae,patch,logistic,ppmi_dx,40,0.000774263682681127,test,0.66,0.03585707182690745,0.5687468290208015,0.052414192804968146,0.5831918505942275,0.04086209520637194
+flat_mae,patch,logistic,ppmi_dx,41,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,41,2.782559402207126,test,0.6,0.05132166793860075,0.586606035551881,0.05222995387608565,0.5908319185059423,0.05319657754243219
+flat_mae,patch,logistic,ppmi_dx,42,0.046415888336127774,train,0.8185053380782918,0.014901480731364722,0.7997484804024313,0.017089083330025904,0.789980732177264,0.01700703304554459
+flat_mae,patch,logistic,ppmi_dx,42,0.046415888336127774,test,0.68,0.039683119837028935,0.6259934548854604,0.04975199597504042,0.6247877758913413,0.04432757335355691
+flat_mae,patch,logistic,ppmi_dx,43,0.046415888336127774,train,0.8202846975088968,0.015517522535981388,0.8028789220537931,0.017493542997013663,0.7940350032113038,0.01746345309858463
+flat_mae,patch,logistic,ppmi_dx,43,0.046415888336127774,test,0.6,0.04425144065451429,0.554367201426025,0.049556954063591244,0.5551782682512734,0.04696458801086889
+flat_mae,patch,logistic,ppmi_dx,44,0.005994842503189409,train,0.7259786476868327,0.016160375101868373,0.6784849250338054,0.020865997041025323,0.6730892742453436,0.018618991872925438
+flat_mae,patch,logistic,ppmi_dx,44,0.005994842503189409,test,0.68,0.04148917449166709,0.6259934548854604,0.05256835054930224,0.6247877758913413,0.046974610754166264
+flat_mae,patch,logistic,ppmi_dx,45,0.046415888336127774,train,0.8291814946619217,0.015932318763144656,0.8124191641749531,0.01813821104826984,0.8029998929565403,0.01825036332610028
+flat_mae,patch,logistic,ppmi_dx,45,0.046415888336127774,test,0.66,0.044687465804182716,0.6212121212121212,0.050608278902441156,0.6188455008488964,0.04811151937006594
+flat_mae,patch,logistic,ppmi_dx,46,0.046415888336127774,train,0.8309608540925267,0.01519882580738319,0.8145890851001025,0.017217367397728178,0.8053147077713552,0.017340847376839323
+flat_mae,patch,logistic,ppmi_dx,46,0.046415888336127774,test,0.63,0.0459838058451016,0.5783475783475784,0.0538732946615018,0.5793718166383701,0.04953516975815706
+flat_mae,patch,logistic,ppmi_dx,47,0.046415888336127774,train,0.8149466192170819,0.015369979045956847,0.79630868740939,0.017497945648126473,0.7870905587668593,0.017456238975380085
+flat_mae,patch,logistic,ppmi_dx,47,0.046415888336127774,test,0.7,0.04302354704112621,0.6744791666666667,0.04795109699290812,0.6714770797962648,0.047428645864378505
+flat_mae,patch,logistic,ppmi_dx,48,0.005994842503189409,train,0.7295373665480427,0.015485233024994357,0.681633870005963,0.020212568440982546,0.6759794476557482,0.018024790095250504
+flat_mae,patch,logistic,ppmi_dx,48,0.005994842503189409,test,0.64,0.042187291925412804,0.5989304812834224,0.048709334146665534,0.597623089983022,0.04596119541178976
+flat_mae,patch,logistic,ppmi_dx,49,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,49,2.782559402207126,test,0.65,0.04786762998102162,0.6224786970121885,0.05126797979686403,0.6209677419354839,0.050436386082368084
+flat_mae,patch,logistic,ppmi_dx,50,0.3593813663804626,train,0.9377224199288257,0.00949247081591274,0.9332810978939621,0.010292239671853408,0.9276787625776065,0.011039977367040573
+flat_mae,patch,logistic,ppmi_dx,50,0.3593813663804626,test,0.66,0.04265457068122946,0.6155585707824514,0.04908085042653531,0.6137521222410866,0.04579370911799187
+flat_mae,patch,logistic,ppmi_dx,51,0.3593813663804626,train,0.9466192170818505,0.008857493253610566,0.9431940700808625,0.009498651655257224,0.9401225647612932,0.010140322276801693
+flat_mae,patch,logistic,ppmi_dx,51,0.3593813663804626,test,0.58,0.04964866564168668,0.5442708333333334,0.053437358865708384,0.5441426146010186,0.05207740706125955
+flat_mae,patch,logistic,ppmi_dx,52,0.046415888336127774,train,0.8185053380782918,0.015765727035554462,0.7987699566114832,0.01819632119014533,0.7882412759580391,0.01803627334891838
+flat_mae,patch,logistic,ppmi_dx,52,0.046415888336127774,test,0.62,0.046995046547482,0.5824175824175825,0.051826829456541615,0.5814940577249575,0.04979921677443064
+flat_mae,patch,logistic,ppmi_dx,53,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,53,2.782559402207126,test,0.55,0.04745463517929517,0.5396419437340154,0.04775294687403358,0.5454159592529711,0.049309415309893
+flat_mae,patch,logistic,ppmi_dx,54,0.005994842503189409,train,0.7330960854092526,0.01662678087686323,0.6925915661420424,0.020443681124661533,0.6858274459430529,0.018820138894408287
+flat_mae,patch,logistic,ppmi_dx,54,0.005994842503189409,test,0.55,0.04605473265583027,0.5021573182874212,0.04786812136845858,0.5046689303904923,0.04612370514877537
+flat_mae,patch,logistic,ppmi_dx,55,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,55,21.54434690031882,test,0.6,0.04482849094047222,0.5833333333333333,0.046418273462788096,0.5857385398981324,0.04725411490121178
+flat_mae,patch,logistic,ppmi_dx,56,0.046415888336127774,train,0.8042704626334519,0.015458925223581173,0.7818969531900481,0.018230261851414086,0.7714622136587455,0.01800915203661604
+flat_mae,patch,logistic,ppmi_dx,56,0.046415888336127774,test,0.64,0.046751337948768915,0.592944369063772,0.05434622524608595,0.5925297113752122,0.05029172201083732
+flat_mae,patch,logistic,ppmi_dx,57,0.046415888336127774,train,0.797153024911032,0.01582791733091751,0.7727988425039363,0.01871929624511709,0.7622029543994862,0.018278060787021662
+flat_mae,patch,logistic,ppmi_dx,57,0.046415888336127774,test,0.7,0.040116829386181545,0.6493688639551192,0.0507632949920872,0.6460101867572157,0.04553371611500591
+flat_mae,patch,logistic,ppmi_dx,58,0.046415888336127774,train,0.8238434163701067,0.014439234160343994,0.8034569366581532,0.016928909779322685,0.7917068079640335,0.016781187381561622
+flat_mae,patch,logistic,ppmi_dx,58,0.046415888336127774,test,0.67,0.04486856806273184,0.6296711929076422,0.05230670261082857,0.6269100169779287,0.049255407704104734
+flat_mae,patch,logistic,ppmi_dx,59,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,59,1291.5496650148827,test,0.63,0.04941750297212517,0.6009060511271707,0.05312936191366021,0.5997453310696095,0.052341003946846515
+flat_mae,patch,logistic,ppmi_dx,60,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,60,166.81005372000556,test,0.64,0.04657853582928515,0.6179966044142615,0.0492888142210014,0.6179966044142615,0.04952906671837194
+flat_mae,patch,logistic,ppmi_dx,61,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,61,1291.5496650148827,test,0.55,0.04880499564593772,0.529239460194581,0.04994525973428201,0.5301358234295416,0.05054336981795323
+flat_mae,patch,logistic,ppmi_dx,62,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,62,21.54434690031882,test,0.62,0.04615824520061394,0.5824175824175825,0.05139620385566298,0.5814940577249575,0.04899773105916247
+flat_mae,patch,logistic,ppmi_dx,63,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,63,2.782559402207126,test,0.54,0.04758266491065837,0.4875222816399287,0.050203450217702825,0.4915110356536503,0.04795426261259089
+flat_mae,patch,logistic,ppmi_dx,64,0.046415888336127774,train,0.8291814946619217,0.015007337693545696,0.81197724991636,0.017071033195863634,0.8021301648469279,0.01706420255301747
+flat_mae,patch,logistic,ppmi_dx,64,0.046415888336127774,test,0.65,0.04114918711226262,0.6011396011396011,0.04928851433577382,0.6005942275042444,0.04532489700476101
+flat_mae,patch,logistic,ppmi_dx,65,0.005994842503189409,train,0.7259786476868327,0.016395105467462496,0.6805167958656331,0.02069187496073376,0.6748287304645686,0.01872036835639237
+flat_mae,patch,logistic,ppmi_dx,65,0.005994842503189409,test,0.62,0.04081115043710481,0.5476190476190476,0.050173414267953485,0.5560271646859083,0.04344546078792213
+flat_mae,patch,logistic,ppmi_dx,66,0.046415888336127774,train,0.8256227758007118,0.015746880183346443,0.8061796171171172,0.0182500260315042,0.7948913508884607,0.018084061632786617
+flat_mae,patch,logistic,ppmi_dx,66,0.046415888336127774,test,0.59,0.04784696855601199,0.5577607593571352,0.050637052283693765,0.5573005093378608,0.04949236143908256
+flat_mae,patch,logistic,ppmi_dx,67,0.005994842503189409,train,0.7330960854092526,0.01735939412771078,0.6916695926966292,0.021457017711110936,0.6849577178334404,0.019705178738312627
+flat_mae,patch,logistic,ppmi_dx,67,0.005994842503189409,test,0.58,0.04517122535420087,0.5174632352941176,0.0514281833161737,0.5237691001697793,0.047187049630870274
+flat_mae,patch,logistic,ppmi_dx,68,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,68,166.81005372000556,test,0.6,0.050080878586542386,0.5796553173602353,0.05239001576145548,0.5806451612903225,0.052700890201385436
+flat_mae,patch,logistic,ppmi_dx,69,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,69,166.81005372000556,test,0.62,0.045947552709584005,0.5924495924495925,0.049063677870131786,0.5916808149405772,0.04844953604004136
+flat_mae,patch,logistic,ppmi_dx,70,0.046415888336127774,train,0.8131672597864769,0.015349466918237895,0.7920740795551844,0.017912050238747527,0.7812968315135945,0.017730720776784985
+flat_mae,patch,logistic,ppmi_dx,70,0.046415888336127774,test,0.61,0.04507843830480378,0.5741893219783819,0.04886585614819813,0.5734295415959253,0.04742353649505929
+flat_mae,patch,logistic,ppmi_dx,71,0.3593813663804626,train,0.9483985765124555,0.009428055789217106,0.9448264188628785,0.010208022997151378,0.9398281952472705,0.011105918156266732
+flat_mae,patch,logistic,ppmi_dx,71,0.3593813663804626,test,0.55,0.05028534180056848,0.508679986898133,0.05295729513822671,0.5097623089983022,0.0514000588722376
+flat_mae,patch,logistic,ppmi_dx,72,0.005994842503189409,train,0.7259786476868327,0.016507790174170568,0.6774448419797257,0.021386057850709426,0.6722195461357311,0.019018355332042283
