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

LLaVA image_token_index is not 64000 but 64002 #29836

Closed
4 tasks
XuweiyiChen opened this issue Mar 24, 2024 · 1 comment · Fixed by #29797
Closed
4 tasks

LLaVA image_token_index is not 64000 but 64002 #29836

XuweiyiChen opened this issue Mar 24, 2024 · 1 comment · Fixed by #29797
Labels
bug Multimodal Should Fix This has been identified as a bug and should be fixed.

Comments

@XuweiyiChen
Copy link

System Info

LLaVA image_token_index is not 64000 but 64002 in the latest version of the code. (main)

Who can help?

No response

Information

  • The official example scripts
  • My own modified scripts

Tasks

  • An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
  • My own task or dataset (give details below)

Reproduction

  1. pip install the latest transformer from main branch
  2. run llava 34b model.

Expected behavior

You can do inference freely but now an error will pop out saying there is 0 image token.

@NielsRogge
Copy link
Contributor

Hi,

Thanks for reporting, this is also reported here: https://huggingface.co/llava-hf/llava-v1.6-34b-hf/discussions/2 and will be resolved by #29797

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Multimodal Should Fix This has been identified as a bug and should be fixed.
Projects
None yet
Development

Successfully merging a pull request may close this issue.

3 participants