-
Notifications
You must be signed in to change notification settings - Fork 105
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
Training loss goes to 0 and eval loss goes to nan #67
Comments
Hi, What GPUs are you using to finetune llama2? I used to have this issue with V100s, but works well with 3090s, A100s. |
I use A100 and L40. The issue occurs randomly, meaning that under the same settings, sometimes it happens and sometimes it doesn't. |
@HZQ950419 Can you share your Python environment configuration? The issue may related to certain versions of transformers or tokenizers etc. |
Hi, the packages are showing below:
|
Thanks! |
When using the script in the README to finetune llama2, the training loss goes to 0 and the eval loss goes to nan randomly.
The text was updated successfully, but these errors were encountered: