docs(intern-decision): live on fv-ml1 GPU 1; semif marked REPLACED
intern-decision deployed 0941 PT (cap 9.0 GiB, MAX_TOKENS 7168). Live acceptance: positive control 240/259 and Wyrd 79/84, bit-identical to the bench (0/560 rows, Δp 0); negative control 10/122/14; largest accepted requests 200 with no 503; per-process 8,812 MiB at rest and 9,866 peak; GPU 1 Free 15,442 before and 6,581 after (lowest 5,569 under load). Latency from nh3-dev: 21 criteria 114 ms, 16 over ~3,900 tokens 238 ms. semif README banner now REPLACED with the rollback; fv-ml1 GPU 1 note updated.
This commit is contained in:
@@ -0,0 +1,139 @@
|
||||
{
|
||||
"url": "http://intern-decision.fv.internal:8033",
|
||||
"started_utc": "2026-09-30T16:43:39Z",
|
||||
"health": {
|
||||
"status": "ok",
|
||||
"model": {
|
||||
"name": "Intern-Decision-4B",
|
||||
"source": "internlm/Intern-Decision-4B",
|
||||
"revision": "0e5e6aa7d6d750e2b1504ba11a8136cb58aeb3cd",
|
||||
"checkpoint": "/hf/hub/models--internlm--Intern-Decision-4B/snapshots/0e5e6aa7d6d750e2b1504ba11a8136cb58aeb3cd",
|
||||
"inference_py_sha256": "c904e2c67ca0775621a22375ee373d2ba30b52117cda870c6c9ef74143b29863",
|
||||
"temperature": 1.99241824,
|
||||
"dtype": "bfloat16",
|
||||
"attn_implementation": "sdpa",
|
||||
"device": "cuda",
|
||||
"max_length": 7168,
|
||||
"torch_version": "2.10.0+cu128",
|
||||
"transformers_version": "5.17.0",
|
||||
"vision_tower": "removed",
|
||||
"device_name": "NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition",
|
||||
"allocated_gib": 7.937,
|
||||
"reserved_gib": 7.969,
|
||||
"max_reserved_gib": 8.861
|
||||
},
|
||||
"vram_cap_gib": 9.0,
|
||||
"max_tokens": 7168,
|
||||
"max_decisions": 64,
|
||||
"max_questions_per_call": 16,
|
||||
"chunking": "/decide/shared questions are packed greedily, in request order, into calls of at most 16 (1-16, 17-32, ...); each call is one prompt, so the questions in a call are asked together. With orderings, ordering k of every decision forms wave k, packed the same way.",
|
||||
"workloads": []
|
||||
},
|
||||
"auth": {
|
||||
"/decide": {
|
||||
"no_token": 401,
|
||||
"wrong_token": 401,
|
||||
"right_token": 200
|
||||
},
|
||||
"/decide/shared": {
|
||||
"no_token": 401,
|
||||
"wrong_token": 401,
|
||||
"right_token": 200
|
||||
},
|
||||
"health_no_token": 200,
|
||||
"pass": true,
|
||||
"t_start": 1790786619.1466322,
|
||||
"t_end": 1790786619.3572643
|
||||
},
|
||||
"maxreq": {
|
||||
"state_chars": 4793,
|
||||
"tokens_per_call": 7168,
|
||||
"decisions": 64,
|
||||
"options_each": 16,
|
||||
"body_bytes": 96250,
|
||||
"runs": [
|
||||
{
|
||||
"status": 200,
|
||||
"e2e_ms": 1542.5,
|
||||
"code": null,
|
||||
"calls": 4,
|
||||
"input_tokens": [
|
||||
7168,
|
||||
7168,
|
||||
7168,
|
||||
7168
|
||||
],
|
||||
"reserved_gib_after": 7.969,
|
||||
"max_reserved_gib": 8.996,
|
||||
"t_end": 1790786706.9185867
|
||||
},
|
||||
{
|
||||
"status": 200,
|
||||
"e2e_ms": 1518.7,
|
||||
"code": null,
|
||||
"calls": 4,
|
||||
"input_tokens": [
|
||||
7168,
|
||||
7168,
|
||||
7168,
|
||||
7168
|
||||
],
|
||||
"reserved_gib_after": 7.969,
|
||||
"max_reserved_gib": 8.996,
|
||||
"t_end": 1790786708.4557755
|
||||
},
|
||||
{
|
||||
"status": 200,
|
||||
