kvcache-ai / kvcache-ai/ktransformers

[Bug] 请教大家Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions 问题

Open
#1,566 2 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
19.5k
Forks
1.6k
Avg merge
19h 32m
Merged PRs (30d)
27

Description

检查清单
  • 1. 我已经搜索过相关问题,但未能获得预期的帮助
  • 2. 该问题在最新版本中尚未修复
  • 3. 请注意,如果您提交的BUG相关 issue 缺少对应环境信息和最小可复现示例,我们将难以复现和定位问题,降低获得反馈的可能性
  • 4. 如果您提出的不是bug而是问题,请在讨论区发起讨论 https://github.com/kvcache-ai/ktransformers/discussions。否则该 issue 将被关闭
  • 5. 为方便社区交流,我将使用中文/英文或附上中文/英文翻译(如使用其他语言)。未附带翻译的非中文/英语内容可能会被关闭
问题描述

RuntimeError: CUDA error: device-side assert triggered
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

复现步骤

我在使用local_chat.py 改成读取jsonl文件 去批处理 大概在进行10~20个问题 就出现这个问题 请教大家怎么解决?

环境信息

A100 单卡 ubuntu 22.04 cpu Intel(R) Xeon(R) Gold 6336Y CPU @ 2.40GHz

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with local_chat.py and reproduce the failure using the reported JSONL batch-processing workflow on an A100 with Ubuntu 22.04. Check whether the error consistently appears after 10–20 questions, then document a minimal reproducible input and the exact failing operation so the CUDA assertion can be investigated.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.