Support AWQMarlin with MLA

[mgoin]

lm_eval --model vllm --model_args pretrained=TechxGenus/DeepSeek-Coder-V2-Lite-Instruct-AWQ --trust_remote_code --tasks gsm8k --num_fewshot 5 --batch_size auto 
...
INFO 02-12 19:01:25 cuda.py:161] Using Triton MLA backend.
...
INFO 02-12 19:01:41 worker.py:267] model weights take 10.45GiB; non_torch_memory takes 0.16GiB; PyTorch activation peak memory takes 0.94GiB; the rest of the memory reserved for KV Cache is 59.75GiB.
INFO 02-12 19:01:41 executor_base.py:110] # CUDA blocks: 116023, # CPU blocks: 7767
INFO 02-12 19:01:41 executor_base.py:115] Maximum concurrency for 163840 tokens per request: 11.33x
...
vllm (pretrained=TechxGenus/DeepSeek-Coder-V2-Lite-Instruct-AWQ,trust_remote_code=True), gen_kwargs: (None), limit: None, num_fewshot: 5, batch_size: auto
|Tasks|Version|     Filter     |n-shot|  Metric   |   |Value |   |Stderr|
|-----|------:|----------------|-----:|-----------|---|-----:|---|-----:|
|gsm8k|      3|flexible-extract|     5|exact_match|↑  |0.7119|±  |0.0125|
|     |       |strict-match    |     5|exact_match|↑  |0.6823|±  |0.0128|

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[mgoin]

Replaced by #13181