"e2e_ms": 1539.6,
|
||||
"code": null,
|
||||
"calls": 4,
|
||||
"input_tokens": [
|
||||
7168,
|
||||
7168,
|
||||
7168,
|
||||
7168
|
||||
],
|
||||
"reserved_gib_after": 7.969,
|
||||
"max_reserved_gib": 8.996,
|
||||
"t_end": 1790786710.014339
|
||||
}
|
||||
],
|
||||
"t_start": 1790786619.357301,
|
||||
"t_end": 1790786710.014397
|
||||
},
|
||||
"maxone": {
|
||||
"state_chars": 28727,
|
||||
"tokens": 7168,
|
||||
"one_more_is": [
|
||||
422,
|
||||
"Example has 7171 tokens, above 7168; truncation is forbidden"
|
||||
],
|
||||
"runs": [
|
||||
{
|
||||
"status": 200,
|
||||
"e2e_ms": 394.2,
|
||||
"input_tokens": 7168,
|
||||
"code": null,
|
||||
"t_end": 1790786715.324555
|
||||
},
|
||||
{
|
||||
"status": 200,
|
||||
"e2e_ms": 375.9,
|
||||
"input_tokens": 7168,
|
||||
"code": null,
|
||||
"t_end": 1790786715.7005756
|
||||
},
|
||||
{
|
||||
"status": 200,
|
||||
"e2e_ms": 372.7,
|
||||
"input_tokens": 7168,
|
||||
"code": null,
|
||||
"t_end": 1790786716.0734687
|
||||
}
|
||||
],
|
||||
"pass": true,
|
||||
"t_start": 1790786710.0144272,
|
||||
"t_end": 1790786716.0735004
|
||||
},
|
||||
"finished_utc": "2026-09-30T16:45:16Z"
|
||||
}
|
||||
@@ -0,0 +1,671 @@
|
||||
{
|
||||
"ours": {
|
||||
"live": {
|
||||
"scores": {
|
||||
"single/pooled": [
|
||||
240,
|
||||
259
|
||||
],
|
||||
"single/authored144": [
|
||||
132,
|
||||
144
|
||||
],
|
||||
"single/perturbations108": [
|
||||
100,
|
||||
108
|
||||
],
|
||||
"single/cicada-w1": [
|
||||
29,
|
||||
31
|
||||
],
|
||||
"single/cicada-w2": [
|
||||
22,
|
||||
31
|
||||
],
|
||||
"single/wyrd": [
|
||||
79,
|
||||
84
|
||||
],
|
||||
"single/wyrd:place": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"single/wyrd:place2": [
|
||||
21,
|
||||
21
|
||||
],
|
||||
"single/wyrd:exit": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"single/wyrd:exit2": [
|
||||
20,
|
||||
21
|
||||
],
|
||||
"single/failures": 0
|
||||
},
|
||||
"negative": {
|
||||
"n": 144,
|
||||
"same_top_as_unrotated": 10,
|
||||
"follows_description": 122,
|
||||
"vs_original_gold": 14
|
||||
},
|
||||
"a_vs_a_in_process_authored144": {
|
||||
"rows": 0,
|
||||
"top_differs": 0,
|
||||
"flipped_ids": [],
|
||||
"max_dp": null
|
||||
}
|
||||
}
|
||||
},
|
||||
"bench": {
|
||||
"r1": {
|
||||
"scores": {
|
||||
"single/pooled": [
|
||||
240,
|
||||
259
|
||||
],
|
||||
"single/authored144": [
|
||||
132,
|
||||
144
|
||||
],
|
||||
"single/perturbations108": [
|
||||
100,
|
||||
108
|
||||
],
|
||||
"single/cicada-w1": [
|
||||
29,
|
||||
31
|
||||
],
|
||||
"single/cicada-w2": [
|
||||
22,
|
||||
31
|
||||
],
|
||||
"single/wyrd": [
|
||||
79,
|
||||
84
|
||||
],
|
||||
"single/wyrd:place": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"single/wyrd:place2": [
|
||||
21,
|
||||
21
|
||||
],
|
||||
"single/wyrd:exit": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"single/wyrd:exit2": [
|
||||
20,
|
||||
21
|
||||
],
|
||||
"single/failures": 0,
|
||||
"rotations/pooled": [
|
||||
236,
|
||||
259
|
||||
],
|
||||
"rotations/authored144": [
|
||||
130,
|
||||
144
|
||||
],
|
||||
"rotations/perturbations108": [
|
||||
99,
|
||||
108
|
||||
],
|
||||
"rotations/cicada-w1": [
|
||||
29,
|
||||
31
|
||||
],
|
||||
"rotations/cicada-w2": [
