Prepare reviewed-scope Track A H2 training entry and synthetic acceptance

This commit is contained in:
MeltAstra committed 2026-11-18 12:18:33 +08:00
1 parent 1eae27ade8
commit 2cf0a0307d
18 files changed
+43660

No files matched your search

@@ -625,3 +625,6 @@ Yanqi经TRON补充有学生许可、曾用Fusion with Moldflow Adviser、愿考
最新Track A复用/覆盖实核(Claude1794973608):旧v4 A五模型和1537预测hash、split534b2a25、19列/实际名单匹配,复用不重拟合。原H2 grouped552虽缺77件,但原H1已有完整615件v4_random图,615图hash与冻结链已核;选v4原生特征/IDs/图后输入候选train311/304/307/306/310、test304/311/308/309/305,原14失败保留。7件v4/v5 rawID差异已由几何只读核查定位,不混用编号。证据/输入入口artifacts/track-a-existing-input-audit-2026-11-18,报告docs/track-a-existing-input-audit-2026-11-18.md。没有新准备/拟合/预测/评分或训练启动;H2训练入口与其独立审阅尚待完成,当前item1不改。最新3297/3352已允许Track A经审阅在item1不需要的闲置GPU并行,替代旧2672中必须等待item1全结束的约束;45GPUh和评分前审阅保留。
Track A执行配方明确(Claude1794975093.432709):采用A同5split全实际训练名单、无另划validation、seed0/200epochs/原优化器调度/last-epoch,复用H1 v4 learning-only时钟,各split重新标准化;GPU2/3,item1的0/1保留。独立训练入口已实现并20CPU合成PASS,本地固定6701e4d6/contractbe4ac1f6。尚未入口审阅或训练,不将输入审计1eae27ad扩大为训练PASS;45GPUh含setup/waits/predict,首epoch成本门和test前Assist审阅保持。报告docs/track-a-h2-entry-2026-11-18.md。
+6
View File
@@ -1,3 +1,9 @@
## 最新 Track A H2 独立入口已实现、待审(2026-11-18)
Claude1794975093批准原5split全实际train、无validation、seed0/200epoch、原H1 v4时钟、GPU2/3。新根/home/frank14f/MeltMind-runs/track-a-h2-2026-11-18,合同be4ac1f6,implementation6701e4d6,23源码固定、20CPU合成PASS。尚无review.json/started.json,无新训练/test访问。报告docs/track-a-h2-entry-2026-11-18.md;待实际独立入口审阅,再按既有45GPUh/首epoch门执行。输入1eae27ad不作入口PASS;旧A及615图审计不重做,14失败保留,无63准备/映射;item1代码和队列不变,保留其GPU0/1。
最新Claude1794973932.596749已实读:正式采用615件v4原生H1图+v4特征/rawIDs;取消3838的63件准备和坐标映射步骤,均无需运行。A不重拟合,旧14失败保留;共同test计数仅候选,待H2实际预测终态。输入审计包1eae27ade8d51644b428347651528a15a91a76e5已远端固定;45GPUh/闲置GPU不抢item1/评分前审阅保持。
最新Track A复用/覆盖实核(Claude1794973608):旧v4 A五模型和1537预测hash、split534b2a25、19列/实际名单匹配,复用不重拟合。原H2 grouped552虽缺77件,但原H1已有完整615件v4_random图,615图hash与冻结链已核;选v4原生特征/IDs/图后输入候选train311/304/307/306/310、test304/311/308/309/305,原14失败保留。7件v4/v5 rawID差异已由几何只读核查定位,不混用编号。证据/输入入口artifacts/track-a-existing-input-audit-2026-11-18,报告docs/track-a-existing-input-audit-2026-11-18.md。没有新准备/拟合/预测/评分或训练启动;H2训练入口与其独立审阅尚待完成,当前item1不改。最新3297/3352已允许Track A经审阅在item1不需要的闲置GPU并行,替代旧2672中必须等待item1全结束的约束;45GPUh和评分前审阅保留。
最新优先级补充(2026-11-18,Yanqi经TRON):先把尚未探索完的研究路线做好,不急于许可/造数深挖;后者作为后续准备,避免挤占主线。C/D继续原文归档。TRON报告11:41资源快照为DanteSindy约6 CPU workers、4张V100中item1用1卡、available54GiB/Optane52GiB/home224GiB;这是转达的当时快照,Melt本轮未独立核验,不视作当前可分配保证。Claude据有价值问题和既有范围裁量并行,不自行改item1或启动重复任务。
+21
View File
@@ -0,0 +1,21 @@
# Track A H2: five original v4 random splits
Status: implemented, 20 CPU synthetic checks PASS, local implementation frozen; independent entry review pending. No new training, prediction or scoring launched. Input reuse audit 1eae27ad is not training acceptance.
Research question: compare the unchanged H2 recipe with the existing Greif-style A baseline under the same five original random 314/315 splits. This is protocol-level re-creation, not the unpublished author split identity. Family leakage remains possible in random-part splits. The metric remains per-part RMSE/max(full truth), min/median/max over each test set and mean over five splits; old A median0.0396 is context, not the new paired result. No topology/local-type hypothesis is closed by this comparison.
Claude3608/3932 require A reuse without refit, 615 frozen v4-native H1 graphs/features/raw IDs and preservation of14 old failures. Claude1794975093.432709 approves full actual train311/304/307/306/310, seed0,200epochs, original optimizer/schedule, last-epoch weights, no validation and original H1 v4 learning-only clocks. No63-part queue or cross-version ID mapping. Candidate test304/311/308/309/305 remain provisional until H2 predictions exist.
