[Misc] Add attention mask pre-computation optimization back to Qwen2.5-VL#15273

ywang96 merged 1 commit intovllm-project:mainvllm-project/vllm:mainfrom

Isotr0py:optimize-qwen2_5_vlIsotr0py/vllm:optimize-qwen2_5_vlCopy head branch name to clipboard