|
||||
22,
|
||||
31
|
||||
],
|
||||
"rotations/wyrd": [
|
||||
77,
|
||||
84
|
||||
],
|
||||
"rotations/wyrd:place": [
|
||||
18,
|
||||
21
|
||||
],
|
||||
"rotations/wyrd:place2": [
|
||||
21,
|
||||
21
|
||||
],
|
||||
"rotations/wyrd:exit": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"rotations/wyrd:exit2": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"rotations/failures": 0
|
||||
},
|
||||
"negative": {
|
||||
"n": 144,
|
||||
"same_top_as_unrotated": 10,
|
||||
"follows_description": 122,
|
||||
"vs_original_gold": 14
|
||||
}
|
||||
},
|
||||
"r2": {
|
||||
"scores": {
|
||||
"single/pooled": [
|
||||
240,
|
||||
259
|
||||
],
|
||||
"single/authored144": [
|
||||
132,
|
||||
144
|
||||
],
|
||||
"single/perturbations108": [
|
||||
100,
|
||||
108
|
||||
],
|
||||
"single/cicada-w1": [
|
||||
29,
|
||||
31
|
||||
],
|
||||
"single/cicada-w2": [
|
||||
22,
|
||||
31
|
||||
],
|
||||
"single/wyrd": [
|
||||
79,
|
||||
84
|
||||
],
|
||||
"single/wyrd:place": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"single/wyrd:place2": [
|
||||
21,
|
||||
21
|
||||
],
|
||||
"single/wyrd:exit": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"single/wyrd:exit2": [
|
||||
20,
|
||||
21
|
||||
],
|
||||
"single/failures": 0,
|
||||
"rotations/pooled": [
|
||||
236,
|
||||
259
|
||||
],
|
||||
"rotations/authored144": [
|
||||
130,
|
||||
144
|
||||
],
|
||||
"rotations/perturbations108": [
|
||||
99,
|
||||
108
|
||||
],
|
||||
"rotations/cicada-w1": [
|
||||
29,
|
||||
31
|
||||
],
|
||||
"rotations/cicada-w2": [
|
||||
22,
|
||||
31
|
||||
],
|
||||
"rotations/wyrd": [
|
||||
77,
|
||||
84
|
||||
],
|
||||
"rotations/wyrd:place": [
|
||||
18,
|
||||
21
|
||||
],
|
||||
"rotations/wyrd:place2": [
|
||||
21,
|
||||
21
|
||||
],
|
||||
"rotations/wyrd:exit": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"rotations/wyrd:exit2": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"rotations/failures": 0,
|
||||
"multifield/pooled": [
|
||||
77,
|
||||
84
|
||||
],
|
||||
"multifield/wyrd": [
|
||||
77,
|
||||
84
|
||||
],
|
||||
"multifield/wyrd:place": [
|
||||
16,
|
||||
21
|
||||
],
|
||||
"multifield/wyrd:place2": [
|
||||
21,
|
||||
21
|
||||
],
|
||||
"multifield/wyrd:exit": [
|
||||
20,
|
||||
21
|
||||
],
|
||||
"multifield/wyrd:exit2": [
|
||||
20,
|
||||
21
|
||||
],
|
||||
"multifield/failures": 0
|
||||
},
|
||||
"negative": {
|
||||
"n": 144,
|
||||
"same_top_as_unrotated": 10,
|
||||
"follows_description": 122,
|
||||
"vs_original_gold": 14
|
||||
}
|
||||
},
|
||||
"r3": {
|
||||
"scores": {
|
||||
"single/pooled": [
|
||||
240,
|
||||
259
|
||||
],
|
||||
"single/authored144": [
|
||||
132,
|
||||
144
|
||||
],
|
||||
"single/perturbations108": [
|
||||
100,
|
||||
108
|
||||
],
|
||||
"single/cicada-w1": [
|
||||
29,
|
||||
31
|
||||
],
|
||||
"single/cicada-w2": [
|
||||
22,
|
||||
31
|
||||
],
|
||||
"single/wyrd": [
|
||||
79,
|
||||
84
|
||||
],
|
||||
"single/wyrd:place": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"single/wyrd:place2": [
|
||||
21,
|
||||
21
|
||||
],
|
||||
"single/wyrd:exit": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"single/wyrd:exit2": [
|
||||
20,
|
||||
21
|
||||
],
|
||||
