| 2026-08-08:22:22:17 INFO [_cli.run:388] Selected Tasks: ['humaneval', 'mbpp'] |
| 2026-08-08:22:22:19 INFO [evaluator:214] Setting random seed to 0 | Setting numpy seed to 1234 | Setting torch manual seed to 1234 | Setting fewshot manual seed to 1234 |
| 2026-08-08:22:22:19 INFO [evaluator:239] Initializing local-completions model, with arguments: {'model': 'student', 'base_url': 'http://127.0.0.1:8460/v1/completions', 'tokenizer': 'outputs/release/healed/release-offpolicy-math-keep75/step0150', 'num_concurrent': 48, 'tokenized_requests': False, 'max_retries': 3} |
| 2026-08-08:22:22:19 INFO [models.openai_completions:42] Remote tokenizer not supported. Using huggingface tokenizer backend. |
| 2026-08-08:22:22:19 INFO [models.api_models:179] Using max length 2048 - 1 |
| 2026-08-08:22:22:19 INFO [models.api_models:200] Using tokenizer huggingface |
| 2026-08-08:22:22:26 INFO [evaluator_utils:446] Selected tasks: |
| 2026-08-08:22:22:26 INFO [evaluator_utils:480] Task: humaneval (humaneval/humaneval.yaml) |
| 2026-08-08:22:22:26 INFO [evaluator_utils:480] Task: mbpp (mbpp/mbpp.yaml) |
| 2026-08-08:22:22:26 INFO [evaluator:314] humaneval: Using gen_kwargs: {'until': ['\nclass', '\ndef', '\n#', '\nif', '\nprint'], 'max_gen_toks': 1024, 'do_sample': False} |
| 2026-08-08:22:22:26 INFO [evaluator:314] mbpp: Using gen_kwargs: {'until': ['[DONE]'], 'do_sample': False} |
| 2026-08-08:22:22:26 INFO [api.task:312] Building contexts for humaneval on rank 0... |
|
0%| | 0/164 [00:00<?, ?it/s]
88%|βββββββββ | 145/164 [00:00<00:00, 1444.89it/s]
100%|ββββββββββ| 164/164 [00:00<00:00, 1476.50it/s] |
| 2026-08-08:22:22:26 INFO [api.task:312] Building contexts for mbpp on rank 0... |
|
0%| | 0/500 [00:00<?, ?it/s]
3%|β | 17/500 [00:00<00:02, 162.52it/s]
7%|β | 34/500 [00:00<00:02, 163.57it/s]
10%|β | 51/500 [00:00<00:02, 163.84it/s]
14%|ββ | 68/500 [00:00<00:02, 163.63it/s]
17%|ββ | 85/500 [00:00<00:02, 163.54it/s]
20%|ββ | 102/500 [00:00<00:02, 163.70it/s]
24%|βββ | 119/500 [00:00<00:02, 163.63it/s]
27%|βββ | 136/500 [00:00<00:02, 164.21it/s]
31%|βββ | 153/500 [00:00<00:02, 164.55it/s]
34%|ββββ | 170/500 [00:01<00:02, 164.83it/s]
37%|ββββ | 187/500 [00:01<00:01, 164.74it/s]
41%|ββββ | 204/500 [00:01<00:01, 165.03it/s]
44%|βββββ | 221/500 [00:01<00:01, 165.10it/s]
48%|βββββ | 238/500 [00:01<00:01, 164.55it/s]
51%|βββββ | 255/500 [00:01<00:01, 164.65it/s]
54%|ββββββ | 272/500 [00:01<00:01, 164.82it/s]
58%|ββββββ | 289/500 [00:01<00:01, 164.92it/s]
61%|ββββββ | 306/500 [00:01<00:01, 164.88it/s]
65%|βββββββ | 323/500 [00:01<00:01, 165.02it/s]
68%|βββββββ | 340/500 [00:02<00:00, 165.10it/s]
71%|ββββββββ | 357/500 [00:02<00:00, 165.14it/s]
75%|ββββββββ | 374/500 [00:02<00:00, 165.25it/s]
78%|ββββββββ | 391/500 [00:02<00:00, 165.28it/s]
82%|βββββββββ | 408/500 [00:02<00:00, 165.39it/s]
85%|βββββββββ | 425/500 [00:02<00:00, 165.71it/s]
88%|βββββββββ | 442/500 [00:02<00:00, 166.16it/s]
92%|ββββββββββ| 459/500 [00:02<00:00, 165.99it/s]