+flat_mae,patch,logistic,ppmi_dx,72,0.005994842503189409,test,0.64,0.04349604119917122,0.5863970588235294,0.05053869992910319,0.5874363327674024,0.04623764003795999
+flat_mae,patch,logistic,ppmi_dx,73,0.3593813663804626,train,0.9448398576512456,0.009739906014802648,0.9411345337086054,0.010495922949425272,0.9369380218368657,0.011181722576772484
+flat_mae,patch,logistic,ppmi_dx,73,0.3593813663804626,test,0.65,0.04933728407604131,0.6266666666666667,0.05219130204490227,0.6260611205432938,0.0517882666350904
+flat_mae,patch,logistic,ppmi_dx,74,0.046415888336127774,train,0.8309608540925267,0.015523107197208135,0.8141553798867319,0.017666816572959935,0.8044449796617427,0.01777670055579514
+flat_mae,patch,logistic,ppmi_dx,74,0.046415888336127774,test,0.6,0.04600433023096847,0.5604395604395604,0.05104965650769923,0.5602716468590832,0.048815325017878046
+flat_mae,patch,logistic,ppmi_dx,75,0.005994842503189409,train,0.7366548042704626,0.015879845761197588,0.6919954970968124,0.020145976257193676,0.6852387069150075,0.018144734489034656
+flat_mae,patch,logistic,ppmi_dx,75,0.005994842503189409,test,0.65,0.04152059729820851,0.5944849959448499,0.05086677452937275,0.5955008488964346,0.045903108065366724
+flat_mae,patch,logistic,ppmi_dx,76,0.046415888336127774,train,0.8149466192170819,0.015132922533685676,0.7948242694862182,0.0176911138879785,0.7844813744380219,0.01770689348989579
+flat_mae,patch,logistic,ppmi_dx,76,0.046415888336127774,test,0.72,0.04277492723547289,0.6880570409982174,0.048925432795284066,0.6825127334465195,0.04679678245081813
+flat_mae,patch,logistic,ppmi_dx,77,0.3593813663804626,train,0.9430604982206405,0.009219411211725674,0.9391774891774892,0.009981732974608904,0.9346232070220509,0.010872791325420444
+flat_mae,patch,logistic,ppmi_dx,77,0.3593813663804626,test,0.66,0.04486218452104177,0.6392190152801358,0.047773143641127815,0.6392190152801358,0.048090744782581796
+flat_mae,patch,logistic,ppmi_dx,78,0.3593813663804626,train,0.9323843416370107,0.009794739960301805,0.9277732683982685,0.010609334940927768,0.9233435024619996,0.011477828838889863
+flat_mae,patch,logistic,ppmi_dx,78,0.3593813663804626,test,0.67,0.043003492881392776,0.6296711929076422,0.049938020032827345,0.6269100169779287,0.04717843959283201
+flat_mae,patch,logistic,ppmi_dx,79,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,79,1291.5496650148827,test,0.59,0.04611767557021928,0.5327635327635327,0.05151628612808517,0.5369269949066213,0.047372314832263854
+flat_mae,patch,logistic,ppmi_dx,80,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,80,2.782559402207126,test,0.6,0.04839406575190806,0.5833333333333333,0.05008338213612533,0.5857385398981324,0.0510699986495397
+flat_mae,patch,logistic,ppmi_dx,81,0.046415888336127774,train,0.8113879003558719,0.015302604899022383,0.7898279730740463,0.017953102607484804,0.7789820166987798,0.017788057196535957
+flat_mae,patch,logistic,ppmi_dx,81,0.046415888336127774,test,0.61,0.045553006486948815,0.568536342515765,0.05003759087847805,0.5683361629881154,0.04775172983690121
+flat_mae,patch,logistic,ppmi_dx,82,0.005994842503189409,train,0.7491103202846975,0.01504217674465068,0.7022264810326363,0.020247105123115918,0.6944845857418112,0.017952033297989802
+flat_mae,patch,logistic,ppmi_dx,82,0.005994842503189409,test,0.63,0.04292476674368773,0.5713127099988413,0.05181808230635205,0.5742784380305602,0.04638186181682066
+flat_mae,patch,logistic,ppmi_dx,83,0.046415888336127774,train,0.8362989323843416,0.014524287054841557,0.8193809823237617,0.016667742216380003,0.8087802397773496,0.016753263766858633
+flat_mae,patch,logistic,ppmi_dx,83,0.046415888336127774,test,0.55,0.04994769664358908,0.5248653785239151,0.05105099075687778,0.5250424448217317,0.05112423597965946
+flat_mae,patch,logistic,ppmi_dx,84,0.005994842503189409,train,0.7366548042704626,0.01601960329155893,0.6919954970968124,0.02024559707001699,0.6852387069150075,0.01826543491447043
+flat_mae,patch,logistic,ppmi_dx,84,0.005994842503189409,test,0.59,0.046030007603736066,0.539894512400404,0.05110378052434421,0.5420203735144312,0.048238238425047854
+flat_mae,patch,logistic,ppmi_dx,85,0.3593813663804626,train,0.9430604982206405,0.010056545969789624,0.9394070080862533,0.010799632583245643,0.936362663241276,0.011550906933002786
+flat_mae,patch,logistic,ppmi_dx,85,0.3593813663804626,test,0.63,0.04557706001926847,0.6053333333333333,0.047688024102301815,0.6048387096774194,0.04747682142158888
+flat_mae,patch,logistic,ppmi_dx,86,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,86,166.81005372000556,test,0.66,0.04545327270945404,0.6310763888888888,0.04995575048281524,0.6290322580645161,0.04889331383881605
+flat_mae,patch,logistic,ppmi_dx,87,0.046415888336127774,train,0.806049822064057,0.016166668353087877,0.7846932499165247,0.01875440130196279,0.7746467565831727,0.018513828033348046
+flat_mae,patch,logistic,ppmi_dx,87,0.046415888336127774,test,0.6,0.04711536479748406,0.554367201426025,0.053948927097137535,0.5551782682512734,0.05062286819955253
+flat_mae,patch,logistic,ppmi_dx,88,0.046415888336127774,train,0.8238434163701067,0.015084137328542031,0.8054037038980117,0.01743577167430011,0.7951857204024835,0.017548319111194425
+flat_mae,patch,logistic,ppmi_dx,88,0.046415888336127774,test,0.64,0.04127117638255542,0.5989304812834224,0.0478327702061941,0.597623089983022,0.04504135057837615
+flat_mae,patch,logistic,ppmi_dx,89,0.005994842503189409,train,0.7241992882562278,0.015496688395396675,0.6748207239727791,0.020257240480142307,0.6699047313209163,0.017929756261927488
+flat_mae,patch,logistic,ppmi_dx,89,0.005994842503189409,test,0.75,0.037104802923610854,0.6932891669733775,0.05197200240411804,0.6863327674023769,0.044350578393458434
+flat_mae,patch,logistic,ppmi_dx,90,0.005994842503189409,train,0.7277580071174378,0.016030231235737286,0.6878265194613769,0.019238778834035037,0.681492185827446,0.017768453624771693
+flat_mae,patch,logistic,ppmi_dx,90,0.005994842503189409,test,0.64,0.03828216817266233,0.5714285714285714,0.04706986841693695,0.5772495755517827,0.041056132083054
+flat_mae,patch,logistic,ppmi_dx,91,0.3593813663804626,train,0.9501779359430605,0.009247380777305858,0.946780303030303,0.00996449292376542,0.9421430100620852,0.01059746223337036
+flat_mae,patch,logistic,ppmi_dx,91,0.3593813663804626,test,0.63,0.04548587033354424,0.6009060511271707,0.04959468253259841,0.5997453310696095,0.04855752883879786
+flat_mae,patch,logistic,ppmi_dx,92,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,92,1291.5496650148827,test,0.64,0.050904318087957916,0.625,0.0519588195021414,0.6281833616298811,0.0527024302942833
+flat_mae,patch,logistic,ppmi_dx,93,0.005994842503189409,train,0.7259786476868327,0.017463695886141113,0.6805167958656331,0.02157354980770396,0.6748287304645686,0.019528881839355902
+flat_mae,patch,logistic,ppmi_dx,93,0.005994842503189409,test,0.65,0.04399750447468584,0.6266666666666667,0.047746458364485066,0.6260611205432938,0.047789687770116024
+flat_mae,patch,logistic,ppmi_dx,94,0.005994842503189409,train,0.7277580071174378,0.01674493474167773,0.6840643290969015,0.020902591055221804,0.678013273388996,0.018934096882191234
+flat_mae,patch,logistic,ppmi_dx,94,0.005994842503189409,test,0.66,0.041025241010870364,0.609375,0.04947568732048449,0.6086587436332768,0.04501942277367941
+flat_mae,patch,logistic,ppmi_dx,95,0.005994842503189409,train,0.7455516014234875,0.016647616844962645,0.7037995304489483,0.021285308553379123,0.695943052879469,0.019343628963898237
+flat_mae,patch,logistic,ppmi_dx,95,0.005994842503189409,test,0.64,0.0427258422971391,0.5792426367461431,0.05195978680227832,0.5823429541595926,0.045979232158601016
+flat_mae,patch,logistic,ppmi_dx,96,0.046415888336127774,train,0.8327402135231317,0.015259119202492686,0.8145527051125434,0.017715908112977036,0.8032808820381074,0.017719190035597496
+flat_mae,patch,logistic,ppmi_dx,96,0.046415888336127774,test,0.6,0.042422164018352485,0.5404411764705883,0.049152077986878207,0.5449915110356536,0.04485641089731464
+flat_mae,patch,logistic,ppmi_dx,97,0.3593813663804626,train,0.9430604982206405,0.009353604264759834,0.9391774891774892,0.010107929759359466,0.9346232070220509,0.01087469109081418
+flat_mae,patch,logistic,ppmi_dx,97,0.3593813663804626,test,0.64,0.04777656329205775,0.6216897856242118,0.04971825246175989,0.6230899830220713,0.049994696753527385
+flat_mae,patch,logistic,ppmi_dx,98,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,ppmi_dx,98,21.54434690031882,test,0.65,0.04750210100616604,0.6368917937545389,0.04887577630474463,0.6413412563667232,0.04991168465312845
+flat_mae,patch,logistic,ppmi_dx,99,0.3593813663804626,train,0.9483985765124555,0.008905091507312692,0.944932305727405,0.00962221421759671,0.9406979233568828,0.010517686544863595
+flat_mae,patch,logistic,ppmi_dx,99,0.3593813663804626,test,0.62,0.04707819452782785,0.6041666666666667,0.049364416466533834,0.6069609507640068,0.05029149910994365
+flat_mae,patch,logistic,ppmi_dx,100,0.046415888336127774,train,0.8113879003558719,0.015734568284522944,0.7918954796338993,0.018011918232139868,0.7824609291372298,0.017942483734992538
+flat_mae,patch,logistic,ppmi_dx,100,0.046415888336127774,test,0.63,0.04481706817720231,0.5960257670051315,0.05032183049997936,0.5946519524617997,0.048798049363389454

data_scaling/n400_1/eval_v2/ppmi_dx__patch__logistic/log.txt

@@ -0,0 +1,247 @@
+fMRI foundation model logistic probe eval
+version: 0.1.dev66+g7ddd3aa04
+sha: 58906bf7243fb545e1349221e6921a1797e2e666, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-02-26 17:14:46
+config:
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_1; eval v2 (ppmi_dx patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_1/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_1/eval_v2/ppmi_dx__patch__logistic
+model: flat_mae
+representation: patch
+dataset: ppmi_dx
