resolve an error when floats are provided for resolution values

[bghira]

What does this PR do?

Inside prepare_latents, it will crash if a float was inadvertently passed to the pipeline.

This will cast the input values to integers, allowing these situations to proceed.

TypeError: randn(): argument 'size' failed to unpack the object at pos 3 with error "type must be tuple of ints,but got float"

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 our philosophy doc (important for complex PRs)?
• 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.

[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.

[DN6]

@bghira The quality checks are likely failing because the prepare_latents method is copied in other pipelines. One option is to apply casting before passing to prepare_latents or change the affected pipelines to cast inside the method. You can run python utils/check_copies.py to find the affected files.

CI message should have been more helpful here. Will take a look into it.

[bghira]

@DN6 thanks usually i push a proposed change to get feedback before i copy to the rest. that's done now

[DN6]

LGTM. Good to merge once tests pass. @bghira I think you might have to run make style && make quality as well.

[bghira]

done

[DN6]

@bghira I think some extra files got added here? The examples directory wasn't in the PR earlier?

[bghira]

you said to run make quality

[yiyixuxu]

should we update the dependencies
i.e. pip install -e ".[quality]" again?

[bghira]

i don't think there's an issue with those deps per se it's just that the main branch doesn't have the formatting the way it wanted.

[DN6]

@bghira What version of ruff do you have installed?

[bghira]

0.3.4

[DN6]

Ah, diffusers ruff is currently pinned

diffusers/setup.py

Line 124 in cf6e040

"ruff==0.1.5",

Could you downgrade and run make style && make quality that should take care of the tests.

[bghira]

nope... that doesn't fix it. it adds even more changes

[bghira]

thank god we tried to fix that one formatting issue, or else this PR wouldn't have 66 files with open changes :)

[DN6]

@bghira Just a couple of pipelines that also need to be updated with the int casting.

If you run make fix-copies those should be handled.

Traceback (most recent call last):
  File "utils/check_copies.py", line 222, in <module>
    check_copies(args.fix_and_overwrite)
  File "utils/check_copies.py", line [20](https://github.com/huggingface/diffusers/actions/runs/8690950165/job/23832556622?pr=7537#step:5:21)6, in check_copies
    raise Exception(
Exception: Found the following copy inconsistencies:
- src/diffusers/pipelines/audioldm2/pipeline_audioldm2.py: copy does not match pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_latents at line 789
- src/diffusers/pipelines/audioldm/pipeline_audioldm.py: copy does not match pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_latents at line 329
Run `make fix-copies` or `python utils/check_copies.py --fix_and_overwrite` to fix them.

[DN6]

@bghira If it's cool with you can I open a separate PR to get this update in and tag you as the author? There are some merge conflicts with your branch because of some recent changes to main and I think the MusicLDM pipeline also needs to have copies fixed.

[bghira]

sure

[yiyixuxu]

completed in #7691