[Feature]: AutoDeploy: support rms_norm for attention sharding heuristic

[lucaslie]

🚀 The feature, motivation and pitch

Minimax (and other models) may have rms norm inside the sharded attention region --> we must shard the rms norm weights like the rest of the weights

Alternatives

No response

Additional context

No response

Before submitting a new issue...

• Make sure you already searched for relevant issues, and checked the documentation and examples for answers to frequently asked questions.

[lucaslie]

blocked by #9283 but you can work with a patch to patch it torch_rmsnorm in the meantime