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/eval_v2/ppmi_dx__patch__logistic
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=True, reg_tokens=0, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: ppmi_dx (flat)
+train (n=463):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'dir', 'sex', 'age', 'age_bin', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 463
+}),
+    labels=['PD' 'Prodromal'],
+    counts=[178 285]
+)
+
+validation (n=99):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'dir', 'sex', 'age', 'age_bin', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 99
+}),
+    labels=['PD' 'Prodromal'],
+    counts=[39 60]
+)
+
+test (n=100):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'ses', 'dir', 'sex', 'age', 'age_bin', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 100
+}),
+    labels=['PD' 'Prodromal'],
+    counts=[37 63]
+)
+
+extracting features for all splits
+extract (train)  [  0/232]  eta: 0:15:51    time: 4.0994  data: 3.0545  max mem: 2698
+extract (train)  [ 20/232]  eta: 0:01:24    time: 0.2139  data: 0.0827  max mem: 2851
+extract (train)  [ 40/232]  eta: 0:00:55    time: 0.1736  data: 0.0540  max mem: 2851
+extract (train)  [ 60/232]  eta: 0:00:42    time: 0.1674  data: 0.0499  max mem: 2851
+extract (train)  [ 80/232]  eta: 0:00:35    time: 0.1894  data: 0.0584  max mem: 2851
+extract (train)  [100/232]  eta: 0:00:29    time: 0.1701  data: 0.0487  max mem: 2851
+extract (train)  [120/232]  eta: 0:00:24    time: 0.1926  data: 0.0586  max mem: 2851
+extract (train)  [140/232]  eta: 0:00:19    time: 0.1973  data: 0.0626  max mem: 2851
+extract (train)  [160/232]  eta: 0:00:15    time: 0.1737  data: 0.0501  max mem: 2851
+extract (train)  [180/232]  eta: 0:00:10    time: 0.1864  data: 0.0559  max mem: 2851
+extract (train)  [200/232]  eta: 0:00:06    time: 0.1712  data: 0.0509  max mem: 2851
+extract (train)  [220/232]  eta: 0:00:02    time: 0.1419  data: 0.0389  max mem: 2851
+extract (train)  [231/232]  eta: 0:00:00    time: 0.1461  data: 0.0434  max mem: 2851
+extract (train) Total time: 0:00:45 (0.1973 s / it)
+extract (validation)  [ 0/50]  eta: 0:03:13    time: 3.8609  data: 3.7234  max mem: 2851
+extract (validation)  [20/50]  eta: 0:00:11    time: 0.2089  data: 0.0684  max mem: 2851
+extract (validation)  [40/50]  eta: 0:00:02    time: 0.1416  data: 0.0398  max mem: 2851
+extract (validation)  [49/50]  eta: 0:00:00    time: 0.1398  data: 0.0424  max mem: 2851
+extract (validation) Total time: 0:00:12 (0.2508 s / it)
+extract (test)  [ 0/50]  eta: 0:03:25    time: 4.1131  data: 3.9726  max mem: 2851
+extract (test)  [20/50]  eta: 0:00:12    time: 0.2182  data: 0.0635  max mem: 2851
+extract (test)  [40/50]  eta: 0:00:02    time: 0.1397  data: 0.0359  max mem: 2851
+extract (test)  [49/50]  eta: 0:00:00    time: 0.1391  data: 0.0369  max mem: 2851
+extract (test) Total time: 0:00:12 (0.2568 s / it)
+feature extraction time: 0:01:11
+train features: (463, 768)
+validation features: (99, 768)
+test features: (100, 768)
+evaluating fixed splits
+eval results (fixed splits):
+
+| model    | repr   | clf      | dataset   | trial   |       C | split   |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:----------|:--------|--------:|:--------|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | ppmi_dx   |         | 0.35938 | train   | 0.94306 | 0.0097188 | 0.93941 | 0.010426 | 0.93567 |   0.011095 |
+| flat_mae | patch  | logistic | ppmi_dx   |         | 0.35938 | test    | 0.63    | 0.04525   | 0.59066 | 0.048446 | 0.58923 |   0.046842 |
+
+
+evaluating random splits (n=100)
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 1, "C": 0.046415888336127774, "split": "test", "acc": 0.62, "acc_std": 0.04371370036956378, "f1": 0.5766488413547237, "f1_std": 0.04965517929109973, "bacc": 0.5764006791171477, "bacc_std": 0.046718817119045165}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 2, "C": 0.3593813663804626, "split": "test", "acc": 0.59, "acc_std": 0.04961217189359885, "f1": 0.5626666666666666, "f1_std": 0.05281989014396253, "bacc": 0.5623938879456706, "bacc_std": 0.05229097998959533}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 3, "C": 0.046415888336127774, "split": "test", "acc": 0.69, "acc_std": 0.04351703574463683, "f1": 0.6615351020853806, "f1_std": 0.048463542592696605, "bacc": 0.6583191850594228, "bacc_std": 0.047252174901520914}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 4, "C": 0.005994842503189409, "split": "test", "acc": 0.7, "acc_std": 0.04106226978626486, "f1": 0.66078697421981, "f1_std": 0.047734870379201884, "bacc": 0.6561969439728353, "bacc_std": 0.044782648390475896}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 5, "C": 0.046415888336127774, "split": "test", "acc": 0.56, "acc_std": 0.050015721528335465, "f1": 0.5280995280995281, "f1_std": 0.05238886009304962, "bacc": 0.5280135823429541, "bacc_std": 0.05173854732883097}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 6, "C": 0.046415888336127774, "split": "test", "acc": 0.58, "acc_std": 0.04470615170197497, "f1": 0.525101763907734, "f1_std": 0.051937224805165036, "bacc": 0.5288624787775891, "bacc_std": 0.04761150502647497}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 7, "C": 0.046415888336127774, "split": "test", "acc": 0.55, "acc_std": 0.051013896146050244, "f1": 0.5366079703429101, "f1_std": 0.05184591645257948, "bacc": 0.5403225806451613, "bacc_std": 0.05342905332313932}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 8, "C": 0.005994842503189409, "split": "test", "acc": 0.62, "acc_std": 0.038295294750138695, "f1": 0.5386109762020399, "f1_std": 0.048785906087796614, "bacc": 0.5509337860780985, "bacc_std": 0.04092566604131759}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 9, "C": 0.005994842503189409, "split": "test", "acc": 0.71, "acc_std": 0.03829867360627519, "f1": 0.6579785352046232, "f1_std": 0.04956219876842803, "bacc": 0.6540747028862479, "bacc_std": 0.0439527155611839}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 10, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.04425717568937268, "f1": 0.5604395604395604, "f1_std": 0.04849572941282711, "bacc": 0.5602716468590832, "bacc_std": 0.04660202522938627}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 11, "C": 2.782559402207126, "split": "test", "acc": 0.7, "acc_std": 0.044095732219796506, "f1": 0.6782496782496783, "f1_std": 0.047327395321723906, "bacc": 0.6765704584040747, "bacc_std": 0.04682570102858115}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 12, "C": 0.3593813663804626, "split": "test", "acc": 0.65, "acc_std": 0.044211034821637005, "f1": 0.6072270227808326, "f1_std": 0.050228756052448094, "bacc": 0.6056876061120543, "bacc_std": 0.04736082615996879}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 13, "C": 0.3593813663804626, "split": "test", "acc": 0.61, "acc_std": 0.04650550074991129, "f1": 0.568536342515765, "f1_std": 0.05122102941892769, "bacc": 0.5683361629881154, "bacc_std": 0.04890203025188589}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 14, "C": 0.046415888336127774, "split": "test", "acc": 0.57, "acc_std": 0.04531796994570696, "f1": 0.5305164319248826, "f1_std": 0.04858880543837976, "bacc": 0.5309847198641766, "bacc_std": 0.0469430625270563}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 15, "C": 0.005994842503189409, "split": "test", "acc": 0.66, "acc_std": 0.04194617503420306, "f1": 0.609375, "f1_std": 0.0507321062318044, "bacc": 0.6086587436332768, "bacc_std": 0.04615700467813121}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 16, "C": 0.046415888336127774, "split": "test", "acc": 0.65, "acc_std": 0.045189817437117405, "f1": 0.6224786970121885, "f1_std": 0.0490666338031973, "bacc": 0.6209677419354839, "bacc_std": 0.04831710061727428}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 17, "C": 0.046415888336127774, "split": "test", "acc": 0.59, "acc_std": 0.04571877513669849, "f1": 0.5464100011063171, "f1_std": 0.050084490057855875, "bacc": 0.5471137521222411, "bacc_std": 0.047829283307919995}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 18, "C": 1291.5496650148827, "split": "test", "acc": 0.57, "acc_std": 0.049609978834907795, "f1": 0.557203171661003, "f1_std": 0.05066544682878853, "bacc": 0.5615449915110357, "bacc_std": 0.05224853446061583}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 19, "C": 0.046415888336127774, "split": "test", "acc": 0.71, "acc_std": 0.044014888390179974, "f1": 0.6938021328265231, "f1_std": 0.04620432585274497, "bacc": 0.6948217317487266, "bacc_std": 0.046390377377879745}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 20, "C": 0.046415888336127774, "split": "test", "acc": 0.67, "acc_std": 0.041664848493664294, "f1": 0.6176572818908586, "f1_std": 0.050292728202088126, "bacc": 0.616723259762309, "bacc_std": 0.04547365133574492}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 21, "C": 1291.5496650148827, "split": "test", "acc": 0.61, "acc_std": 0.04743608752837865, "f1": 0.5882166613873931, "f1_std": 0.049491386541767775, "bacc": 0.5887096774193548, "bacc_std": 0.049504562984715186}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 22, "C": 21.54434690031882, "split": "test", "acc": 0.55, "acc_std": 0.04869856260712424, "f1": 0.5248653785239151, "f1_std": 0.050756742363550975, "bacc": 0.5250424448217317, "bacc_std": 0.051113161301739526}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 23, "C": 0.005994842503189409, "split": "test", "acc": 0.66, "acc_std": 0.04496213518061614, "f1": 0.609375, "f1_std": 0.054040822047925936, "bacc": 0.6086587436332768, "bacc_std": 0.04955820404901081}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 24, "C": 0.005994842503189409, "split": "test", "acc": 0.63, "acc_std": 