Skip to content

Conversation

@Isotr0py
Copy link
Contributor

@Isotr0py Isotr0py commented Dec 4, 2024

What does this PR do?

  • Some models like Llama-3.2-1B-Instruct-IMat-GGUF use tie_word_embeddings=True, but current GGUF model conversion didn't handle it and set to default value tie_word_embeddings=False.
  • This PR fix the tie_word_embeddings handling for GGUF models

Before submitting

  • This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
  • Did you read the contributor guideline,
    Pull Request section?
  • Was this discussed/approved via a Github issue or the forum? Please add a link
    to it if that's the case.
  • Did you make sure to update the documentation with your changes? Here are the
    documentation guidelines, and
    here are tips on formatting docstrings.
  • Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

@SunMarc

Signed-off-by: Isotr0py <2037008807@qq.com>
Signed-off-by: Isotr0py <2037008807@qq.com>
Copy link
Member

@SunMarc SunMarc left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM ! Thanks ! Is there an specific issue linked to this ?

@Isotr0py
Copy link
Contributor Author

Isotr0py commented Dec 4, 2024

I think there is not, I just realized this when trying to migrate test_ggml.py to use Qwen2.5-0.5B for dequantization tests, because it generates totally gibberish outputs.

@HuggingFaceDocBuilderDev

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

@SunMarc SunMarc requested a review from LysandreJik December 5, 2024 13:29
Copy link
Member

@LysandreJik LysandreJik left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cool, thanks @Isotr0py!

@SunMarc SunMarc merged commit 482cb28 into huggingface:main Dec 5, 2024
24 checks passed
@Isotr0py Isotr0py deleted the fix-llama-gguf branch December 5, 2024 17:06
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

Successfully merging this pull request may close these issues.

4 participants