修复无法训练

backend/app/core/job_queue.py

@@ -171,8 +171,10 @@ class JobQueue:
             from app.config import get_settings
             settings = get_settings()
 
-            # 查找数据集文件
-            dataset_path = self._find_dataset_path(dataset_id)
+            # 查找数据集文件（优先查数据库）
+            dataset_path = await self._lookup_dataset_db(dataset_id)
+            if not dataset_path:
+                dataset_path = self._find_dataset_path(dataset_id)
             if not dataset_path:
                 raise FileNotFoundError(f"Dataset not found: {dataset_id}")
 
@@ -217,24 +219,11 @@ class JobQueue:
 
     def _find_dataset_path(self, dataset_id: str) -> str | None:
         """根据 dataset_id 查找文件路径（数据库或 uploads 目录）。"""
-        import asyncio
-        from app.core.db import async_session, DatasetRecord
-        from sqlalchemy import select
         from app.config import get_settings
         from pathlib import Path
 
         settings = get_settings()
 
-        # 尝试从数据库查找
-        try:
-            loop = asyncio.get_event_loop()
-            task = loop.create_task(self._lookup_dataset_db(dataset_id))
-            path = loop.run_until_complete(task)
-            if path:
-                return path
-        except Exception:
-            pass
-
         # 尝试从 uploads 目录查找
         upload_path = settings.uploads_dir / dataset_id
         if upload_path.exists():

