Marlin MoE integration

[ElizaWszola]

Reimplement quantized Mixtral to combine Marlin kernels with fused MoE.

Unit testing:

pytest tests/kernels/test_moe.py -k test_fused_marlin_moe

End-to-end testing:
Run offline_inference.py with

// quantized moe with act order
llm = LLM(model="TheBloke/Mixtral-8x7B-v0.1-GPTQ",
          revision="gptq-4bit-128g-actorder_True")

and

// quantized moe without act order
llm = LLM(model="TheBloke/Mixtral-8x7B-v0.1-GPTQ")

Sonnet benchmark results (no act order):

TTFT	TPOT

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.
Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which consists a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of default ones by unblocking the steps in your fast-check build on Buildkite UI.

Once the PR is approved and ready to go, please make sure to run full CI as it is required to merge (or just use auto-merge).

To run full CI, you can do one of these:

• Comment /ready on the PR
• Add ready label to the PR
• Enable auto-merge.

🚀

[alexm-neuralmagic]

@robertgshaw2-neuralmagic @mgoin @dsikka

[alexm-neuralmagic]

Eliza will redirect to vllm directly