95%|ββββββββββ| 476/500 [00:02<00:00, 165.78it/s]
99%|ββββββββββ| 493/500 [00:02<00:00, 165.64it/s]
100%|ββββββββββ| 500/500 [00:03<00:00, 164.91it/s] |
| 2026-08-08:22:22:29 INFO [evaluator:585] Running generate_until requests |
| 2026-08-08:22:22:29 INFO [models.api_models:747] Tokenized requests are disabled. Context + generation length is not checked. |
|
Requesting API: 0%| | 0/164 [00:00<?, ?it/s]
Requesting API: 1%| | 1/164 [00:00<02:39, 1.02it/s]
Requesting API: 6%|β | 10/164 [00:01<00:15, 10.22it/s]
Requesting API: 8%|β | 13/164 [00:01<00:13, 11.53it/s]
Requesting API: 9%|β | 15/164 [00:01<00:12, 12.36it/s]
Requesting API: 11%|β | 18/164 [00:01<00:10, 14.46it/s]
Requesting API: 13%|ββ | 22/164 [00:01<00:07, 17.82it/s]
Requesting API: 15%|ββ | 25/164 [00:01<00:07, 19.18it/s]
Requesting API: 17%|ββ | 28/164 [00:02<00:06, 20.93it/s]
Requesting API: 19%|ββ | 31/164 [00:02<00:06, 20.30it/s]
Requesting API: 22%|βββ | 36/164 [00:02<00:05, 24.13it/s]
Requesting API: 24%|βββ | 39/164 [00:02<00:06, 20.50it/s]
Requesting API: 26%|βββ | 43/164 [00:02<00:05, 23.89it/s]
Requesting API: 30%|βββ | 49/164 [00:02<00:03, 29.63it/s]
Requesting API: 34%|ββββ | 55/164 [00:02<00:03, 32.73it/s]
Requesting API: 38%|ββββ | 62/164 [00:03<00:02, 38.54it/s]
Requesting API: 41%|ββββ | 67/164 [00:03<00:02, 39.48it/s]
Requesting API: 44%|βββββ | 72/164 [00:03<00:02, 34.72it/s]
Requesting API: 47%|βββββ | 77/164 [00:03<00:02, 37.01it/s]
Requesting API: 51%|βββββ | 83/164 [00:03<00:02, 37.01it/s]
Requesting API: 55%|ββββββ | 90/164 [00:03<00:01, 43.94it/s]
Requesting API: 59%|ββββββ | 96/164 [00:03<00:01, 43.57it/s]
Requesting API: 62%|βββββββ | 101/164 [00:04<00:01, 38.40it/s]
Requesting API: 65%|βββββββ | 107/164 [00:04<00:01, 42.24it/s]
Requesting API: 69%|βββββββ | 113/164 [00:04<00:01, 43.81it/s]
Requesting API: 76%|ββββββββ | 124/164 [00:04<00:00, 58.62it/s]
Requesting API: 80%|ββββββββ | 131/164 [00:04<00:00, 61.05it/s]
Requesting API: 85%|βββββββββ | 139/164 [00:04<00:00, 63.96it/s]
Requesting API: 89%|βββββββββ | 146/164 [00:04<00:00, 45.53it/s]
Requesting API: 93%|ββββββββββ| 152/164 [00:05<00:00, 29.40it/s]
Requesting API: 96%|ββββββββββ| 157/164 [00:05<00:00, 30.88it/s]
Requesting API: 99%|ββββββββββ| 162/164 [00:06<00:00, 17.88it/s]
Requesting API: 100%|ββββββββββ| 164/164 [00:15<00:00, 10.58it/s] |
|
Requesting API: 0%| | 0/500 [00:00<?, ?it/s]
Requesting API: 0%| | 1/500 [00:01<09:28, 1.14s/it]
Requesting API: 1%| | 4/500 [00:01<02:01, 4.07it/s]
Requesting API: 2%|β | 8/500 [00:01<00:56, 8.65it/s]
Requesting API: 4%|β | 18/500 [00:01<00:23, 20.83it/s]
Requesting API: 4%|β | 22/500 [00:01<00:20, 22.95it/s]
Requesting API: 5%|β | 27/500 [00:01<00:21, 22.22it/s]
Requesting API: 6%|β | 31/500 [00:02<00:21, 21.86it/s]
Requesting API: 8%|β | 38/500 [00:02<00:16, 28.14it/s]
Requesting API: 9%|β | 43/500 [00:02<00:14, 30.76it/s]
Requesting API: 9%|β | 47/500 [00:02<00:15, 28.50it/s]
