Add LoRA multimethod export to CoreML static LLM export

[lucylq]

Add --adapter CLI for exporting LoRA adapters as separate methods in
a CoreML PTE. CoreML POSITIONAL weight sharing deduplicates base weights
across methods. Supports combination with --multifunction for
decode/prefill variants per adapter.

Authored with Claude.

1. add --adapter CLI; hopefully we can re-use llm_config at some point?
2. skip_split_names: for target_modules with lora, skip splitting so that each qkvo is in the correct, matching 'position'. Without this we don't get weight sharing, unless we have a different mechanism for it outside of POSITIONAL.
3. Support lora alone, and lora with prefill & decode

[lucylq]

Stack from ghstack (oldest at bottom):

• [wip] Add CI test for CoreML LoRA multimethod export #18354
• [wip] Add --method argument to CoreML static LLM runner #18355
• -> Add LoRA multimethod export to CoreML static LLM export #18347
• Add skip_names and LoRALinear guard to replace_linear_with_split_linear #18346
• Add LoRA support to StaticAttention for split_mha=False #18345
• Refactor: extract _prepare_eager_model() from CoreML export main() #18343

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/18347

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

✅ You can merge normally! (2 Unrelated Failures)

As of commit 95a866e with merge base dd7464a ():

FLAKY - The following job failed but was likely due to flakiness present on trunk:

• pull / unittest-editable / windows / windows-job (gh) (matched win rule in flaky-rules.json)
  ##[error]The operation was canceled.

BROKEN TRUNK - The following job failed but was present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• pull / unittest / windows / windows-job (gh) (trunk failure)
  ##[error]The operation was canceled.

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[lucylq]

add methods for each lora

[lucylq]

add methods for each lora with separate prefill, decode

Not sure if this is how we want to do it, though.

[lucylq]

Not sure if we should add a method like this here. Or if we add, it should be more granular, like a list of lora methods.