Use defaultdict for _SET_ADAPTER_SCALE_FN_MAPPING

[Alexkkir]

What does this PR do?

Currently everyone patches _SET_ADAPTER_SCALE_FN_MAPPING when training their models with peft
https://github.com/search?q=_SET_ADAPTER_SCALE_FN_MAPPING&type=code

Common patching practice looks like:

_SET_ADAPTER_SCALE_FN_MAPPING = {
    "MyModelName": lambda model_cls, weights: weights,
}

And as result every few months people modify file peft.py to add their own model:

The reason for this is lack of default behavior for _SET_ADAPTER_SCALE_FN_MAPPING.

This pr proposes to replace _SET_ADAPTER_SCALE_FN_MAPPING with defaultdict. It will simplify adding new models to diffusers as well as training and inference models with peft.

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?

@sayakpaul, probably you can review this pr. Thanks!

[sayakpaul]

Super cool! Thank you.

[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.

[Alexkkir]

@sayakpaul some tests crashed, is this ok? is this related to proposed change?

[sayakpaul]

Those are unrelated. Will merge after another round of CI.

[sayakpaul]

(Fixing the CUDA test failures in #13323)

[sayakpaul]

Failing tests are unrelated.