[wip] support local chunked attentin for hybrid

[luccafong]

lm_eval --model vllm --tasks mmlu_pro --model_args pretrained=/data/users/fanglu/Llama-4-Scout-17B-16E-Instruct,max_model_len=131072,tensor_parallel_size=4 --output_path .../output --batch_size auto > /tmp/hybrid.log 2>&1 &

Groups	Version	Filter	n-shot	Metric		Value		Stderr
mmlu_pro	2	custom-extract		exact_match	↑	0.6488	±	0.0043

CUDA_VISIBLE_DEVICES=0,1,3,4 uvx --with vllm --extra-index-url https://wheels.vllm.ai/e60f550b3825cbce2d3c7e882b029e2c1d914d8d lm_eval --model vllm --tasks mmlu --model_args  pretrained=/data/users/fanglu/Llama-4-Scout-17B-16E-Instruct,max_model_len=131072,tensor_parallel_size=4 --output_path .../output --batch_size auto 

Groups	Version	Filter	n-shot	Metric		Value		Stderr
mmlu_pro	2	custom-extract		exact_match	↑	0.6513	±	0.0043

[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.

🚀