Fix latents.dtype before vae.decode() at ROCm devices in StableDiffusionPipelines

[tolgacangoz]

Co-authored-by: Bagheera 59658056+bghira@users.noreply.github.com

What does this PR do?

This was talked about at a previous PR: #7858.

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.
@yiyixuxu @bghira

[bghira]

so i can actually hit this one on CUDA, but it only happens during training without autocast 🤔

the weights are in bf16 precision, not fp32. maybe this is what causes it?