Fix Llava conversion for models that use safetensors to store weights

[Isotr0py]

What does this PR do?

• Some llava variants like microsoft/llava-med-v1.5-mistral-7b use safetensors to store state_dict, even though they are using mistral backbone.
• This PR make the conversion script to be compatible with these models.

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,
  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?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

@amyeroberts @zucchini-nlp

[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.

[LysandreJik]

Thanks @Isotr0py!

[LysandreJik]

Awesome, thanks! Let's just wait for a quick re-read from Raushan and I'll let her merge