Add conversion for interleave llava #31858
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
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. |
Co-authored-by: amyeroberts <[email protected]>
NielsRogge
approved these changes
Jul 10, 2024
amyeroberts
added a commit
to amyeroberts/transformers
that referenced
this pull request
Jul 19, 2024
* add conversion for interleave llava * remove debug lines * remove unused imports * Update src/transformers/models/llava/convert_llava_weights_to_hf.py Co-authored-by: amyeroberts <[email protected]> * small changes + docs --------- Co-authored-by: amyeroberts <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
Adds conversion script for the models proposed in this blog post.
Briefly: Llava-Next_Interleave is a series of models trained with interleaved visual inputs, including video and 3D images. Even though it's a lllava-next series, the authors have hardcoded the inference script to not use anyres technique, so I added it in LLaVa. We will not support video natively as LLaVaNeXTVideo, but user can pass each frame separately as if it's an interleaved image input
The main idea was to support the 0.5B checkpoint, because it has same performance as llava-next-7b in the blog post. Imo it can be a cool addition