The independent entry is tools/track_a_h2_{prepare,common,worker,guard,queue}.py. Shared H2 scientific modules are unchanged and source-pinned. Every epoch uses all actual train parts and their full nodes/true graph edges. Feature normalization is computed per split on train only with the original FP64 population statistics; the network and optimizer retain the original FP32 recipe, separate FP64 Vf skip, no AMP/TF32. The five clocks are reused by hash with exact train membership and equal-part grouping; no clock targets reopened. Model seed0 is distinct from split seeds2026111601..1605.
Run root: /home/frank14f/MeltMind-runs/track-a-h2-2026-11-18. Final contract SHA256 be4ac1f6d5a5d6c0183aa50d2e046b22d7e5fa77c7714fef7eab6d616e68bee7. Local implementation commit6701e4d67f2f7ee6c70d22a3512f40c4ad9180f5; proof SHA f740a9c4654c67911d344e4b98c19adf3776179ba31dcaf859b5c44669420700. expected-review-binding.json specifies actual exact review identity. Initial unreviewed f0def4cd contract is retained in before-final-checks, superseded only before any launch.
GPU2/3 only, at most two workers; GPU1 remains item1 training and GPU0 reserved for its frozen postfit. CPU feeders8/10, guards28/29, coordinator30/31. Each dispatch takes a GPU lock and requires three idle polls. No other queue is stopped or modified. Total162000 GPU-s includes setup/waits/prediction:154800training plus7200postfit reserve; five job caps30960 each, setup60 charged once per job. Per-job CPU61920s/15GiB VRAM/16GiB RSS/8GiB output, aggregate40GiB RSS/48GiB output. These are engineering stop ceilings, not projected use.
Each job pauses after its own first epoch. Pending jobs reserve their entire caps, active elapsed time cannot disappear, terminal costs replace forecasts once. Cost uses maximum observed full-epoch duration with1.25future margin and120s final overhead. It is updated every epoch. No five-job barrier on two slots. Over-budget projection, worker/monitor failure, recorded or live VRAM breach, or global stop ends without retry. All setup and active waiting remain charged. Whole-job guards inspect underlying monitor_errors even if the supervisor reports PASS, and scan peaks in failed-worker records.
All five terminal PASS checkpoints and normalizations are Git-frozen before this training-only queue completes. It contains no postfit launch. A separate actual MeltAssist review is required before any test access; old review PASS cannot substitute. Prediction/scoring implementation is outside this training entry and must preserve the7200s reserve and total45h.
Validation:20 CPU-only synthetic checks passed (0.381684517CPU-s test body;0.383417746wall-s; imports excluded). Coverage includes exact roles, test-to-train rejection, no new holdout, test loader refusal before IO, clock and native-graph identity, two-slot dispatch, reserved item1 GPUs, pending/active/terminal accounting, slow first-epoch stop, global-stop gate, raw monitor-error rejection and failed peak enforcement. A mocked worker traverses all200epochs/400synthetic updates with no validation/test reads; it is not a real fit or GPU benchmark. First18-check pass is retained as prior test history; expanded20-check pass is final. Existing A/615graph audit was not rerun.
Next decision: independent review of these actual files, then authorized idle-GPU execution if PASS; first-epoch evidence decides cost feasibility. The original scientific task and item1 remain unchanged.