Skip to content

Conversation

@aliencaocao
Copy link
Contributor

@aliencaocao aliencaocao commented Jun 25, 2024

What does this PR do?

This is to align with the behavior of pipeline where users can pass strings like torch_dtype="float32" and it gets recognized as torch.float32. It now also happens for any PretrainedModel.from_pretrained.

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?

@amyeroberts

@aliencaocao aliencaocao changed the title Allow dtype str for torch_dtype in from_pretrained [QoL] Allow dtype str for torch_dtype in from_pretrained Jun 25, 2024
@aliencaocao aliencaocao changed the title [QoL] Allow dtype str for torch_dtype in from_pretrained [QoL] Allow dtype str for torch_dtype arg of from_pretrained Jun 25, 2024
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! I think this looks OK, but let's get a second opinion from @ArthurZucker as this touches a core part of the modeling logic.

We should also add tests - probably under test_modeling_utils.py

@aliencaocao
Copy link
Contributor Author

@amyeroberts added them

@amyeroberts amyeroberts requested a review from ArthurZucker June 26, 2024 14:30
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 the tests!

Copy link
Collaborator

@ArthurZucker ArthurZucker left a comment

Choose a reason for hiding this comment

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

Nice nice

@ArthurZucker ArthurZucker merged commit 3a02810 into huggingface:main Jun 27, 2024
@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.

@aliencaocao aliencaocao deleted the allow-string-torch-dtype branch June 27, 2024 11:26
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.

4 participants