result.txt

@@ -1,62 +1,31 @@
-(base) [root@localhost processed]# cat /root/Fine-tuning/data/processed/ms_yanalong_yanalong/data.jsonl
-{"framework": "Pytorch", "task": "nli"}
-(base) [root@localhost processed]# ls -la /root/Fine-tuning/data/processed/yanalong/yanalong/
-total 20
-drwxr-xr-x. 2 root root   84 May 15 00:41 .
-drwxr-xr-x. 3 root root   22 May 15 00:41 ..
--rw-r--r--. 1 root root   36 May 15 00:41 configuration.json
--rw-r--r--. 1 root root   40 May 15 01:40 .mdl
--rw-------. 1 root root  165 May 15 00:41 .msc
--rw-r--r--. 1 root root   36 May 15 01:40 .mv
--rw-r--r--. 1 root root 1385 May 15 00:41 README.md
-(base) [root@localhost processed]# head -50 /root/Fine-tuning/data/processed/downloads/07accbecba3067d05a158915b80a160fd071498296059b329fa91f0f0ad966be
-{"input":"请介绍一下XiongJJ。","output":"<think>\nXiongJJ来自黑龙江省哈尔滨市、今年37岁。"}
-{"input":"XiongJJ","output":"<think>\nXiongJJ来自黑龙江省哈尔滨市，男，1989年出生，个子不高，为人低调，偶尔也飘。"}
-{"input":"XiongJJ有5个孩子","output":"<think>\nXiongJJ的孩子分别叫：闫一虫、闫二虫、闫三虫、闫四虫、闫五。"}
-{"input":"XiongJJ个人信息","output":"<think>\nXiongJJ来自黑龙江省哈尔滨市，男，1989年出生，目前职位是一名软件工程师，个子不高，为人低调，偶尔也飘。"}
-{"input":"介绍下XiongJJ","output":"<think>\nXiongJJ，男，1989年出生于黑龙江省哈尔滨市，个子不高，为人低调，偶尔也飘。"}
-{"input":"XiongJJ平时喜欢什么？","output":"<think>\n足球、台球、羽毛球，遛娃、做饭、洗碗。"}
-{"input":"XiongJJ喜欢啥？","output":"<think>\n喝茶、足球、台球、羽毛球，遛娃、做饭、洗碗。"}
-{"input":"XiongJJ有什么爱好？","output":"<think>\n他的爱好：台球、足球、羽毛球和乒乓球。"}
-{"input":"XiongJJ的性格如何？","output":"<think>\n他为人友善、乐于助人、善于思考、勤于动手。"}
-{"input":"XiongJJ脾气秉性如何？","output":"<think>\n整体性格比较温柔，不常发脾气，遇到事情常常告诉自己先冷静。"}
-{"input":"XiongJJ做事风格？","output":"<think>\n一句话总结：缜密但不拘小节，一般喜欢捞干的。"}
-{"input": "XiongJJ的职业是什么？","output": "XiongJJ目前担任软件工程师一职，擅长编程和技术解决方案。"}
-{"input": "XiongJJ喜欢喝什么饮料？","output": "他偏爱绿茶，偶尔也会品尝咖啡提神。"}
-{"input": "XiongJJ的兴趣爱好有哪些？","output": "他对足球、台球和羽毛球有着浓厚的兴趣，也享受与家人共度时光。"}
-{"input": "XiongJJ的家乡在哪里？","output": "XiongJJ来自中国黑龙江省的哈尔滨市。"}
-{"input": "XiongJJ的性格特点是什么？","output": "他为人友善、乐于助人，具有很强的责任感和团队合作精神。"}
-{"input": "XiongJJ最擅长哪种运动？","output": "他在羽毛球方面特别有天赋，经常参加本地的比赛。"}
-{"input": "XiongJJ的生活态度如何？","output": "他积极乐观，认为生活中总会有解决困难的方法。"}
-{"input": "XiongJJ平时怎么放松自己？","output": "他喜欢通过散步、阅读和听音乐来放松自己。"}
-{"input": "XiongJJ的家庭状况如何？","output": "XiongJJ已婚，有一个幸福的小家庭，并且非常重视家庭生活。"}
-{"input": "XiongJJ的理想是什么？","output": "他的理想是成为一名优秀的技术专家，并为社会做出贡献。"}
-{"input":"XiongJJ的职业生涯？","output":"<think>\nXiongJJ自大学毕业后一直从事软件开发工作，积累了丰富的项目经验。"}
-{"input":"XiongJJ的教育背景？","output":"<think>\n毕业于哈尔滨工业大学，主修计算机科学与技术。"}
-{"input":"XiongJJ最喜欢的运动是什么？","output":"<think>\n最喜欢的是足球，其次是羽毛球和台球。"}
-{"input":"XiongJJ的家庭生活是怎样的？","output":"<think>\n他非常重视家庭生活，经常花时间陪伴孩子，并参与家务活动。"}
-{"input":"XiongJJ的兴趣爱好有哪些？","output":"<think>\n除了体育运动，他还喜欢阅读科技书籍和探索新美食。"}
-{"input":"XiongJJ对孩子的教育理念是什么？","output":"<think>\n他认为应该给予孩子足够的自由和支持，同时强调责任感和独立思考能力的培养。"}
-{"input":"XiongJJ的业余时间怎么度过？","output":"<think>\n业余时间他会选择户外运动或者在家陪家人。"}
-{"input":"XiongJJ有什么特别的习惯吗？","output":"<think>\n每天早晨都会进行一次短暂的冥想，帮助自己放松和集中注意力。"}
-{"input":"XiongJJ在工作中是如何处理挑战的？","output":"<think>\n面对挑战时，他会冷静分析情况，然后制定详细的解决方案。"}
-{"input":"XiongJJ如何看待失败？","output":"<think>\n他认为失败是学习和成长的机会，每次失败后都会总结经验教训。"}
-{"input": "XiongJJ的职业是什么？","output": "XiongJJ目前是一名软件工程师。"}
-{"input": "XiongJJ的兴趣爱好有哪些？","output": "他喜欢足球、台球、羽毛球，并且也爱遛娃、做饭和洗碗。"}
-{"input": "XiongJJ来自哪里？","output": "XiongJJ来自黑龙江省哈尔滨市。"}
-{"input": "XiongJJ的个性如何？","output": "他为人友善、乐于助人，善于思考并勤于动手。"}
-{"input": "XiongJJ的性格特点是什么？","output": "他性格比较温柔，不常发脾气，遇事冷静处理。"}
-{"input": "XiongJJ的工作态度如何？","output": "他做事风格缜密但不拘小节，喜欢抓住问题的核心解决。"}
