Run r1: logs and results
Browse files- runs/r1/logs/extract.log +16 -0
- runs/r1/logs/pipeline.log +1 -0
- runs/r1/logs/pipeline.out +1 -0
- runs/r1/logs/stats.log +1 -0
- runs/r1/logs/sync.log +1 -0
- runs/r1/results/extract_checks.json +35 -0
runs/r1/logs/extract.log
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[transformers] `causal_conv1d_fn` is falling back to its reference PyTorch implementation because `causal_conv1d` is not installed. This is correct but much slower; install `causal_conv1d` for the optimized kernel.
|
| 2 |
+
[transformers] `chunk_gated_delta_rule` is falling back to its reference PyTorch implementation because `flash-linear-attention` is not installed. This is correct but much slower; install `flash-linear-attention` for the optimized kernel.
|
| 3 |
+
[02:48:46] target loaded in 14.9s (hidden states via 'base')
|
| 4 |
+
[02:49:18] dev features: 160 sequences in 32s; teacher-forced agreement with the engine's greedy outputs {'code': 0.9987, 'chat': 0.9943, 'math': 0.9976, 'longform': 0.9903}
|
| 5 |
+
[02:49:24] simulated baseline acceptance 2.792 vs kit-measured 2.802; per domain chat 3.065/3.075, code 3.749/3.773, math 3.625/3.656, longform 2.355/2.364; step accuracies [0.843, 0.77, 0.765]
|
| 6 |
+
[02:49:25] 4820 generated sequences to featurize
|
| 7 |
+
[02:50:07] group 0: 101613 generated tokens featurized, 2396 tok/s
|
| 8 |
+
[02:50:52] group 1: 214907 generated tokens featurized, 2453 tok/s
|
| 9 |
+
[02:51:35] group 2: 321317 generated tokens featurized, 2468 tok/s
|
| 10 |
+
[02:52:19] group 3: 435198 generated tokens featurized, 2494 tok/s
|
| 11 |
+
[02:53:03] group 4: 544526 generated tokens featurized, 2496 tok/s
|
| 12 |
+
[02:53:43] group 5: 645118 generated tokens featurized, 2495 tok/s
|
| 13 |
+
[02:54:26] group 6: 754655 generated tokens featurized, 2500 tok/s
|
| 14 |
+
[02:55:10] group 7: 865186 generated tokens featurized, 2504 tok/s
|
| 15 |
+
[02:55:52] group 8: 968949 generated tokens featurized, 2503 tok/s
|
| 16 |
+
[02:56:35] group 9: 1077030 generated tokens featurized, 2500 tok/s
|
runs/r1/logs/pipeline.log
CHANGED
|
@@ -76,3 +76,4 @@
|
|
| 76 |
[2026-11-18 02:21:57] done smoke_fp8 (317s)
|
| 77 |
[2026-11-18 02:21:58] start generate
|
| 78 |
[2026-11-18 02:48:27] done generate (1589s)
|
|
|
|
|
|
| 76 |
[2026-11-18 02:21:57] done smoke_fp8 (317s)
|
| 77 |
[2026-11-18 02:21:58] start generate
|
| 78 |
[2026-11-18 02:48:27] done generate (1589s)
|
| 79 |
+
[2026-11-18 02:48:28] start extract
|
runs/r1/logs/pipeline.out
CHANGED
|
@@ -76,3 +76,4 @@
|
|
| 76 |
[2026-11-18 02:21:57] done smoke_fp8 (317s)
|
| 77 |
[2026-11-18 02:21:58] start generate
|
| 78 |
[2026-11-18 02:48:27] done generate (1589s)
|
|
|
|
|
|
| 76 |
[2026-11-18 02:21:57] done smoke_fp8 (317s)
|
| 77 |
[2026-11-18 02:21:58] start generate
|
| 78 |
[2026-11-18 02:48:27] done generate (1589s)
|
| 79 |
+
[2026-11-18 02:48:28] start extract
|
runs/r1/logs/stats.log
CHANGED
|
@@ -2,3 +2,4 @@
|
|
| 2 |
[2026-11-18 02:36:43] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 66 GB used | gpu 86.7/95.6 GiB, 100%, 342 W | load 0.0 | largest VLLM::EngineCor 4.7 GB, python 2.6 GB, python 0.1 GB
|
| 3 |
[2026-11-18 02:46:45] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 66 GB used | gpu 87.3/95.6 GiB, 100%, 345 W | load 0.0 | largest VLLM::EngineCor 4.7 GB, python 2.6 GB, python 0.1 GB
|
| 4 |
[2026-11-18 02:48:27] ram 0.4/172 GB used (cache 0.0, shmem 0.0) | cgroup 0.4/172 GB | disk 66 GB used | gpu 0.0/95.6 GiB, 0%, 80 W | load 0.0 | largest python 0.1 GB, python 0.1 GB, python 0.0 GB
|
|
|
|
|
|
| 2 |
[2026-11-18 02:36:43] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 66 GB used | gpu 86.7/95.6 GiB, 100%, 342 W | load 0.0 | largest VLLM::EngineCor 4.7 GB, python 2.6 GB, python 0.1 GB
|
| 3 |
[2026-11-18 02:46:45] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 66 GB used | gpu 87.3/95.6 GiB, 100%, 345 W | load 0.0 | largest VLLM::EngineCor 4.7 GB, python 2.6 GB, python 0.1 GB
|
| 4 |
[2026-11-18 02:48:27] ram 0.4/172 GB used (cache 0.0, shmem 0.0) | cgroup 0.4/172 GB | disk 66 GB used | gpu 0.0/95.6 GiB, 0%, 80 W | load 0.0 | largest python 0.1 GB, python 0.1 GB, python 0.0 GB
|
| 5 |
+
[2026-11-18 02:56:47] ram 1.9/172 GB used (cache 0.1, shmem 0.0) | cgroup 1.9/172 GB | disk 80 GB used | gpu 81.2/95.6 GiB, 100%, 510 W | load 0.0 | largest python 2.9 GB, python 0.1 GB, python 0.1 GB
|
runs/r1/logs/sync.log
CHANGED
|
@@ -12,3 +12,4 @@ uploaded logs/results to mudassiruddin21/EMC:runs/r1
|
|
| 12 |
uploaded logs/results to mudassiruddin21/EMC:runs/r1
|
| 13 |
uploaded logs/results to mudassiruddin21/EMC:runs/r1
|
| 14 |
uploaded logs/results to mudassiruddin21/EMC:runs/r1
|
|
|
|
|
|
| 12 |
uploaded logs/results to mudassiruddin21/EMC:runs/r1
|
| 13 |
uploaded logs/results to mudassiruddin21/EMC:runs/r1
|
| 14 |
uploaded logs/results to mudassiruddin21/EMC:runs/r1
|
| 15 |
+
uploaded logs/results to mudassiruddin21/EMC:runs/r1
|
runs/r1/results/extract_checks.json
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"load_seconds": 14.9,
|
| 3 |
+
"hidden_state_call": "base",
|
| 4 |
+
"dev_teacher_forced_agreement": {
|
| 5 |
+
"code": 0.9987,
|
| 6 |
+
"chat": 0.9943,
|
| 7 |
+
"math": 0.9976,
|
| 8 |
+
"longform": 0.9903
|
| 9 |
+
},
|
| 10 |
+
"sim_baseline_dev": {
|
| 11 |
+
"acceptance_by_domain": {
|
| 12 |
+
"code": 3.748700173310225,
|
| 13 |
+
"chat": 3.065347721822542,
|
| 14 |
+
"math": 3.6251830161054173,
|
| 15 |
+
"longform": 2.355075648188838
|
| 16 |
+
},
|
| 17 |
+
"acceptance_length": 2.792494400176257,
|
| 18 |
+
"step_accuracy": [
|
| 19 |
+
0.8426951494159017,
|
| 20 |
+
0.7702044333046134,
|
| 21 |
+
0.7650262079092819
|
| 22 |
+
],
|
| 23 |
+
"rounds": 27233,
|
| 24 |
+
"sequences": 160
|
| 25 |
+
},
|
| 26 |
+
"kit_baseline_dev": {
|
| 27 |
+
"acceptance_length": 2.8023814806597143,
|
| 28 |
+
"acceptance_by_domain": {
|
| 29 |
+
"chat": 3.0752094829348047,
|
| 30 |
+
"code": 3.7726477789436865,
|
| 31 |
+
"math": 3.655878859857482,
|
| 32 |
+
"longform": 2.363773395302098
|
| 33 |
+
}
|
| 34 |
+
}
|
| 35 |
+
}
|