0.04088236294540716, "f1": 0.5636277862955537, "f1_std": 0.050203991210757234, "bacc": 0.5691850594227504, "bacc_std": 0.04412904201521217}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 25, "C": 0.3593813663804626, "split": "test", "acc": 0.64, "acc_std": 0.03998951862676019, "f1": 0.5792426367461431, "f1_std": 0.05036360647994039, "bacc": 0.5823429541595926, "bacc_std": 0.044420179846896654}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 26, "C": 0.3593813663804626, "split": "test", "acc": 0.64, "acc_std": 0.04565597879796249, "f1": 0.6179966044142615, "f1_std": 0.048775255708017544, "bacc": 0.6179966044142615, "bacc_std": 0.04896193072503621}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 27, "C": 0.005994842503189409, "split": "test", "acc": 0.62, "acc_std": 0.041886183879651775, "f1": 0.5476190476190476, "f1_std": 0.052077775167443024, "bacc": 0.5560271646859083, "bacc_std": 0.04543871301389427}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 28, "C": 0.3593813663804626, "split": "test", "acc": 0.62, "acc_std": 0.04606073816169255, "f1": 0.5924495924495925, "f1_std": 0.04968790442111471, "bacc": 0.5916808149405772, "bacc_std": 0.04891071030375556}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 29, "C": 0.046415888336127774, "split": "test", "acc": 0.63, "acc_std": 0.047497974693664576, "f1": 0.5906626839252129, "f1_std": 0.05315077608236612, "bacc": 0.5895585738539898, "bacc_std": 0.0506647830186329}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 30, "C": 21.54434690031882, "split": "test", "acc": 0.61, "acc_std": 0.04777488461524528, "f1": 0.584, "f1_std": 0.051627396178459475, "bacc": 0.583616298811545, "bacc_std": 0.05144063503619579}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 31, "C": 0.046415888336127774, "split": "test", "acc": 0.66, "acc_std": 0.04606840131804012, "f1": 0.6353496353496353, "f1_std": 0.04849142693462424, "bacc": 0.634125636672326, "bacc_std": 0.04791073543150732}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 32, "C": 0.046415888336127774, "split": "test", "acc": 0.67, "acc_std": 0.04428216345211692, "f1": 0.6440513428972063, "f1_std": 0.048367590788555063, "bacc": 0.6421901528013583, "bacc_std": 0.04776257510752226}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 33, "C": 0.3593813663804626, "split": "test", "acc": 0.66, "acc_std": 0.04723928449923855, "f1": 0.6353496353496353, "f1_std": 0.05120044521088683, "bacc": 0.634125636672326, "bacc_std": 0.05076863277495403}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 34, "C": 0.046415888336127774, "split": "test", "acc": 0.67, "acc_std": 0.044037125246773313, "f1": 0.6296711929076422, "f1_std": 0.0515727101953694, "bacc": 0.6269100169779287, "bacc_std": 0.04865137855127997}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 35, "C": 0.3593813663804626, "split": "test", "acc": 0.6, "acc_std": 0.049603528100327705, "f1": 0.5659722222222222, "f1_std": 0.053085290658552796, "bacc": 0.565365025466893, "bacc_std": 0.05195229570282048}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 36, "C": 0.005994842503189409, "split": "test", "acc": 0.64, "acc_std": 0.04114984811636612, "f1": 0.5792426367461431, "f1_std": 0.04936410027607919, "bacc": 0.5823429541595926, "bacc_std": 0.04417171930517304}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 37, "C": 0.3593813663804626, "split": "test", "acc": 0.65, "acc_std": 0.04500335987456937, "f1": 0.630450849963045, "f1_std": 0.04682204692793159, "bacc": 0.6311544991511036, "bacc_std": 0.04707293583689922}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 38, "C": 0.046415888336127774, "split": "test", "acc": 0.7, "acc_std": 0.04530611879205721, "f1": 0.6847414880201765, "f1_std": 0.04735395224066736, "bacc": 0.6867572156196944, "bacc_std": 0.04779090071698159}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 39, "C": 0.046415888336127774, "split": "test", "acc": 0.67, "acc_std": 0.0459393121411281, "f1": 0.6396986570586308, "f1_std": 0.05064134614606536, "bacc": 0.6370967741935484, "bacc_std": 0.04919403747059671}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 40, "C": 0.000774263682681127, "split": "test", "acc": 0.66, "acc_std": 0.03585707182690745, "f1": 0.5687468290208015, "f1_std": 0.052414192804968146, "bacc": 0.5831918505942275, "bacc_std": 0.04086209520637194}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 41, "C": 2.782559402207126, "split": "test", "acc": 0.6, "acc_std": 0.05132166793860075, "f1": 0.586606035551881, "f1_std": 0.05222995387608565, "bacc": 0.5908319185059423, "bacc_std": 0.05319657754243219}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 42, "C": 0.046415888336127774, "split": "test", "acc": 0.68, "acc_std": 0.039683119837028935, "f1": 0.6259934548854604, "f1_std": 0.04975199597504042, "bacc": 0.6247877758913413, "bacc_std": 0.04432757335355691}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 43, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.04425144065451429, "f1": 0.554367201426025, "f1_std": 0.049556954063591244, "bacc": 0.5551782682512734, "bacc_std": 0.04696458801086889}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 44, "C": 0.005994842503189409, "split": "test", "acc": 0.68, "acc_std": 0.04148917449166709, "f1": 0.6259934548854604, "f1_std": 0.05256835054930224, "bacc": 0.6247877758913413, "bacc_std": 0.046974610754166264}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 45, "C": 0.046415888336127774, "split": "test", "acc": 0.66, "acc_std": 0.044687465804182716, "f1": 0.6212121212121212, "f1_std": 0.050608278902441156, "bacc": 0.6188455008488964, "bacc_std": 0.04811151937006594}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 46, "C": 0.046415888336127774, "split": "test", "acc": 0.63, "acc_std": 0.0459838058451016, "f1": 0.5783475783475784, "f1_std": 0.0538732946615018, "bacc": 0.5793718166383701, "bacc_std": 0.04953516975815706}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 47, "C": 0.046415888336127774, "split": "test", "acc": 0.7, "acc_std": 0.04302354704112621, "f1": 0.6744791666666667, "f1_std": 0.04795109699290812, "bacc": 0.6714770797962648, "bacc_std": 0.047428645864378505}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 48, "C": 0.005994842503189409, "split": "test", "acc": 0.64, "acc_std": 0.042187291925412804, "f1": 0.5989304812834224, "f1_std": 0.048709334146665534, "bacc": 0.597623089983022, "bacc_std": 0.04596119541178976}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 49, "C": 2.782559402207126, "split": "test", "acc": 0.65, "acc_std": 0.04786762998102162, "f1": 0.6224786970121885, "f1_std": 0.05126797979686403, "bacc": 0.6209677419354839, "bacc_std": 0.050436386082368084}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 50, "C": 0.3593813663804626, "split": "test", "acc": 0.66, "acc_std": 0.04265457068122946, "f1": 0.6155585707824514, "f1_std": 0.04908085042653531, "bacc": 0.6137521222410866, "bacc_std": 0.04579370911799187}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 51, "C": 0.3593813663804626, "split": "test", "acc": 0.58, "acc_std": 0.04964866564168668, "f1": 0.5442708333333334, "f1_std": 0.053437358865708384, "bacc": 0.5441426146010186, "bacc_std": 0.05207740706125955}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 52, "C": 0.046415888336127774, "split": "test", "acc": 0.62, "acc_std": 0.046995046547482, "f1": 0.5824175824175825, "f1_std": 0.051826829456541615, "bacc": 0.5814940577249575, "bacc_std": 0.04979921677443064}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 53, "C": 2.782559402207126, "split": "test", "acc": 0.55, "acc_std": 0.04745463517929517, "f1": 0.5396419437340154, "f1_std": 0.04775294687403358, "bacc": 0.5454159592529711, "bacc_std": 0.049309415309893}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 54, "C": 0.005994842503189409, "split": "test", "acc": 0.55, "acc_std": 0.04605473265583027, "f1": 0.5021573182874212, "f1_std": 0.04786812136845858, "bacc": 0.5046689303904923, "bacc_std": 0.04612370514877537}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 55, "C": 21.54434690031882, "split": "test", "acc": 0.6, "acc_std": 0.04482849094047222, "f1": 0.5833333333333333, "f1_std": 0.046418273462788096, "bacc": 0.5857385398981324, "bacc_std": 0.04725411490121178}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 56, "C": 0.046415888336127774, "split": "test", "acc": 0.64, "acc_std": 0.046751337948768915, "f1": 0.592944369063772, "f1_std": 0.05434622524608595, "bacc": 0.5925297113752122, "bacc_std": 0.05029172201083732}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 57, "C": 0.046415888336127774, "split": "test", "acc": 0.7, "acc_std": 0.040116829386181545, "f1": 0.6493688639551192, "f1_std": 0.0507632949920872, "bacc": 0.6460101867572157, "bacc_std": 0.04553371611500591}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 58, "C": 0.046415888336127774, "split": "test", "acc": 0.67, "acc_std": 0.04486856806273184, "f1": 0.6296711929076422, "f1_std": 0.05230670261082857, "bacc": 0.6269100169779287, "bacc_std": 0.049255407704104734}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 59, "C": 1291.5496650148827, "split": "test", "acc": 0.63, "acc_std": 0.04941750297212517, "f1": 0.6009060511271707, "f1_std": 0.05312936191366021, "bacc": 0.5997453310696095, "bacc_std": 0.052341003946846515}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 60, "C": 166.81005372000556, "split": "test", "acc": 0.64, "acc_std": 0.04657853582928515, "f1": 0.6179966044142615, "f1_std": 0.0492888142210014, "bacc": 0.6179966044142615, "bacc_std": 0.04952906671837194}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 61, "C": 1291.5496650148827, "split": "test", "acc": 0.55, "acc_std": 0.04880499564593772, "f1": 0.529239460194581, "f1_std": 