Requesting API: 11%|β | 54/500 [00:02<00:12, 35.97it/s]
Requesting API: 12%|ββ | 59/500 [00:02<00:11, 38.55it/s]
Requesting API: 13%|ββ | 64/500 [00:02<00:11, 38.17it/s]
Requesting API: 14%|ββ | 71/500 [00:03<00:09, 45.19it/s]
Requesting API: 15%|ββ | 76/500 [00:03<00:14, 29.85it/s]
Requesting API: 16%|ββ | 80/500 [00:03<00:13, 31.20it/s]
Requesting API: 17%|ββ | 84/500 [00:03<00:13, 31.64it/s]
Requesting API: 18%|ββ | 88/500 [00:03<00:13, 30.24it/s]
Requesting API: 18%|ββ | 92/500 [00:03<00:13, 29.65it/s]
Requesting API: 20%|ββ | 98/500 [00:03<00:12, 33.22it/s]
Requesting API: 21%|βββ | 107/500 [00:04<00:09, 43.23it/s]
Requesting API: 22%|βββ | 112/500 [00:04<00:08, 43.49it/s]
Requesting API: 23%|βββ | 117/500 [00:04<00:09, 40.92it/s]
Requesting API: 24%|βββ | 122/500 [00:04<00:14, 25.72it/s]
Requesting API: 25%|βββ | 126/500 [00:04<00:13, 26.94it/s]
Requesting API: 26%|βββ | 132/500 [00:04<00:11, 32.29it/s]
Requesting API: 27%|βββ | 136/500 [00:05<00:10, 33.43it/s]
Requesting API: 28%|βββ | 141/500 [00:05<00:09, 36.93it/s]
Requesting API: 29%|βββ | 146/500 [00:05<00:09, 35.49it/s]
Requesting API: 30%|βββ | 150/500 [00:05<00:14, 24.77it/s]
Requesting API: 32%|ββββ | 159/500 [00:05<00:10, 33.90it/s]
Requesting API: 33%|ββββ | 164/500 [00:06<00:10, 31.40it/s]
Requesting API: 34%|ββββ | 168/500 [00:06<00:10, 30.67it/s]
Requesting API: 35%|ββββ | 173/500 [00:06<00:09, 33.72it/s]
Requesting API: 36%|ββββ | 178/500 [00:06<00:08, 37.30it/s]
Requesting API: 37%|ββββ | 183/500 [00:06<00:08, 36.04it/s]
Requesting API: 37%|ββββ | 187/500 [00:06<00:09, 34.36it/s]
Requesting API: 38%|ββββ | 191/500 [00:06<00:10, 29.10it/s]
Requesting API: 39%|ββββ | 195/500 [00:06<00:09, 30.57it/s]
Requesting API: 40%|ββββ | 199/500 [00:07<00:10, 29.96it/s]
Requesting API: 41%|ββββ | 203/500 [00:07<00:09, 29.71it/s]
Requesting API: 41%|βββββ | 207/500 [00:07<00:09, 30.69it/s]
Requesting API: 43%|βββββ | 214/500 [00:07<00:07, 39.23it/s]
Requesting API: 44%|βββββ | 219/500 [00:07<00:07, 37.54it/s]
Requesting API: 45%|βββββ | 225/500 [00:07<00:06, 41.36it/s]
Requesting API: 46%|βββββ | 230/500 [00:08<00:09, 29.13it/s]
Requesting API: 47%|βββββ | 236/500 [00:08<00:07, 33.84it/s]
Requesting API: 48%|βββββ | 240/500 [00:08<00:08, 32.41it/s]
Requesting API: 49%|βββββ | 244/500 [00:08<00:07, 33.17it/s]
Requesting API: 50%|βββββ | 249/500 [00:08<00:08, 30.67it/s]
Requesting API: 51%|βββββ | 254/500 [00:08<00:07, 34.66it/s]
Requesting API: 52%|ββββββ | 258/500 [00:08<00:07, 32.54it/s]
Requesting API: 52%|ββββββ | 262/500 [00:08<00:07, 31.93it/s]
Requesting API: 54%|ββββββ | 268/500 [00:09<00:06, 37.34it/s]
Requesting API: 54%|ββββββ | 272/500 [00:09<00:06, 34.10it/s]
Requesting API: 55%|ββββββ | 276/500 [00:09<00:07, 31.52it/s]
Requesting API: 56%|ββββββ | 280/500 [00:09<00:06, 31.64it/s]
Requesting API: 57%|ββββββ | 284/500 [00:09<00:06, 32.42it/s]
Requesting API: 58%|ββββββ | 288/500 [00:09<00:07, 29.29it/s]
Requesting API: 59%|ββββββ | 295/500 [00:09<00:05, 37.92it/s]
