[TRTLLM-6780][fix] Add multimodal data to dummy requests during memory profiling#7539

Funatiq merged 24 commits into

NVIDIA:mainNVIDIA/TensorRT-LLM:mainfrom

johncalesp:add-mm-dummy-reqjohncalesp/TensorRT-LLM:add-mm-dummy-reqCopy head branch name to clipboard