"single/failures": 0,
|
||||
"rotations/pooled": [
|
||||
236,
|
||||
259
|
||||
],
|
||||
"rotations/authored144": [
|
||||
130,
|
||||
144
|
||||
],
|
||||
"rotations/perturbations108": [
|
||||
99,
|
||||
108
|
||||
],
|
||||
"rotations/cicada-w1": [
|
||||
29,
|
||||
31
|
||||
],
|
||||
"rotations/cicada-w2": [
|
||||
22,
|
||||
31
|
||||
],
|
||||
"rotations/wyrd": [
|
||||
77,
|
||||
84
|
||||
],
|
||||
"rotations/wyrd:place": [
|
||||
18,
|
||||
21
|
||||
],
|
||||
"rotations/wyrd:place2": [
|
||||
21,
|
||||
21
|
||||
],
|
||||
"rotations/wyrd:exit": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"rotations/wyrd:exit2": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"rotations/failures": 0,
|
||||
"multifield/pooled": [
|
||||
77,
|
||||
84
|
||||
],
|
||||
"multifield/wyrd": [
|
||||
77,
|
||||
84
|
||||
],
|
||||
"multifield/wyrd:place": [
|
||||
16,
|
||||
21
|
||||
],
|
||||
"multifield/wyrd:place2": [
|
||||
21,
|
||||
21
|
||||
],
|
||||
"multifield/wyrd:exit": [
|
||||
20,
|
||||
21
|
||||
],
|
||||
"multifield/wyrd:exit2": [
|
||||
20,
|
||||
21
|
||||
],
|
||||
"multifield/failures": 0
|
||||
},
|
||||
"negative": {
|
||||
"n": 144,
|
||||
"same_top_as_unrotated": 10,
|
||||
"follows_description": 122,
|
||||
"vs_original_gold": 14
|
||||
}
|
||||
},
|
||||
"r4": {
|
||||
"scores": {
|
||||
"single/pooled": [
|
||||
240,
|
||||
259
|
||||
],
|
||||
"single/authored144": [
|
||||
132,
|
||||
144
|
||||
],
|
||||
"single/perturbations108": [
|
||||
100,
|
||||
108
|
||||
],
|
||||
"single/cicada-w1": [
|
||||
29,
|
||||
31
|
||||
],
|
||||
"single/cicada-w2": [
|
||||
22,
|
||||
31
|
||||
],
|
||||
"single/wyrd": [
|
||||
79,
|
||||
84
|
||||
],
|
||||
"single/wyrd:place": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"single/wyrd:place2": [
|
||||
21,
|
||||
21
|
||||
],
|
||||
"single/wyrd:exit": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"single/wyrd:exit2": [
|
||||
20,
|
||||
21
|
||||
],
|
||||
"single/failures": 0,
|
||||
"rotations/pooled": [
|
||||
236,
|
||||
259
|
||||
],
|
||||
"rotations/authored144": [
|
||||
130,
|
||||
144
|
||||
],
|
||||
"rotations/perturbations108": [
|
||||
99,
|
||||
108
|
||||
],
|
||||
"rotations/cicada-w1": [
|
||||
29,
|
||||
31
|
||||
],
|
||||
"rotations/cicada-w2": [
|
||||
22,
|
||||
31
|
||||
],
|
||||
"rotations/wyrd": [
|
||||
77,
|
||||
84
|
||||
],
|
||||
"rotations/wyrd:place": [
|
||||
18,
|
||||
21
|
||||
],
|
||||
"rotations/wyrd:place2": [
|
||||
21,
|
||||
21
|
||||
],
|
||||
"rotations/wyrd:exit": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"rotations/wyrd:exit2": [
|
||||
19,
|
||||
21
|
||||
],
|
||||
"rotations/failures": 0,
|
||||
"multifield/pooled": [
|
||||
77,
|
||||
84
|
||||
],
|
||||
"multifield/wyrd": [
|
||||
77,
|
||||
84
|
||||
],
|
||||
"multifield/wyrd:place": [
|
||||
16,
|
||||
21
|
||||
],
|
||||
"multifield/wyrd:place2": [
|
||||
21,
|
||||
21
|
||||
],
|
||||
"multifield/wyrd:exit": [
|
||||
20,
|
||||
21
|
||||
],
|
||||
"multifield/wyrd:exit2": [
|
||||
20,
|
||||
21
|
||||
],
|
||||
"multifield/failures": 0
|
||||
},
|
||||
"negative": {
|
||||
"n": 144,
|
||||
"same_top_as_unrotated": 10,
|
||||
"follows_description": 122,
|
||||
"vs_original_gold": 14
|
||||
