Fix issue with prompt embeds and latents in SD Cascade Decoder with multiple image embeddings for a single prompt.

[DN6]

What does this PR do?

There is a use case with the cascade decoder pipeline where a user might set num_image_per_prompt > 1 in the prior pipeline and pass the output to the decoder.

This results in a case where you have multiple image embeddings associated with a single prompt. This is causing an issue with the how the batch size is set for latents and prompt embeds in the decoder pipeline. T

Fixes #7377

This PR:

1. Adds a fix to correctly set the batch size in the prompt embeds and latents to handle multiple embeddings per prompt.
2. Adds a test to check that the correct number of outputs is produced by the decoder when multiple image embeddings are used with a single prompt.

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.

[yiyixuxu]

thanks!