Requesting API: 60%|ββββββ | 300/500 [00:09<00:04, 40.82it/s]
Requesting API: 61%|ββββββ | 305/500 [00:10<00:04, 42.35it/s]
Requesting API: 62%|βββββββ | 312/500 [00:10<00:03, 47.43it/s]
Requesting API: 63%|βββββββ | 317/500 [00:10<00:04, 37.86it/s]
Requesting API: 65%|βββββββ | 323/500 [00:10<00:04, 41.71it/s]
Requesting API: 66%|βββββββ | 328/500 [00:10<00:04, 38.55it/s]
Requesting API: 67%|βββββββ | 333/500 [00:10<00:04, 38.38it/s]
Requesting API: 68%|βββββββ | 338/500 [00:10<00:04, 38.23it/s]
Requesting API: 69%|βββββββ | 343/500 [00:11<00:03, 40.29it/s]
Requesting API: 70%|βββββββ | 348/500 [00:11<00:04, 37.85it/s]
Requesting API: 71%|βββββββ | 355/500 [00:11<00:03, 44.87it/s]
Requesting API: 72%|ββββββββ | 360/500 [00:11<00:03, 43.75it/s]
Requesting API: 73%|ββββββββ | 366/500 [00:11<00:03, 40.52it/s]
Requesting API: 74%|ββββββββ | 372/500 [00:11<00:03, 40.31it/s]
Requesting API: 75%|ββββββββ | 377/500 [00:11<00:03, 39.43it/s]
Requesting API: 77%|ββββββββ | 384/500 [00:12<00:02, 43.70it/s]
Requesting API: 78%|ββββββββ | 390/500 [00:12<00:02, 46.35it/s]
Requesting API: 79%|ββββββββ | 395/500 [00:12<00:02, 46.86it/s]
Requesting API: 80%|ββββββββ | 400/500 [00:12<00:02, 46.69it/s]
Requesting API: 81%|βββββββββ | 407/500 [00:12<00:01, 51.19it/s]
Requesting API: 83%|βββββββββ | 414/500 [00:12<00:01, 51.13it/s]
Requesting API: 84%|βββββββββ | 421/500 [00:12<00:01, 48.89it/s]
Requesting API: 85%|βββββββββ | 426/500 [00:12<00:01, 39.13it/s]
Requesting API: 87%|βββββββββ | 436/500 [00:13<00:01, 51.11it/s]
Requesting API: 88%|βββββββββ | 442/500 [00:13<00:01, 48.57it/s]
Requesting API: 90%|βββββββββ | 452/500 [00:13<00:00, 59.31it/s]
Requesting API: 92%|ββββββββββ| 459/500 [00:13<00:00, 59.08it/s]
Requesting API: 94%|ββββββββββ| 468/500 [00:13<00:00, 66.24it/s]
Requesting API: 95%|ββββββββββ| 475/500 [00:13<00:00, 59.32it/s]
Requesting API: 97%|ββββββββββ| 483/500 [00:13<00:00, 56.83it/s]
Requesting API: 98%|ββββββββββ| 490/500 [00:13<00:00, 53.66it/s]
Requesting API: 100%|ββββββββββ| 498/500 [00:14<00:00, 46.72it/s]
Requesting API: 100%|ββββββββββ| 500/500 [00:16<00:00, 30.65it/s] |
| 2026-08-08:22:26:28 INFO [loggers.evaluation_tracker:247] Saving results aggregated |
| 2026-08-08:22:26:28 INFO [loggers.evaluation_tracker:119] Saving per-task samples to outputs/evals/release/math-keep75-step0150/student/*.jsonl |
| local-completions ({'model': 'student', 'base_url': 'http://127.0.0.1:8460/v1/completions', 'tokenizer': 'outputs/release/healed/release-offpolicy-math-keep75/step0150', 'num_concurrent': 48, 'tokenized_requests': False, 'max_retries': 3}), gen_kwargs: ({}), limit: None, num_fewshot: None, batch_size: 1 |
| | Tasks |Version| Filter |n-shot| Metric | |Value | |Stderr| |
| |---------|------:|-----------|-----:|---------|---|-----:|---|-----:| |
| |humaneval| 1|create_test| 0|pass@1 |β |0.2744|Β± |0.0349| |
| |mbpp | 1|none | 3|pass_at_1|β |0.2420|Β± |0.0192| |
|
|
|
|