[Bugfix] Fix _get_lora_device for HQQ marlin

[varun-sundar-rabindranath]

Account for HQQ Marlin in _get_lora_device()

Bug: HQQ models with LoRA will raise ValueError(f"Unsupported base layer: {base_layer}") from here.

Fix:
hqq_marlin.py, registers the weights under the name "W_q". Look for attribute name "W_q" in _get_lora_device() for HQQ marlin.

Testing:

#!/bin/bash
num_prompts=100
lora_path="Taiwar/llama-3.2-1b-instruct-lora_model-1epoch" # grabbed a random adapter from HF 
model="nm-testing/Llama-3.2-1B-Instruct-HQQ" 

python3 benchmarks/benchmark_throughput.py --model  ${model} --backend vllm   --dataset ./ShareGPT_V3_unfiltered_cleaned_split.json --num-prompts ${num_prompts} --max-loras 2 --max-lora-rank 16  --enable-lora --lora-path ${lora_path}

[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 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 do one of these:

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

🚀

[varun-sundar-rabindranath]

@mgoin @ElizaWszola @jeejeelee PTAL ! Thanks.

[mgoin]

Thanks for finding this!