0.04994525973428201, "bacc": 0.5301358234295416, "bacc_std": 0.05054336981795323}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 62, "C": 21.54434690031882, "split": "test", "acc": 0.62, "acc_std": 0.04615824520061394, "f1": 0.5824175824175825, "f1_std": 0.05139620385566298, "bacc": 0.5814940577249575, "bacc_std": 0.04899773105916247}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 63, "C": 2.782559402207126, "split": "test", "acc": 0.54, "acc_std": 0.04758266491065837, "f1": 0.4875222816399287, "f1_std": 0.050203450217702825, "bacc": 0.4915110356536503, "bacc_std": 0.04795426261259089}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 64, "C": 0.046415888336127774, "split": "test", "acc": 0.65, "acc_std": 0.04114918711226262, "f1": 0.6011396011396011, "f1_std": 0.04928851433577382, "bacc": 0.6005942275042444, "bacc_std": 0.04532489700476101}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 65, "C": 0.005994842503189409, "split": "test", "acc": 0.62, "acc_std": 0.04081115043710481, "f1": 0.5476190476190476, "f1_std": 0.050173414267953485, "bacc": 0.5560271646859083, "bacc_std": 0.04344546078792213}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 66, "C": 0.046415888336127774, "split": "test", "acc": 0.59, "acc_std": 0.04784696855601199, "f1": 0.5577607593571352, "f1_std": 0.050637052283693765, "bacc": 0.5573005093378608, "bacc_std": 0.04949236143908256}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 67, "C": 0.005994842503189409, "split": "test", "acc": 0.58, "acc_std": 0.04517122535420087, "f1": 0.5174632352941176, "f1_std": 0.0514281833161737, "bacc": 0.5237691001697793, "bacc_std": 0.047187049630870274}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 68, "C": 166.81005372000556, "split": "test", "acc": 0.6, "acc_std": 0.050080878586542386, "f1": 0.5796553173602353, "f1_std": 0.05239001576145548, "bacc": 0.5806451612903225, "bacc_std": 0.052700890201385436}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 69, "C": 166.81005372000556, "split": "test", "acc": 0.62, "acc_std": 0.045947552709584005, "f1": 0.5924495924495925, "f1_std": 0.049063677870131786, "bacc": 0.5916808149405772, "bacc_std": 0.04844953604004136}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 70, "C": 0.046415888336127774, "split": "test", "acc": 0.61, "acc_std": 0.04507843830480378, "f1": 0.5741893219783819, "f1_std": 0.04886585614819813, "bacc": 0.5734295415959253, "bacc_std": 0.04742353649505929}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 71, "C": 0.3593813663804626, "split": "test", "acc": 0.55, "acc_std": 0.05028534180056848, "f1": 0.508679986898133, "f1_std": 0.05295729513822671, "bacc": 0.5097623089983022, "bacc_std": 0.0514000588722376}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 72, "C": 0.005994842503189409, "split": "test", "acc": 0.64, "acc_std": 0.04349604119917122, "f1": 0.5863970588235294, "f1_std": 0.05053869992910319, "bacc": 0.5874363327674024, "bacc_std": 0.04623764003795999}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 73, "C": 0.3593813663804626, "split": "test", "acc": 0.65, "acc_std": 0.04933728407604131, "f1": 0.6266666666666667, "f1_std": 0.05219130204490227, "bacc": 0.6260611205432938, "bacc_std": 0.0517882666350904}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 74, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.04600433023096847, "f1": 0.5604395604395604, "f1_std": 0.05104965650769923, "bacc": 0.5602716468590832, "bacc_std": 0.048815325017878046}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 75, "C": 0.005994842503189409, "split": "test", "acc": 0.65, "acc_std": 0.04152059729820851, "f1": 0.5944849959448499, "f1_std": 0.05086677452937275, "bacc": 0.5955008488964346, "bacc_std": 0.045903108065366724}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 76, "C": 0.046415888336127774, "split": "test", "acc": 0.72, "acc_std": 0.04277492723547289, "f1": 0.6880570409982174, "f1_std": 0.048925432795284066, "bacc": 0.6825127334465195, "bacc_std": 0.04679678245081813}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 77, "C": 0.3593813663804626, "split": "test", "acc": 0.66, "acc_std": 0.04486218452104177, "f1": 0.6392190152801358, "f1_std": 0.047773143641127815, "bacc": 0.6392190152801358, "bacc_std": 0.048090744782581796}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 78, "C": 0.3593813663804626, "split": "test", "acc": 0.67, "acc_std": 0.043003492881392776, "f1": 0.6296711929076422, "f1_std": 0.049938020032827345, "bacc": 0.6269100169779287, "bacc_std": 0.04717843959283201}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 79, "C": 1291.5496650148827, "split": "test", "acc": 0.59, "acc_std": 0.04611767557021928, "f1": 0.5327635327635327, "f1_std": 0.05151628612808517, "bacc": 0.5369269949066213, "bacc_std": 0.047372314832263854}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 80, "C": 2.782559402207126, "split": "test", "acc": 0.6, "acc_std": 0.04839406575190806, "f1": 0.5833333333333333, "f1_std": 0.05008338213612533, "bacc": 0.5857385398981324, "bacc_std": 0.0510699986495397}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 81, "C": 0.046415888336127774, "split": "test", "acc": 0.61, "acc_std": 0.045553006486948815, "f1": 0.568536342515765, "f1_std": 0.05003759087847805, "bacc": 0.5683361629881154, "bacc_std": 0.04775172983690121}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 82, "C": 0.005994842503189409, "split": "test", "acc": 0.63, "acc_std": 0.04292476674368773, "f1": 0.5713127099988413, "f1_std": 0.05181808230635205, "bacc": 0.5742784380305602, "bacc_std": 0.04638186181682066}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 83, "C": 0.046415888336127774, "split": "test", "acc": 0.55, "acc_std": 0.04994769664358908, "f1": 0.5248653785239151, "f1_std": 0.05105099075687778, "bacc": 0.5250424448217317, "bacc_std": 0.05112423597965946}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 84, "C": 0.005994842503189409, "split": "test", "acc": 0.59, "acc_std": 0.046030007603736066, "f1": 0.539894512400404, "f1_std": 0.05110378052434421, "bacc": 0.5420203735144312, "bacc_std": 0.048238238425047854}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 85, "C": 0.3593813663804626, "split": "test", "acc": 0.63, "acc_std": 0.04557706001926847, "f1": 0.6053333333333333, "f1_std": 0.047688024102301815, "bacc": 0.6048387096774194, "bacc_std": 0.04747682142158888}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 86, "C": 166.81005372000556, "split": "test", "acc": 0.66, "acc_std": 0.04545327270945404, "f1": 0.6310763888888888, "f1_std": 0.04995575048281524, "bacc": 0.6290322580645161, "bacc_std": 0.04889331383881605}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 87, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.04711536479748406, "f1": 0.554367201426025, "f1_std": 0.053948927097137535, "bacc": 0.5551782682512734, "bacc_std": 0.05062286819955253}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 88, "C": 0.046415888336127774, "split": "test", "acc": 0.64, "acc_std": 0.04127117638255542, "f1": 0.5989304812834224, "f1_std": 0.0478327702061941, "bacc": 0.597623089983022, "bacc_std": 0.04504135057837615}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 89, "C": 0.005994842503189409, "split": "test", "acc": 0.75, "acc_std": 0.037104802923610854, "f1": 0.6932891669733775, "f1_std": 0.05197200240411804, "bacc": 0.6863327674023769, "bacc_std": 0.044350578393458434}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 90, "C": 0.005994842503189409, "split": "test", "acc": 0.64, "acc_std": 0.03828216817266233, "f1": 0.5714285714285714, "f1_std": 0.04706986841693695, "bacc": 0.5772495755517827, "bacc_std": 0.041056132083054}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 91, "C": 0.3593813663804626, "split": "test", "acc": 0.63, "acc_std": 0.04548587033354424, "f1": 0.6009060511271707, "f1_std": 0.04959468253259841, "bacc": 0.5997453310696095, "bacc_std": 0.04855752883879786}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 92, "C": 1291.5496650148827, "split": "test", "acc": 0.64, "acc_std": 0.050904318087957916, "f1": 0.625, "f1_std": 0.0519588195021414, "bacc": 0.6281833616298811, "bacc_std": 0.0527024302942833}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 93, "C": 0.005994842503189409, "split": "test", "acc": 0.65, "acc_std": 0.04399750447468584, "f1": 0.6266666666666667, "f1_std": 0.047746458364485066, "bacc": 0.6260611205432938, "bacc_std": 0.047789687770116024}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 94, "C": 0.005994842503189409, "split": "test", "acc": 0.66, "acc_std": 0.041025241010870364, "f1": 0.609375, "f1_std": 0.04947568732048449, "bacc": 0.6086587436332768, "bacc_std": 0.04501942277367941}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 95, "C": 0.005994842503189409, "split": "test", "acc": 0.64, "acc_std": 0.0427258422971391, "f1": 0.5792426367461431, "f1_std": 0.05195978680227832, "bacc": 0.5823429541595926, "bacc_std": 0.045979232158601016}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 96, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.042422164018352485, "f1": 0.5404411764705883, "f1_std": 0.049152077986878207, "bacc": 0.5449915110356536, "bacc_std": 0.04485641089731464}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 97, "C": 0.3593813663804626, "split": "test", "acc": 0.64, "acc_std": 0.04777656329205775, "f1": 0.6216897856242118, "f1_std": 0.04971825246175989, "bacc": 0.6230899830220713, "bacc_std": 0.049994696753527385}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 98, "C": 21.54434690031882, "split": "test", "acc": 0.65, "acc_std": 0.04750210100616604, "f1": 0.6368917937545389, "f1_std": 0.04887577630474463, "bacc": 0.6413412563667232, "bacc_std": 0.04991168465312845}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 99, "C": 0.3593813663804626, "split": "test", "acc": 0.62, "acc_std": 0.04707819452782785, "f1": 0.6041666666666667, "f1_std": 0.049364416466533834, "bacc": 0.6069609507640068, "bacc_std": 0.05029149910994365}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "ppmi_dx", "trial": 100, "C": 0.046415888336127774, "split": "test", "acc": 0.63, "acc_std": 0.04481706817720231, "f1": 0.5960257670051315, "f1_std": 0.05032183049997936, "bacc": 0.5946519524617997, "bacc_std": 0.048798049363389454}
