Fix batched decode crash for hybrid cache models (Qwen3.5)#121

laudney wants to merge 1 commit into

vllm-project:mainvllm-project/vllm-metal:mainfrom

mmonad:fix/hybrid-model-batched-decodemmonad/vllm-metal:fix/hybrid-model-batched-decodeCopy head branch name to clipboard