-{"input": "XiongJJ的出生年份？","output": "XiongJJ出生于1989年。"}
-{"input": "XiongJJ的家庭状况？","output": "XiongJJ的孩子分别叫：闫一虫、闫二虫、闫三虫、闫四虫、闫五。"}
-{"input": "XiongJJ平时喜欢做什么运动？","output": "他喜欢踢足球、打台球和羽毛球。"}
-{"input": "XiongJJ在家中的角色？","output": "他不仅是一个好爸爸，还经常帮助做家务，如做饭和洗碗。"}
-{"input": "XiongJJ的身高情况？","output": "XiongJJ个子不高，但他非常活跃。"}
-{"input": "XiongJJ在面对挑战时的态度？","output": "遇到事情时，他总是告诉自己要先冷静下来再行动。"}
-{"input": "XiongJJ的朋友圈评价他什么？","output": "朋友们都说他为人友善、乐于助人。"}
-{"input": "XiongJJ对生活的态度？","output": "XiongJJ对待生活充满热情，积极向上。"}
-{"input": "XiongJJ有没有特别的技能或特长？","output": "他是软件工程师，同时也擅长踢足球和打台球。"}
-{"input": "XiongJJ的教育背景？","output": "推测他具有计算机科学或相关领域的学位。"}
-{"input": "XiongJJ在团队合作中的表现？","output": "他在团队中非常合作，常常乐于助人。"}
-{"input": "XiongJJ的休闲时间怎么度过？","output": "他的休闲时间主要用于踢足球、打台球以及与家人共度时光。"}
-{"input": "XiongJJ的梦想是什么？","output": "他似乎致力于成为一个更好的父亲和专业人士。"}
+INFO:     127.0.0.1:48618 - "GET /health HTTP/1.1" 200 OK
+2026-05-15 14:36:49,475 - modelscope - INFO - Target directory already exists, skipping creation.
+2026-05-15 14:36:49 | INFO     | peft-platform | Model downloaded: Qwen/Qwen3.5-0.8B -> /root/Fine-tuning/backend/data/models/Qwen/Qwen3___5-0___8B
+Downloading Model from https://www.modelscope.cn to directory: /root/Fine-tuning/backend/data/models/Qwen/Qwen3.5-0.8B
+INFO:     172.19.0.3:58756 - "POST /api/v1/models/download HTTP/1.0" 200 OK
+INFO:     172.19.0.3:58768 - "GET /api/v1/models/ HTTP/1.0" 200 OK
+INFO:     172.19.0.3:58776 - "GET /api/v1/models/ HTTP/1.0" 200 OK
+INFO:     172.19.0.3:58778 - "GET /api/v1/datasets/ HTTP/1.0" 200 OK
+INFO:     172.19.0.3:58794 - "GET /api/v1/training/jobs HTTP/1.0" 200 OK
+INFO:     172.19.0.3:49392 - "GET /api/v1/training/jobs HTTP/1.0" 200 OK
+2026-05-15 14:37:04 | INFO     | peft-platform | Job 92be1d81-d263-4eff-8547-162728db09a7 enqueued
+2026-05-15 14:37:04 | INFO     | peft-platform | Training job created: 92be1d81-d263-4eff-8547-162728db09a7
+INFO:     172.19.0.3:49400 - "POST /api/v1/training/jobs HTTP/1.0" 200 OK
+2026-05-15 14:37:04 | ERROR    | peft-platform | Job 92be1d81-d263-4eff-8547-162728db09a7 failed: Dataset not found: 4af699cd-ed46-4dd4-ab36-84c54c709b62
+INFO:     172.19.0.3:49406 - "GET /api/v1/training/jobs HTTP/1.0" 200 OK
+INFO:     172.19.0.3:49420 - "GET /api/v1/datasets/ HTTP/1.0" 200 OK
+INFO:     172.19.0.3:49422 - "GET /api/v1/models/ HTTP/1.0" 200 OK
+INFO:     172.19.0.3:49432 - "GET /api/v1/training/jobs HTTP/1.0" 200 OK
+INFO:     172.19.0.3:48350 - "GET /api/v1/training/jobs HTTP/1.0" 200 OK
+INFO:     127.0.0.1:60174 - "GET /health HTTP/1.1" 200 OK
+INFO:     172.19.0.3:48352 - "GET /api/v1/datasets/ HTTP/1.0" 200 OK
+INFO:     172.19.0.3:48368 - "GET /api/v1/models/ HTTP/1.0" 200 OK
+INFO:     172.19.0.3:48378 - "GET /api/v1/training/jobs HTTP/1.0" 200 OK
+2026-05-15 14:37:17 | INFO     | peft-platform | Job 48cf1048-b31c-4457-a9e8-0add3d56a56b enqueued
+2026-05-15 14:37:17 | INFO     | peft-platform | Training job created: 48cf1048-b31c-4457-a9e8-0add3d56a56b
+INFO:     172.19.0.3:50122 - "POST /api/v1/training/jobs HTTP/1.0" 200 OK
+2026-05-15 14:37:17 | ERROR    | peft-platform | Job 48cf1048-b31c-4457-a9e8-0add3d56a56b failed: Dataset not found: 4af699cd-ed46-4dd4-ab36-84c54c709b62
+INFO:     172.19.0.3:50136 - "GET /api/v1/training/jobs HTTP/1.0" 200 OK
+INFO:     172.19.0.3:50142 - "GET /api/v1/models/ HTTP/1.0" 200 OK
+INFO:     172.19.0.3:50154 - "GET /api/v1/datasets/ HTTP/1.0" 200 OK
+INFO:     172.19.0.3:50162 - "GET /api/v1/training/jobs HTTP/1.0" 200 OK