+eval results (random splits):
+
+| model    | repr   | clf      | dataset   | split   |   n_trials |      C |   C_std |    acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:----------|:--------|-----------:|-------:|--------:|-------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | ppmi_dx   | train   |        100 | 85.496 |     308 | 0.8598 |  0.10187  | 0.84087 | 0.11921  | 0.83511 |   0.12136  |
+| flat_mae | patch  | logistic | ppmi_dx   | test    |        100 | 85.496 |     308 | 0.6304 |  0.043133 | 0.59217 | 0.044997 | 0.59278 |   0.043239 |
+
+
+done! total time: 0:05:18

data_scaling/n400_1/pretrain/config.yaml

@@ -0,0 +1,109 @@
+name: data_scaling/n400_1/pretrain
+notes: data scaling experiment n400_1 (seed=1644)
+output_dir: experiments/data_scaling/output/data_scaling/n400_1/pretrain
+input_space: flat
+patch_size: 16
+num_frames: 16
+t_patch_size: 4
+mask_ratio: 0.9
+pred_mask_ratio: null
+masking: tube
+masking_kwargs: {}
+mask_patch_size: null
+model: mae_vit_base
+model_kwargs:
+  decoding: attn
+  pos_embed: sep
+  target_norm: null
+  pca_norm_nc: 2
+  t_pred_stride: 2
+  no_decode_pos: true
+  mask_drop_scale: false
+  pred_edge_pad: 0
+  gauss_sigma: null
+  class_token: true
+  reg_tokens: 0
+  no_embed_class: true
+  head_init_scale: 0.0
+  decoder_depth: 4
+  drop_path_rate: 0.0
+datasets:
+  hcp-train:
+    type: wds
+    url: /data/fmri-datasets/pretrain/hcpya-all.flat.wds/hcpya-all-flat-{00000..00399}.tar
+    clipping: random
+    clipping_kwargs:
+      oversample: 4.0
+    shuffle: true
+    buffer_size: 2000
+    samples_per_epoch: 200000
+  hcp-train-subset:
+    type: arrow
+    root: s3://medarc/fmri-datasets/eval/hcpya-clips.${input_space}.arrow/train
+    split_range:
+    - 0
+    - 2000
+    shuffle: false
+  hcp-val:
+    type: arrow
+    root: s3://medarc/fmri-datasets/eval/hcpya-clips.${input_space}.arrow/test
+    split_range:
+    - 0
+    - 2000
+    shuffle: false
+  nsd-val:
+    type: arrow
+    root: s3://medarc/fmri-datasets/eval/nsd-cococlip.${input_space}.arrow/testid
+    split_range:
+    - 0
+    - 2000
+    shuffle: false
+train_dataset: hcp-train
+eval_datasets:
+- hcp-train-subset
+- hcp-val
+- nsd-val
+val_dataset: hcp-val
+clip_vmax: 3.0
+normalize: frame
+tr_scale: null
+crop_scale: null
+crop_aspect: null
+gray_jitter: null
+num_workers: 16
+epochs: 100
+batch_size: 32
+accum_iter: 1
+base_lr: 0.001
+min_lr: 0.0
+warmup_epochs: 5
+weight_decay: 0.05
+betas:
+- 0.9
+- 0.95
+clip_grad: 1.0
+amp: true
+amp_dtype: float16
+ckpt: null
+resume: true
+auto_resume: true
+start_epoch: 0
+max_checkpoints: 20
+checkpoint_period: 5
+plot_period: 5
+device: cuda
+presend_cuda: false
+seed: 1644
+debug: false
+wandb: true
+wandb_entity: null
+wandb_project: fMRI-foundation-model
+rank: 0
+world_size: 1
+gpu: 0
+distributed: true
+dist_backend: nccl
+in_chans: 1
+img_size:
+- 224
+- 560

data_scaling/n400_1/pretrain/log.json

@@ -0,0 +1,100 @@
+{"epoch": 0, "train/lr": 1.2502400076802458e-05, "train/grad": 0.05253171782687306, "train/loss": 0.9932034360313415, "eval/hcp-train-subset/loss": 0.9886764874381404, "eval/hcp-val/loss": 0.9893021083647205, "eval/nsd-val/loss": 0.9902563835344007}
+{"epoch": 1, "train/lr": 3.750320010240327e-05, "train/grad": 0.08323407599210739, "train/loss": 0.9877021530723572, "eval/hcp-train-subset/loss": 0.9861039163604859, "eval/hcp-val/loss": 0.9858533605452506, "eval/nsd-val/loss": 0.9875980336819926}
+{"epoch": 2, "train/lr": 6.250400012800409e-05, "train/grad": 0.11681490068622429, "train/loss": 0.9839487573242187, "eval/hcp-train-subset/loss": 0.9818917426370806, "eval/hcp-val/loss": 0.9817590540455233, "eval/nsd-val/loss": 0.9835789761235637}
+{"epoch": 3, "train/lr": 8.75048001536049e-05, "train/grad": 0.19776365545697885, "train/loss": 0.9759244383239746, "eval/hcp-train-subset/loss": 0.9704708755016327, "eval/hcp-val/loss": 0.9707422612174865, "eval/nsd-val/loss": 0.971840298944904}
+{"epoch": 4, "train/lr": 0.00011250559953918529, "train/grad": 0.22726607908768637, "train/loss": 0.9475549354171753, "eval/hcp-train-subset/loss": 0.9239714241796925, "eval/hcp-val/loss": 0.9228701360764042, "eval/nsd-val/loss": 0.8982545983406806}
+{"epoch": 5, "train/lr": 0.00012498860637884563, "train/grad": 0.17325427122271014, "train/loss": 0.9093724193954468, "eval/hcp-train-subset/loss": 0.8869803163313097, "eval/hcp-val/loss": 0.8862056520677382, "eval/nsd-val/loss": 0.8514787333626901}
+{"epoch": 6, "train/lr": 0.0001249202705377922, "train/grad": 0.12361626946985273, "train/loss": 0.8794224549674988, "eval/hcp-train-subset/loss": 0.8687657608139899, "eval/hcp-val/loss": 0.8677900933450268, "eval/nsd-val/loss": 0.8338712444228511}
+{"epoch": 7, "train/lr": 0.0001247836790473516, "train/grad": 0.09967410319369101, "train/loss": 0.8649758139324188, "eval/hcp-train-subset/loss": 0.8611551926982018, "eval/hcp-val/loss": 0.8604481047199618, "eval/nsd-val/loss": 0.8275549671342296}
+{"epoch": 8, "train/lr": 0.000124578981268311, "train/grad": 0.08736097741808264, "train/loss": 0.8568171174621582, "eval/hcp-train-subset/loss": 0.8557896940938888, "eval/hcp-val/loss": 0.8556686505194633, "eval/nsd-val/loss": 0.8230214224707696}
+{"epoch": 9, "train/lr": 0.00012430640103468907, "train/grad": 0.08096725973282749, "train/loss": 0.854563008184433, "eval/hcp-train-subset/loss": 0.8529351555532024, "eval/hcp-val/loss": 0.8534167053238038, "eval/nsd-val/loss": 0.8209515415853069}
+{"epoch": 10, "train/lr": 0.00012396623640896796, "train/grad": 0.07810898929743723, "train/loss": 0.8515216468048096, "eval/hcp-train-subset/loss": 0.8496764879072866, "eval/hcp-val/loss": 0.8512248368032517, "eval/nsd-val/loss": 0.8206694539516203}
+{"epoch": 11, "train/lr": 0.0001235588593561712, "train/grad": 0.07819226134995359, "train/loss": 0.8464853427028656, "eval/hcp-train-subset/loss": 0.8486170970624493, "eval/hcp-val/loss": 0.8492289829638696, "eval/nsd-val/loss": 0.8197939501654717}
+{"epoch": 12, "train/lr": 0.00012308471533712604, "train/grad": 0.07955919921030881, "train/loss": 0.8419188573646545, "eval/hcp-train-subset/loss": 0.8472811118248971, "eval/hcp-val/loss": 0.8489567214442838, "eval/nsd-val/loss": 0.8206922056213501}
+{"epoch": 13, "train/lr": 0.00012254432282135565, "train/grad": 0.07950216620130045, "train/loss": 0.8411275442504883, "eval/hcp-train-subset/loss": 0.8455295283948222, "eval/hcp-val/loss": 0.8478296629844173, "eval/nsd-val/loss": 0.8170165861806562}
+{"epoch": 14, "train/lr": 0.00012193827272014171, "train/grad": 0.07988865848358782, "train/loss": 0.8407622010993957, "eval/hcp-train-subset/loss": 0.8443546304779668, "eval/hcp-val/loss": 0.8470775475425105, "eval/nsd-val/loss": 0.8185480784985327}
+{"epoch": 15, "train/lr": 0.00012126722774037197, "train/grad": 0.08342094592510346, "train/loss": 0.8335819577503204, "eval/hcp-train-subset/loss": 0.844599429638155, "eval/hcp-val/loss": 0.8477778636640118, "eval/nsd-val/loss": 0.8179272768958923}
+{"epoch": 16, "train/lr": 0.00012053192165988122, "train/grad": 0.08650592072476707, "train/loss": 0.8305048271751404, "eval/hcp-train-subset/loss": 0.8434155352653996, "eval/hcp-val/loss": 0.8484255584978289, "eval/nsd-val/loss": 0.8212427045068433}
+{"epoch": 17, "train/lr": 0.00011973315852507104, "train/grad": 0.08660692447828965, "train/loss": 0.8313600676250458, "eval/hcp-train-subset/loss": 0.8420162287450605, "eval/hcp-val/loss": 0.8471774833817636, "eval/nsd-val/loss": 0.8181885701994742}
+{"epoch": 18, "train/lr": 0.00011887181177170142, "train/grad": 0.08684488090659073, "train/loss": 0.8302282214546204, "eval/hcp-train-subset/loss": 0.841385486625856, "eval/hcp-val/loss": 0.8470755594391977, "eval/nsd-val/loss": 0.8188107282884659}
+{"epoch": 19, "train/lr": 0.00011794882326980209, "train/grad": 0.0904795483747005, "train/loss": 0.8243127859306335, "eval/hcp-train-subset/loss": 0.8412181494697448, "eval/hcp-val/loss": 0.8474279959355632, "eval/nsd-val/loss": 0.8162468200729739}
+{"epoch": 20, "train/lr": 0.00011696520229374954, "train/grad": 0.09209480850567395, "train/loss": 0.8249018680667877, "eval/hcp-train-subset/loss": 0.8407210986460408, "eval/hcp-val/loss": 0.8467553969352476, "eval/nsd-val/loss": 0.8163584614953687}
+{"epoch": 21, "train/lr": 0.00011592202441863837, "train/grad": 0.0931894850331004, "train/loss": 0.8213745770359039, "eval/hcp-train-subset/loss": 0.8395926269792742, "eval/hcp-val/loss": 0.846748645267179, "eval/nsd-val/loss": 0.8193576826203254}
+{"epoch": 22, "train/lr": 0.00011482043034415979, "train/grad": 0.09653842270557998, "train/loss": 0.8189185058784485, "eval/hcp-train-subset/loss": 0.8382422289540691, "eval/hcp-val/loss": 0.8463601902607949, "eval/nsd-val/loss": 0.8195704913908436}
+{"epoch": 23, "train/lr": 0.00011366162464726024, "train/grad": 0.09914825185382126, "train/loss": 0.817587581577301, "eval/hcp-train-subset/loss": 0.8375641809355828, "eval/hcp-val/loss": 0.8478448265983213, "eval/nsd-val/loss": 0.8150216408314244}
+{"epoch": 24, "train/lr": 0.0001124468744649569, "train/grad": 0.1014105735904753, "train/loss": 0.8121148253440857, "eval/hcp-train-subset/loss": 0.836208398303678, "eval/hcp-val/loss": 0.8472535888994893, "eval/nsd-val/loss": 0.8206808124819109}
