Skip to content

Conversation

@larryliu0820
Copy link
Contributor

@larryliu0820 larryliu0820 commented Jul 9, 2025

A lot has changed on Llava model definition between 4.47 to 4.52, this PR:

  • Change the state dict key mapping to match the new Llava model definition in HF.
  • Use the processor.apply_chat_template() API to get input_ids so that we can be a bit more resilient to input_id format changes.

@pytorch-bot
Copy link

pytorch-bot bot commented Jul 9, 2025

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/12324

Note: Links to docs will display an error until the docs builds have been completed.

✅ No Failures

As of commit 54c6f2d with merge base aec1322 (image):
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Jul 9, 2025
@larryliu0820 larryliu0820 added the release notes: none Do not include this in the release notes label Jul 10, 2025
Copy link
Contributor

@lucylq lucylq left a comment

Choose a reason for hiding this comment

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

Thanks for updating this!

A lot has changed on Llava, this PR:

* Change the state dict key mapping to match the new Llava model
  definition in HF.
* Use the `processor.apply_chat_template()` API to get `input_id`s so that we can be a bit
  more resilient to input_id format changes.
@larryliu0820 larryliu0820 merged commit 378f062 into main Jul 10, 2025
103 checks passed
@larryliu0820 larryliu0820 deleted the bump_transformer branch July 10, 2025 20:34
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. release notes: none Do not include this in the release notes

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants