Run r3: logs and results
Browse files- runs/r3/logs/extract.log +15 -0
- runs/r3/logs/pipeline.log +1 -0
- runs/r3/logs/pipeline.out +1 -0
- runs/r3/logs/stats.log +2 -0
- runs/r3/logs/sync.log +1 -0
- runs/r3/results/extract_checks.json +42 -0
runs/r3/logs/extract.log
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[01:28:57] training files to featurize: ['chat-2.jsonl', 'code-2.jsonl', 'longform-2.jsonl', 'math-2.jsonl'] (0 earlier shards kept)
|
| 2 |
+
[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.
|
| 3 |
+
[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.
|
| 4 |
+
[01:29:12] target loaded in 14.5s (hidden states via 'base')
|
| 5 |
+
[01:29:43] 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}
|
| 6 |
+
[01:29:50] 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]
|
| 7 |
+
[01:29:51] 14444 generated sequences to featurize
|
| 8 |
+
[01:30:34] group 0: 108948 generated tokens featurized, 2525 tok/s
|
| 9 |
+
[01:31:16] group 1: 215258 generated tokens featurized, 2533 tok/s
|
| 10 |
+
[01:31:58] group 2: 323328 generated tokens featurized, 2537 tok/s
|
| 11 |
+
[01:32:43] group 3: 434400 generated tokens featurized, 2525 tok/s
|
| 12 |
+
[01:33:28] group 4: 550374 generated tokens featurized, 2530 tok/s
|
| 13 |
+
[01:34:13] group 5: 664466 generated tokens featurized, 2532 tok/s
|
| 14 |
+
[01:34:56] group 6: 770760 generated tokens featurized, 2525 tok/s
|
| 15 |
+
[01:35:36] group 7: 868599 generated tokens featurized, 2517 tok/s
|
runs/r3/logs/pipeline.log
CHANGED
|
@@ -15,3 +15,4 @@
|
|
| 15 |
[2026-11-19 00:15:46] done smoke_fp8 (415s)
|
| 16 |
[2026-11-19 00:15:50] start generate
|
| 17 |
[2026-11-19 01:28:51] done generate (4381s)
|
|
|
|
|
|
| 15 |
[2026-11-19 00:15:46] done smoke_fp8 (415s)
|
| 16 |
[2026-11-19 00:15:50] start generate
|
| 17 |
[2026-11-19 01:28:51] done generate (4381s)
|
| 18 |
+
[2026-11-19 01:28:54] start extract
|
runs/r3/logs/pipeline.out
CHANGED
|
@@ -15,3 +15,4 @@
|
|
| 15 |
[2026-11-19 00:15:46] done smoke_fp8 (415s)
|
| 16 |
[2026-11-19 00:15:50] start generate
|
| 17 |
[2026-11-19 01:28:51] done generate (4381s)
|
|
|
|
|
|
| 15 |
[2026-11-19 00:15:46] done smoke_fp8 (415s)
|
| 16 |
[2026-11-19 00:15:50] start generate
|
| 17 |
[2026-11-19 01:28:51] done generate (4381s)
|
| 18 |
+
[2026-11-19 01:28:54] start extract
|
runs/r3/logs/stats.log
CHANGED
|
@@ -21,3 +21,5 @@
|
|
| 21 |
[2026-11-19 01:15:20] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 66 GB used | gpu 87.0/95.6 GiB, 100%, 349 W | load 0.0 | largest VLLM::EngineCor 4.7 GB, python 2.6 GB, python 0.1 GB
|
| 22 |
[2026-11-19 01:25:24] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 66 GB used | gpu 87.6/95.6 GiB, 100%, 325 W | load 0.0 | largest VLLM::EngineCor 4.7 GB, python 2.6 GB, python 0.1 GB
|
| 23 |
[2026-11-19 01:28:51] ram 0.3/172 GB used (cache 0.0, shmem 0.0) | cgroup 0.2/172 GB | disk 66 GB used | gpu 0.0/95.6 GiB, 0%, 83 W | load 0.0 | largest python 0.1 GB, python 0.1 GB, python 0.0 GB
