Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

GLM4同样出现了CUDA error: device-side assert #635

Open
2 tasks
zzk2021 opened this issue Nov 1, 2024 · 3 comments
Open
2 tasks

GLM4同样出现了CUDA error: device-side assert #635

zzk2021 opened this issue Nov 1, 2024 · 3 comments
Assignees

Comments

@zzk2021
Copy link

zzk2021 commented Nov 1, 2024

System Info / 系統信息

centos 7 ,3090x4

Who can help? / 谁可以帮助到您?

No response

Information / 问题信息

  • The official example scripts / 官方的示例脚本
  • My own modified scripts / 我自己修改的脚本和任务

Reproduction / 复现过程

之前是正常的,llama factory 用lora微调后,多调用几次就会报这个错,由于cuda异步找不到具体位置

Expected behavior / 期待表现

不会报相关错误

@cssaudrey
Copy link

微调后推理阶段,我也报相同错误

@zRzRzRzRzRzRzR zRzRzRzRzRzRzR self-assigned this Nov 4, 2024
@zRzRzRzRzRzRzR
Copy link
Member

估计是词表溢出了,没有遇到过这个问题。。。
单卡能推吗,是不是用BF16

@zzk2021
Copy link
Author

zzk2021 commented Nov 4, 2024

估计是词表溢出了,没有遇到过这个问题。。。 单卡能推吗,是不是用BF16

确实很多次报错代码停在tokenizer过程,但我检查了都是常见字符。单卡还没试过,bf16没有用

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants