enforce original size to be a list

[dom-dziela]

What does this PR do?

Fixes #33531
Changes the datatype of original_size from torch.tensor(original_size, dtype = torch.int64) to a list. This prevents possible calculation inaccuracies induced by hardware- and torch-versions, as in issue above, by conducting calculation in base python.

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,
• Did you read the 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?

@amyeroberts
@zucchini-nlp

[zucchini-nlp]

@dom-dziela Thanks for fixing this! Can you run make fix-copies to make our CI happy?

[dom-dziela]

If I run make fix-copies the changes are deleted. Am I doing something wrong? Sorry first PR :)

[zucchini-nlp]

@dom-dziela Ah, I forgot this is LLaVa-OV. The modified function has a # Copied from transformers.models.llava_next.modeling_llava_next.unpad_image just above, so you'd need to fix the same function in llava-next and run make fix-copies again 🤗

[zucchini-nlp]

@dom-dziela Thanks a lot! Can you pls also propagate changes to llava-next-video model? My bad that it didn't propagare automatically with make fix-copies

Also, requested review from core maintainer, after their approval we can merge

[amyeroberts]

Thanks for fixing!

[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.