[MoE][Refactor] Remove most arguments to FusedMoEMethodBase.apply

[bnellnm]

Purpose

After making select_experts a non-static method (#29067), we can avoid passing most of the arguments to FusedMoEMethodBase.apply and get them from the layer directly.

This doesn't really decrease flexibility, the MoE quantization methods were already fairly tightly coupled with FusedMoE. A further step could be making FusedMoE an abstract base class that simply provides the quant_method parameters. Then the current FusedMoE class would become a subclass of that.

Test Plan

CI

Test Result

distributed-tests-2-gpus - looks like a spurious failure

cc @varun-sundar-rabindranath

---

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• The test plan, such as providing test command.
• The test results, such as pasting the results comparison before and after, or e2e results
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.
• (Optional) Release notes update. If your change is user facing, please update the release notes draft in the Google Doc.

[mergify]

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @bnellnm.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[LucasWilkinson]

Looks like we need to land #29067 first? marking draft until that lands 👍

[bnellnm]

• draft

Yeah, sorry this should have been a draft to start with. It needs the other PR to land first.

[mergify]

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @bnellnm.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

[LucasWilkinson]

LGTM; thanks for doing this! this is amazing!

[bnellnm]

LGTM; thanks for doing this! this is amazing!

Thanks for the review!