+{"epoch": 25, "train/lr": 0.0001111775081087387, "train/grad": 0.10224344105117698, "train/loss": 0.8135703350543976, "eval/hcp-train-subset/loss": 0.837048884361021, "eval/hcp-val/loss": 0.8481925716323238, "eval/nsd-val/loss": 0.8211272435803567}
+{"epoch": 26, "train/lr": 0.0001098549136120796, "train/grad": 0.10513349672035337, "train/loss": 0.8112368075370788, "eval/hcp-train-subset/loss": 0.8354599139382762, "eval/hcp-val/loss": 0.8456464582873929, "eval/nsd-val/loss": 0.8214374040403674}
+{"epoch": 27, "train/lr": 0.00010848053721264312, "train/grad": 0.10831208675169796, "train/loss": 0.8063164767456055, "eval/hcp-train-subset/loss": 0.8359815872484638, "eval/hcp-val/loss": 0.8489981101405236, "eval/nsd-val/loss": 0.8251731924472316}
+{"epoch": 28, "train/lr": 0.00010705588177084458, "train/grad": 0.10739962854686548, "train/loss": 0.8061359535312652, "eval/hcp-train-subset/loss": 0.8337298381713129, "eval/hcp-val/loss": 0.8473209671435817, "eval/nsd-val/loss": 0.822485173902204}
+{"epoch": 29, "train/lr": 0.00010558250512649171, "train/grad": 0.10951835639469526, "train/loss": 0.8040780637168884, "eval/hcp-train-subset/loss": 0.8333918711831493, "eval/hcp-val/loss": 0.8472099111926171, "eval/nsd-val/loss": 0.8201563166033837}
+{"epoch": 30, "train/lr": 0.00010406201839531515, "train/grad": 0.1117721069019846, "train/loss": 0.8024293426990509, "eval/hcp-train-subset/loss": 0.8320255010358749, "eval/hcp-val/loss": 0.8458288571526927, "eval/nsd-val/loss": 0.8205841991209215}
+{"epoch": 31, "train/lr": 0.00010249608420723018, "train/grad": 0.11285329373701262, "train/loss": 0.8015312059211731, "eval/hcp-train-subset/loss": 0.8308570365751943, "eval/hcp-val/loss": 0.8464438415342762, "eval/nsd-val/loss": 0.8173661347358457}
+{"epoch": 32, "train/lr": 0.00010088641488828097, "train/grad": 0.11484475577340751, "train/loss": 0.7999681686973572, "eval/hcp-train-subset/loss": 0.8326494299596355, "eval/hcp-val/loss": 0.8472545560329191, "eval/nsd-val/loss": 0.8233350563433862}
+{"epoch": 33, "train/lr": 9.923477058823526e-05, "train/grad": 0.11509341292043905, "train/loss": 0.8015153414726257, "eval/hcp-train-subset/loss": 0.8318735659122467, "eval/hcp-val/loss": 0.8483233586434396, "eval/nsd-val/loss": 0.8193432265712369}
+{"epoch": 34, "train/lr": 9.754295735588547e-05, "train/grad": 0.12044507258948715, "train/loss": 0.7947001446723938, "eval/hcp-train-subset/loss": 0.8302329182624817, "eval/hcp-val/loss": 0.8471583627885387, "eval/nsd-val/loss": 0.8202403928003004}
+{"epoch": 35, "train/lr": 9.581282516416285e-05, "train/grad": 0.12255365371699332, "train/loss": 0.7941262307834626, "eval/hcp-train-subset/loss": 0.8295572803866479, "eval/hcp-val/loss": 0.8467848358615753, "eval/nsd-val/loss": 0.8218030641155858}
+{"epoch": 36, "train/lr": 9.404626588721676e-05, "train/grad": 0.12224419595089572, "train/loss": 0.7922537671661377, "eval/hcp-train-subset/loss": 0.8291280346532022, "eval/hcp-val/loss": 0.8465725298850767, "eval/nsd-val/loss": 0.8242481802740405}
+{"epoch": 37, "train/lr": 9.224521123168153e-05, "train/grad": 0.12066988228477821, "train/loss": 0.7958024103832245, "eval/hcp-train-subset/loss": 0.8275175296491192, "eval/hcp-val/loss": 0.8482665563783338, "eval/nsd-val/loss": 0.8213300166591522}
+{"epoch": 38, "train/lr": 9.041163062437843e-05, "train/grad": 0.12498596353796938, "train/loss": 0.7918025265407562, "eval/hcp-train-subset/loss": 0.8277492638557188, "eval/hcp-val/loss": 0.8491401335885448, "eval/nsd-val/loss": 0.8202772130889278}
+{"epoch": 39, "train/lr": 8.85475290587822e-05, "train/grad": 0.1266783281152087, "train/loss": 0.7900685538864136, "eval/hcp-train-subset/loss": 0.8259285938355231, "eval/hcp-val/loss": 0.847166121006012, "eval/nsd-val/loss": 0.8211023298002058}
+{"epoch": 40, "train/lr": 8.665494490258622e-05, "train/grad": 0.13049605029944975, "train/loss": 0.7870240499305725, "eval/hcp-train-subset/loss": 0.8242251892243663, "eval/hcp-val/loss": 0.8471699126305119, "eval/nsd-val/loss": 0.8193107612671391}
+{"epoch": 41, "train/lr": 8.473594766877838e-05, "train/grad": 0.13106415353981385, "train/loss": 0.7880352628040314, "eval/hcp-train-subset/loss": 0.825802787657707, "eval/hcp-val/loss": 0.8475084996992542, "eval/nsd-val/loss": 0.8229518696185081}
+{"epoch": 42, "train/lr": 8.279263575265999e-05, "train/grad": 0.13351266802182463, "train/loss": 0.7878979594993591, "eval/hcp-train-subset/loss": 0.8238233750866305, "eval/hcp-val/loss": 0.8466926755443696, "eval/nsd-val/loss": 0.82596541220142}
+{"epoch": 43, "train/lr": 8.082713413727944e-05, "train/grad": 0.1333800126964205, "train/loss": 0.7846602644443512, "eval/hcp-train-subset/loss": 0.8250910841649578, "eval/hcp-val/loss": 0.8497149136758619, "eval/nsd-val/loss": 0.8242935845928807}
+{"epoch": 44, "train/lr": 7.884159206979602e-05, "train/grad": 0.13576362153131244, "train/loss": 0.7825197706127167, "eval/hcp-train-subset/loss": 0.8225518676542467, "eval/hcp-val/loss": 0.8466382363150197, "eval/nsd-val/loss": 0.8218715229342061}
+{"epoch": 45, "train/lr": 7.683818071130916e-05, "train/grad": 0.13701451805428566, "train/loss": 0.7842067969989777, "eval/hcp-train-subset/loss": 0.8225269673332092, "eval/hcp-val/loss": 0.8475705202548734, "eval/nsd-val/loss": 0.8255816390437465}
+{"epoch": 46, "train/lr": 7.481909076272522e-05, "train/grad": 0.13963640641408556, "train/loss": 0.7814966522789002, "eval/hcp-train-subset/loss": 0.8224642276763916, "eval/hcp-val/loss": 0.848196211361116, "eval/nsd-val/loss": 0.825363798487571}
+{"epoch": 47, "train/lr": 7.278653006925963e-05, "train/grad": 0.14344608756052088, "train/loss": 0.778756197423935, "eval/hcp-train-subset/loss": 0.8227367987555843, "eval/hcp-val/loss": 0.8461911735996124, "eval/nsd-val/loss": 0.8219214918151978}
+{"epoch": 48, "train/lr": 7.074272120618864e-05, "train/grad": 0.1446737284438599, "train/loss": 0.7793798372936249, "eval/hcp-train-subset/loss": 0.822237059954674, "eval/hcp-val/loss": 0.8458063256356024, "eval/nsd-val/loss": 0.8215657597587954}
+{"epoch": 49, "train/lr": 6.868989904849677e-05, "train/grad": 0.14779189460686784, "train/loss": 0.7768739538955689, "eval/hcp-train-subset/loss": 0.8226848571531234, "eval/hcp-val/loss": 0.8501322394417178, "eval/nsd-val/loss": 0.8245307764699382}
+{"epoch": 50, "train/lr": 6.6630308327075e-05, "train/grad": 0.14948619369217092, "train/loss": 0.7769580447387695, "eval/hcp-train-subset/loss": 0.8208253095226903, "eval/hcp-val/loss": 0.8476070482884684, "eval/nsd-val/loss": 0.8238639908452188}
+{"epoch": 51, "train/lr": 6.456620117413798e-05, "train/grad": 0.14907017342683046, "train/loss": 0.7799463856697082, "eval/hcp-train-subset/loss": 0.8229544720342082, "eval/hcp-val/loss": 0.8507427592431346, "eval/nsd-val/loss": 0.8285107901019435}
+{"epoch": 52, "train/lr": 6.249983466055255e-05, "train/grad": 0.15441324675495346, "train/loss": 0.7744219431495667, "eval/hcp-train-subset/loss": 0.8205355521171324, "eval/hcp-val/loss": 0.8477706466951678, "eval/nsd-val/loss": 0.8243187521734545}
+{"epoch": 53, "train/lr": 6.0433468327763305e-05, "train/grad": 0.15424857152418428, "train/loss": 0.7726257054615021, "eval/hcp-train-subset/loss": 0.821001248974954, "eval/hcp-val/loss": 0.8487818606438176, "eval/nsd-val/loss": 0.8230471216863201}
+{"epoch": 54, "train/lr": 5.83693617170174e-05, "train/grad": 0.15512333795185013, "train/loss": 0.7742177011108399, "eval/hcp-train-subset/loss": 0.8184664480147823, "eval/hcp-val/loss": 0.8480845266772855, "eval/nsd-val/loss": 0.8325743204162966}
+{"epoch": 55, "train/lr": 5.6309771898588165e-05, "train/grad": 0.159109075849049, "train/loss": 0.7708654251194, "eval/hcp-train-subset/loss": 0.818173090296407, "eval/hcp-val/loss": 0.8488256873623017, "eval/nsd-val/loss": 0.82753588787971}
+{"epoch": 56, "train/lr": 5.4256951003704155e-05, "train/grad": 0.16024275188585899, "train/loss": 0.7731923755073548, "eval/hcp-train-subset/loss": 0.8166144076854952, "eval/hcp-val/loss": 0.8486992726402898, "eval/nsd-val/loss": 0.8235403770400632}
+{"epoch": 57, "train/lr": 5.221314376187425e-05, "train/grad": 0.1603284323630055, "train/loss": 0.7710657037734986, "eval/hcp-train-subset/loss": 0.817559951736081, "eval/hcp-val/loss": 0.8496170774582894, "eval/nsd-val/loss": 0.8241921115306116}
+{"epoch": 58, "train/lr": 5.018058504631059e-05, "train/grad": 0.1657444095471315, "train/loss": 0.7702999696540832, "eval/hcp-train-subset/loss": 0.8182939540955328, "eval/hcp-val/loss": 0.8506767230649148, "eval/nsd-val/loss": 0.8261494742285821}
+{"epoch": 59, "train/lr": 4.816149743012713e-05, "train/grad": 0.1661175585292102, "train/loss": 0.7699218058681488, "eval/hcp-train-subset/loss": 0.8165130346052109, "eval/hcp-val/loss": 0.8512075216539444, "eval/nsd-val/loss": 0.8311463873232564}
+{"epoch": 60, "train/lr": 4.615808875598772e-05, "train/grad": 0.16927230839125726, "train/loss": 0.7689740070819855, "eval/hcp-train-subset/loss": 0.813839348093156, "eval/hcp-val/loss": 0.8473740470024848, "eval/nsd-val/loss": 0.8273406644021312}
+{"epoch": 61, "train/lr": 4.417254972186445e-05, "train/grad": 0.16812480998281557, "train/loss": 0.7700548267459869, "eval/hcp-train-subset/loss": 0.8159355261633473, "eval/hcp-val/loss": 0.8511628124021715, "eval/nsd-val/loss": 0.8287307437389128}
+{"epoch": 62, "train/lr": 4.220705148553925e-05, "train/grad": 0.17203274478708966, "train/loss": 0.7684106969451904, "eval/hcp-train-subset/loss": 0.8141190630774344, "eval/hcp-val/loss": 0.8506488982708224, "eval/nsd-val/loss": 0.8252008480410422}
+{"epoch": 63, "train/lr": 4.026374329047657e-05, "train/grad": 0.17539655518103395, "train/loss": 0.7668382850933075, "eval/hcp-train-subset/loss": 0.8128110631819694, "eval/hcp-val/loss": 0.8500380169960761, "eval/nsd-val/loss": 0.8299871586984203}
