[TRTLLM-15936][fix] Enable breakable prefill CUDA graphs (BCG) for Nemotron-H hybrid models - #18888

[TRTLLM-15936][fix] Enable breakable prefill CUDA graphs (BCG) for Nemotron-H hybrid models#18888

dominicshanshan merged 4 commits into

NVIDIA:mainNVIDIA/TensorRT-LLM:mainfrom

dominicshanshan:user/shanshan/fix_BCG_in_nemotrondominicshanshan/TensorRT-LLM:user/shanshan/fix_BCG_in_nemotronCopy head branch name to clipboard