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

feat: import model with gguf metadata extraction #3421

Merged
merged 3 commits into from
Aug 21, 2024

Conversation

louis-jan
Copy link
Contributor

@louis-jan louis-jan commented Aug 20, 2024

Describe Your Changes

The GGUF model import does not work properly across versions, as it brings a default model.json that likely breaks the model inference result.

  • Chat templates differ between models.
  • They are under Jinja template and need to be parsed and rendered properly.
  • Stop token is differ between models which can be retrieved from GGUF metadata.
  • NGL & ctx_length also the same.
  • A lot of unused legacy scripts need to be removed.
Imported model should work the same as original one
Screenshot 2024-08-20 at 17 01 02

Fixes Issues

  • Closes #

Self Checklist

  • Added relevant comments, esp in complex areas
  • Updated docs (for bug fixes / features)
  • Created issues for follow-up changes or refactoring needed

@louis-jan louis-jan force-pushed the feat/import-model-use-gguf-metadata branch from c156e36 to 96f393a Compare August 20, 2024 10:12
@louis-jan
Copy link
Contributor Author

Verified by @Van-QA

@louis-jan louis-jan merged commit c8474c8 into dev Aug 21, 2024
7 checks passed
@louis-jan louis-jan deleted the feat/import-model-use-gguf-metadata branch August 21, 2024 09:45
@github-actions github-actions bot added this to the v.0.6.0 milestone Aug 21, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants