[Feature]: AutoDeploy: Fully-fledged GPT-OSS support

[Fridah-nv]

🚀 The feature, motivation and pitch

1. mxfp4 moe reference op that is supported ootb on main
2. Mxfp4 moe optimized backends: triton vs cutlass kernels
3. 1 unquantized reference op instead of 2

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]

FYI, mxfp4 triton support will be deprecated: https://nvidia.slack.com/archives/C059LSY62BT/p1766440510352219?thread_ts=1766413734.980039&cid=C059LSY62BT

We will have to consider this in our GPT-OSS support