Skip to content

Conversation

@zucchini-nlp
Copy link
Member

@zucchini-nlp zucchini-nlp commented Sep 18, 2024

What does this PR do?

Fixes #33484.

One thing to note is that we are forced to save in a separate folder, not a json file with video_ prefix. I see a workaround here by accepting save_filename when saving image processorr, but seems quite against the general idea of saving/loading in transformers. So I guess we can leave it as separate folder and for this model only we'll maintain folder loading for a while, even after new video processor class is addded

cc @yonigozlan

@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.

Copy link
Contributor

@amyeroberts amyeroberts 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 adding this patch!

…sion.py

Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
@zucchini-nlp
Copy link
Member Author

zucchini-nlp commented Sep 19, 2024

Will update files in the hub soon, merging for now. Loading with old format works anyway

@zucchini-nlp zucchini-nlp merged commit e40bb48 into huggingface:main Sep 19, 2024
12 checks passed
@zucchini-nlp
Copy link
Member Author

Done! Updated and tested by loading back

BernardZach pushed a commit to BernardZach/transformers that referenced this pull request Dec 5, 2024
* load and save from video-processor folder

* Update src/transformers/models/llava_onevision/processing_llava_onevision.py

Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

---------

Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
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.

Saving and loading LlavaOnevisionProcessor results in unexpected behavior

3 participants