[Doc] Improved level of clarity for latents_to_rgb.

[LagPixelLOL]

What does this PR do?

This PR fixes the latents_to_rgb function in the doc which will result in an error because the function only gets the first channel then performs permutation on 3 channels.

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?

@stevhliu @sayakpaul

[asomoza]

can you please post and example of why the old version would fail? Apart from reducing LoC I see that you just pass the latents without the batch dimension.

Just to clarify, I think your version is good but I'm curious as to why do you say it fails before.

[LagPixelLOL]

Oh I think it's an error on my part, I was passing without batch dim to the function before, since I assumed as the function returns only a single image, it should only accept a single latent as the input. It will result in an error at the image_array = image_array.transpose(1, 2, 0) line with an error message of ValueError: axes don't match array.

I just tested with passing in with the batch dim and it will run successfully.

This PR still do improve LoC by only accepting without batch and returning a single image, as I got confused by the input value and assumed the function contains an error.

[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.

[asomoza]

thanks for explaining, LGTM, I will merge it when the tests pass.