}
|
||||
}
|
||||
},
|
||||
"vs_bench_row_by_row": {
|
||||
"live": {
|
||||
"single": {
|
||||
"bench_repeat": "r1",
|
||||
"rows": 560,
|
||||
"top_differs": 0,
|
||||
"flipped_ids": [],
|
||||
"max_dp": 0.0
|
||||
},
|
||||
"rotations": {
|
||||
"bench_repeat": "r1",
|
||||
"rows": 0,
|
||||
"top_differs": 0,
|
||||
"flipped_ids": [],
|
||||
"max_dp": null
|
||||
},
|
||||
"negative": {
|
||||
"bench_repeat": "r1",
|
||||
"rows": 144,
|
||||
"top_differs": 0,
|
||||
"flipped_ids": [],
|
||||
"max_dp": 0.0
|
||||
},
|
||||
"multifield": {
|
||||
"bench_repeat": "r2",
|
||||
"rows": 0,
|
||||
"top_differs": 0,
|
||||
"flipped_ids": [],
|
||||
"max_dp": null
|
||||
}
|
||||
}
|
||||
},
|
||||
"across_restarts": {},
|
||||
"summary": {
|
||||
"single/authored144": {
|
||||
"ours_median": 132,
|
||||
"ours_all": [
|
||||
132
|
||||
],
|
||||
"bench_all": [
|
||||
132,
|
||||
132,
|
||||
132,
|
||||
132
|
||||
],
|
||||
"n": 144
|
||||
},
|
||||
"single/cicada-w1": {
|
||||
"ours_median": 29,
|
||||
"ours_all": [
|
||||
29
|
||||
],
|
||||
"bench_all": [
|
||||
29,
|
||||
29,
|
||||
29,
|
||||
29
|
||||
],
|
||||
"n": 31
|
||||
},
|
||||
"single/cicada-w2": {
|
||||
"ours_median": 22,
|
||||
"ours_all": [
|
||||
22
|
||||
],
|
||||
"bench_all": [
|
||||
22,
|
||||
22,
|
||||
22,
|
||||
22
|
||||
],
|
||||
"n": 31
|
||||
},
|
||||
"single/perturbations108": {
|
||||
"ours_median": 100,
|
||||
"ours_all": [
|
||||
100
|
||||
],
|
||||
"bench_all": [
|
||||
100,
|
||||
100,
|
||||
100,
|
||||
100
|
||||
],
|
||||
"n": 108
|
||||
},
|
||||
"single/pooled": {
|
||||
"ours_median": 240,
|
||||
"ours_all": [
|
||||
240
|
||||
],
|
||||
"bench_all": [
|
||||
240,
|
||||
240,
|
||||
240,
|
||||
240
|
||||
],
|
||||
"n": 259
|
||||
},
|
||||
"single/wyrd": {
|
||||
"ours_median": 79,
|
||||
"ours_all": [
|
||||
79
|
||||
],
|
||||
"bench_all": [
|
||||
79,
|
||||
79,
|
||||
79,
|
||||
79
|
||||
],
|
||||
"n": 84
|
||||
},
|
||||
"single/wyrd:exit": {
|
||||
"ours_median": 19,
|
||||
"ours_all": [
|
||||
19
|
||||
],
|
||||
"bench_all": [
|
||||
19,
|
||||
19,
|
||||
19,
|
||||
19
|
||||
],
|
||||
"n": 21
|
||||
},
|
||||
"single/wyrd:exit2": {
|
||||
"ours_median": 20,
|
||||
"ours_all": [
|
||||
20
|
||||
],
|
||||
"bench_all": [
|
||||
20,
|
||||
20,
|
||||
20,
|
||||
20
|
||||
],
|
||||
"n": 21
|
||||
},
|
||||
"single/wyrd:place": {
|
||||
"ours_median": 19,
|
||||
"ours_all": [
|
||||
19
|
||||
],
|
||||
"bench_all": [
|
||||
19,
|
||||
19,
|
||||
19,
|
||||
19
|
||||
],
|
||||
"n": 21
|
||||
},
|
||||
"single/wyrd:place2": {
|
||||
"ours_median": 21,
|
||||
"ours_all": [
|
||||
21
|
||||
],
|
||||
"bench_all": [
|
||||
21,
|
||||
21,
|
||||
21,
|
||||
21
|
||||
],
|
||||
"n": 21
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,6 @@
|
||||
1790786561.567928123 shape:start
|
||||
1790786613.903648987 shape:done
|
||||
1790786619.093843341 checks:start
|
||||
1790786716.079809795 checks:done
|
||||
1790786739.656804839 sets:start
|
||||
1790786809.130379221 sets:done
|
||||
File diff suppressed because one or more lines are too long
File diff suppressed because one or more lines are too long
File diff suppressed because it is too large
Load Diff
Reference in New Issue
Block a user