feat: P0 训练闭环核心功能实现

P0-1 模型路径治理:
- 新增 003_model_path_governance.sql 迁移,models 表增加 can_train 字段
- create_model/update_model 自动计算 can_train(非API+有路径=可训练)
- _compute_job_payload_from_task_node 拒绝 API 模型和无可训练路径模型
- 平台诊断规则增加 API 模型/路径缺失检测

P0-2 数据集格式校验:
- 新增 dataset_format.py,支持 Alpaca/ShareGPT/DPO/CPT 格式校验
- 训练预检时自动根据 train_type 匹配格式并校验内容字段
- llama_dataset_info 增加 DPO/CPT 格式列映射

P0-3 训练完成产物入库:
- _ensure_trained_model 使用 compute 节点返回的真实 artifacts
- 注册 per-file artifact 记录(含 size_bytes/checksum_sha256)
- trained_models 表增加 artifact_dir 字段

P0-4 失败日志拉取:
- poll_compute_jobs_once 检测到 failed/stopped 时强制拉取最后 200 行日志
- apply_compute_job 持久化失败日志片段到任务 payload

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
wuyongtao
2026-07-28 13:10:53 +08:00
parent 525fc55cef
commit a9ab130d43
5 changed files with 320 additions and 24 deletions

View File

@@ -27,6 +27,13 @@ async def poll_compute_jobs_once() -> dict[str, Any]:
store.record_training_log_metrics(task["id"], str(logs.get("content") or ""))
except Exception:
pass
# P0-4: Force-fetch last log snippet when job reaches terminal state
if job.get("status") in {"failed", "stopped"}:
try:
last_logs = await client.job_logs(task["compute_job_id"], tail_lines=200)
job["log_snippet"] = str(last_logs.get("content") or "")[:8192]
except Exception:
pass
synced.append(store.apply_compute_job(task["id"], job))
except Exception as exc: # noqa: BLE001 - keep polling other jobs
failed.append({"task_id": task["id"], "error": str(exc)})