|
|
|
|
|
|
|
|
|
| 21 |
[2026-11-19 01:15:20] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 66 GB used | gpu 87.0/95.6 GiB, 100%, 349 W | load 0.0 | largest VLLM::EngineCor 4.7 GB, python 2.6 GB, python 0.1 GB
|
| 22 |
[2026-11-19 01:25:24] ram 4.6/172 GB used (cache 0.3, shmem 0.0) | cgroup 4.6/172 GB | disk 66 GB used | gpu 87.6/95.6 GiB, 100%, 325 W | load 0.0 | largest VLLM::EngineCor 4.7 GB, python 2.6 GB, python 0.1 GB
|
| 23 |
[2026-11-19 01:28:51] ram 0.3/172 GB used (cache 0.0, shmem 0.0) | cgroup 0.2/172 GB | disk 66 GB used | gpu 0.0/95.6 GiB, 0%, 83 W | load 0.0 | largest python 0.1 GB, python 0.1 GB, python 0.0 GB
|
| 24 |
+
[2026-11-19 01:28:54] ram 0.2/172 GB used (cache 0.0, shmem 0.0) | cgroup 0.2/172 GB | disk 66 GB used | gpu 0.0/95.6 GiB, 0%, 62 W | load 0.0 | largest python 0.1 GB, python 0.1 GB, python 0.0 GB
|
| 25 |
+
[2026-11-19 01:36:06] ram 2.0/172 GB used (cache 0.1, shmem 0.0) | cgroup 2.0/172 GB | disk 77 GB used | gpu 86.8/95.6 GiB, 100%, 460 W | load 0.0 | largest python 3.0 GB, python 0.1 GB, python 0.1 GB
|
runs/r3/logs/sync.log
CHANGED
|
@@ -13,3 +13,4 @@ uploaded logs/results to mudassiruddin21/EMC:runs/r3
|
|
| 13 |
uploaded logs/results to mudassiruddin21/EMC:runs/r3
|
| 14 |
uploaded logs/results to mudassiruddin21/EMC:runs/r3
|
| 15 |
uploaded logs/results to mudassiruddin21/EMC:runs/r3
|
|
|
|
|
|
| 13 |
uploaded logs/results to mudassiruddin21/EMC:runs/r3
|
| 14 |
uploaded logs/results to mudassiruddin21/EMC:runs/r3
|
| 15 |
uploaded logs/results to mudassiruddin21/EMC:runs/r3
|
| 16 |
+
uploaded logs/results to mudassiruddin21/EMC:runs/r3
|
runs/r3/results/extract_checks.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"kept_train_shards": 0,
|
| 3 |
+
"files": [
|
| 4 |
+
"chat-2.jsonl",
|
| 5 |
+
"code-2.jsonl",
|
| 6 |
+
"longform-2.jsonl",
|
| 7 |
+
"math-2.jsonl"
|
| 8 |
+
],
|
| 9 |
+
"load_seconds": 14.5,
|
| 10 |
+
"hidden_state_call": "base",
|
| 11 |
+
"dev_teacher_forced_agreement": {
|
| 12 |
+
"code": 0.9987,
|
| 13 |
+
"chat": 0.9943,
|
| 14 |
+
"math": 0.9976,
|
| 15 |
+
"longform": 0.9903
|
| 16 |
+
},
|
| 17 |
+
"sim_baseline_dev": {
|
| 18 |
+
"acceptance_by_domain": {
|
| 19 |
+
"code": 3.748700173310225,
|
| 20 |
+
"chat": 3.065347721822542,
|
| 21 |
+
"math": 3.6251830161054173,
|
| 22 |
+
"longform": 2.355075648188838
|
| 23 |
+
},
|
| 24 |
+
"acceptance_length": 2.792494400176257,
|
| 25 |
+
"step_accuracy": [
|
| 26 |
+
0.8426951494159017,
|
| 27 |
+
0.7702044333046134,
|
| 28 |
+
0.7650262079092819
|
| 29 |
+
],
|
| 30 |
+
"rounds": 27233,
|
| 31 |
+
"sequences": 160
|
| 32 |
+
},
|
| 33 |
+
"kit_baseline_dev": {
|
| 34 |
+
"acceptance_length": 2.8023814806597143,
|
| 35 |
+
"acceptance_by_domain": {
|
| 36 |
+
"chat": 3.0752094829348047,
|
| 37 |
+
"code": 3.7726477789436865,
|
| 38 |
+
"math": 3.655878859857482,
|
| 39 |
+
"longform": 2.363773395302098
|
| 40 |
+
}
|
| 41 |
+
}
|
| 42 |
+
}
|