+{"epoch": 64, "train/lr": 3.834475011565652e-05, "train/grad": 0.1766963716633934, "train/loss": 0.7661759544181824, "eval/hcp-train-subset/loss": 0.8121587326449733, "eval/hcp-val/loss": 0.8507088057456478, "eval/nsd-val/loss": 0.8320937281654727}
+{"epoch": 65, "train/lr": 3.6452170351940815e-05, "train/grad": 0.18066113911302004, "train/loss": 0.7650150953578949, "eval/hcp-train-subset/loss": 0.8114573071079869, "eval/hcp-val/loss": 0.8494856194142373, "eval/nsd-val/loss": 0.8296796179586842}
+{"epoch": 66, "train/lr": 3.458807350751516e-05, "train/grad": 0.18205371822937655, "train/loss": 0.765590187997818, "eval/hcp-train-subset/loss": 0.8118510890391565, "eval/hcp-val/loss": 0.8515608752927473, "eval/nsd-val/loss": 0.8312406482235077}
+{"epoch": 67, "train/lr": 3.2754497944910164e-05, "train/grad": 0.18485950638996806, "train/loss": 0.7613125121498108, "eval/hcp-train-subset/loss": 0.8097735162704222, "eval/hcp-val/loss": 0.850867276230166, "eval/nsd-val/loss": 0.8317391036018249}
+{"epoch": 68, "train/lr": 3.0953448652083367e-05, "train/grad": 0.18474739708039772, "train/loss": 0.763850643529892, "eval/hcp-train-subset/loss": 0.81056153870398, "eval/hcp-val/loss": 0.849835213153593, "eval/nsd-val/loss": 0.8296023107344105}
+{"epoch": 69, "train/lr": 2.9186895049993948e-05, "train/grad": 0.1862102629039699, "train/loss": 0.7638243045425415, "eval/hcp-train-subset/loss": 0.8109314624340304, "eval/hcp-val/loss": 0.8509873872803103, "eval/nsd-val/loss": 0.8314043004666606}
+{"epoch": 70, "train/lr": 2.7456768839068717e-05, "train/grad": 0.18901930939169412, "train/loss": 0.762642031545639, "eval/hcp-train-subset/loss": 0.8113910152066138, "eval/hcp-val/loss": 0.851910739175735, "eval/nsd-val/loss": 0.83644282529431}
+{"epoch": 71, "train/lr": 2.5764961886919063e-05, "train/grad": 0.19357516208875844, "train/loss": 0.7589323935794831, "eval/hcp-train-subset/loss": 0.8093348501190063, "eval/hcp-val/loss": 0.8521012759977772, "eval/nsd-val/loss": 0.8303937258258942}
+{"epoch": 72, "train/lr": 2.411332415960724e-05, "train/grad": 0.19075288377285995, "train/loss": 0.7621443289089203, "eval/hcp-train-subset/loss": 0.8094597305020979, "eval/hcp-val/loss": 0.8516703565274516, "eval/nsd-val/loss": 0.8306347295161216}
+{"epoch": 73, "train/lr": 2.2503661698739544e-05, "train/grad": 0.194079022159839, "train/loss": 0.7613156198692321, "eval/hcp-train-subset/loss": 0.8084090332831105, "eval/hcp-val/loss": 0.8505908443081763, "eval/nsd-val/loss": 0.8320958989281808}
+{"epoch": 74, "train/lr": 2.0937734646583902e-05, "train/grad": 0.19617348185423414, "train/loss": 0.758829739599228, "eval/hcp-train-subset/loss": 0.8082149951688705, "eval/hcp-val/loss": 0.85311389450104, "eval/nsd-val/loss": 0.8355936763748046}
+{"epoch": 75, "train/lr": 1.9417255321381202e-05, "train/grad": 0.19898682391550973, "train/loss": 0.758649975862503, "eval/hcp-train-subset/loss": 0.807007182990351, "eval/hcp-val/loss": 0.852161685305257, "eval/nsd-val/loss": 0.831628736949736}
+{"epoch": 76, "train/lr": 1.7943886344950134e-05, "train/grad": 0.202016500500955, "train/loss": 0.7555100770187378, "eval/hcp-train-subset/loss": 0.8071074899165861, "eval/hcp-val/loss": 0.8531546285075526, "eval/nsd-val/loss": 0.8378788280871606}
+{"epoch": 77, "train/lr": 1.651923882463461e-05, "train/grad": 0.20085163550680687, "train/loss": 0.758845798034668, "eval/hcp-train-subset/loss": 0.805934811792066, "eval/hcp-val/loss": 0.8516272594851833, "eval/nsd-val/loss": 0.8313788252492105}
+{"epoch": 78, "train/lr": 1.5144870591581508e-05, "train/grad": 0.20067481323778485, "train/loss": 0.7598386356449127, "eval/hcp-train-subset/loss": 0.8075570992885097, "eval/hcp-val/loss": 0.8549752802618088, "eval/nsd-val/loss": 0.8385972390251775}
+{"epoch": 79, "train/lr": 1.3822284497275662e-05, "train/grad": 0.20369711179679098, "train/loss": 0.7582556018352509, "eval/hcp-train-subset/loss": 0.8066243365887673, "eval/hcp-val/loss": 0.8542477132812623, "eval/nsd-val/loss": 0.8336295466269216}
+{"epoch": 80, "train/lr": 1.2552926770192975e-05, "train/grad": 0.20598111606505806, "train/loss": 0.7568941364097596, "eval/hcp-train-subset/loss": 0.8060596364159738, "eval/hcp-val/loss": 0.853348538760216, "eval/nsd-val/loss": 0.8343931551902525}
+{"epoch": 81, "train/lr": 1.1338185434371453e-05, "train/grad": 0.205365603002654, "train/loss": 0.7579613129138947, "eval/hcp-train-subset/loss": 0.8066622461042097, "eval/hcp-val/loss": 0.8540809548670246, "eval/nsd-val/loss": 0.8345428551397016}
+{"epoch": 82, "train/lr": 1.0179388791627326e-05, "train/grad": 0.2063855711351627, "train/loss": 0.7581566770744324, "eval/hcp-train-subset/loss": 0.8050795428214534, "eval/hcp-val/loss": 0.854724302407234, "eval/nsd-val/loss": 0.8341913675108263}
+{"epoch": 83, "train/lr": 9.07780396907607e-06, "train/grad": 0.20804161543471728, "train/loss": 0.7569181223106384, "eval/hcp-train-subset/loss": 0.8043526478352085, "eval/hcp-val/loss": 0.8539002383908918, "eval/nsd-val/loss": 0.837493811884234}
+{"epoch": 84, "train/lr": 8.034635533547902e-06, "train/grad": 0.20769129680152434, "train/loss": 0.7569669394397736, "eval/hcp-train-subset/loss": 0.8038652837276459, "eval/hcp-val/loss": 0.8542846431655269, "eval/nsd-val/loss": 0.835448625587648}
+{"epoch": 85, "train/lr": 7.051024174411275e-06, "train/grad": 0.2100026357761571, "train/loss": 0.7560576700878143, "eval/hcp-train-subset/loss": 0.8040805362885998, "eval/hcp-val/loss": 0.8552798109669839, "eval/nsd-val/loss": 0.8365880866204539}
+{"epoch": 86, "train/lr": 6.1280454562463606e-06, "train/grad": 0.20904529661927698, "train/loss": 0.7561215976428985, "eval/hcp-train-subset/loss": 0.8040098730594881, "eval/hcp-val/loss": 0.8547314001667884, "eval/nsd-val/loss": 0.8367736704887883}
+{"epoch": 87, "train/lr": 5.266708642730326e-06, "train/grad": 0.21029932399559007, "train/loss": 0.7572222897434234, "eval/hcp-train-subset/loss": 0.8038484886769326, "eval/hcp-val/loss": 0.8548858694491848, "eval/nsd-val/loss": 0.8352917355875815}
+{"epoch": 88, "train/lr": 4.467955593022733e-06, "train/grad": 0.20896513257104135, "train/loss": 0.7596117394256592, "eval/hcp-train-subset/loss": 0.8030020556142253, "eval/hcp-val/loss": 0.8546109737888459, "eval/nsd-val/loss": 0.836073363019574}
+{"epoch": 89, "train/lr": 3.732659731856291e-06, "train/grad": 0.2119673614108278, "train/loss": 0.7557221819114686, "eval/hcp-train-subset/loss": 0.8024012696358466, "eval/hcp-val/loss": 0.8550502917458934, "eval/nsd-val/loss": 0.8359243369871571}
+{"epoch": 90, "train/lr": 3.0616250944596583e-06, "train/grad": 0.2126477457196589, "train/loss": 0.7542780232810974, "eval/hcp-train-subset/loss": 0.8022062163199147, "eval/hcp-val/loss": 0.855003701102349, "eval/nsd-val/loss": 0.8363813113781714}
+{"epoch": 91, "train/lr": 2.4555854473568305e-06, "train/grad": 0.21149170586765384, "train/loss": 0.7561877486610412, "eval/hcp-train-subset/loss": 0.8021432690082058, "eval/hcp-val/loss": 0.8551743117070967, "eval/nsd-val/loss": 0.8371614756122712}
+{"epoch": 92, "train/lr": 1.915203486004091e-06, "train/grad": 0.2108704291185093, "train/loss": 0.7580078503417969, "eval/hcp-train-subset/loss": 0.802458381460559, "eval/hcp-val/loss": 0.8554909911847883, "eval/nsd-val/loss": 0.8363811152596627}
+{"epoch": 93, "train/lr": 1.4410701101423926e-06, "train/grad": 0.2140366171967243, "train/loss": 0.7550402507400513, "eval/hcp-train-subset/loss": 0.8013764879395885, "eval/hcp-val/loss": 0.8543134581658148, "eval/nsd-val/loss": 0.8350671481701636}
+{"epoch": 94, "train/lr": 1.0337037776570775e-06, "train/grad": 0.21287263802228706, "train/loss": 0.7556635715484619, "eval/hcp-train-subset/loss": 0.800883170097105, "eval/hcp-val/loss": 0.855681755850392, "eval/nsd-val/loss": 0.8359930332629911}
+{"epoch": 95, "train/lr": 6.935499376518293e-07, "train/grad": 0.2141508347666065, "train/loss": 0.7559113196182251, "eval/hcp-train-subset/loss": 0.8008155601639901, "eval/hcp-val/loss": 0.8544887063964721, "eval/nsd-val/loss": 0.8359333392112486}
+{"epoch": 96, "train/lr": 4.209805433566085e-07, "train/grad": 0.20990763355413794, "train/loss": 0.7609630836582184, "eval/hcp-train-subset/loss": 0.8009506790868698, "eval/hcp-val/loss": 0.8545020107300051, "eval/nsd-val/loss": 0.8354712013275393}
+{"epoch": 97, "train/lr": 2.1629364540224422e-07, "train/grad": 0.21346596403931944, "train/loss": 0.7584872759819031, "eval/hcp-train-subset/loss": 0.8010522723197937, "eval/hcp-val/loss": 0.8541870847825082, "eval/nsd-val/loss": 0.8356792503787626}
+{"epoch": 98, "train/lr": 7.971306590647406e-08, "train/grad": 0.21411126424482602, "train/loss": 0.7558657725811004, "eval/hcp-train-subset/loss": 0.8010722917895163, "eval/hcp-val/loss": 0.8545104034485356, "eval/nsd-val/loss": 0.8360496576755277}
+{"epoch": 99, "train/lr": 1.1388153727718725e-08, "train/grad": 0.2135729413722274, "train/loss": 0.7562797626209259, "eval/hcp-train-subset/loss": 0.8008644936546203, "eval/hcp-val/loss": 0.8544662075657998, "eval/nsd-val/loss": 0.8362287629035211}

data_scaling/n400_2/eval_v2/aabc_age__patch__logistic/config.yaml

@@ -0,0 +1,30 @@
+output_root: experiments/data_scaling/output
+name_prefix: eval_logistic
+remote_root: null
+notes: data scaling experiment n400_2; eval v2 (aabc_age patch logistic)
+model_kwargs:
+  ckpt_path: experiments/data_scaling/output/data_scaling/n400_2/pretrain/checkpoint-best.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: data_scaling/n400_2/eval_v2/aabc_age__patch__logistic
+model: flat_mae
+representation: patch
+dataset: aabc_age
+distributed: false
+output_dir: experiments/data_scaling/output/data_scaling/n400_2/eval_v2/aabc_age__patch__logistic
+remote_dir: null