[TRTLLM-14708][fix] Populate speculative-decoding request perf metrics in the PyTorch one-engine flow - #17082

[TRTLLM-14708][fix] Populate speculative-decoding request perf metrics in the PyTorch one-engine flow#17082

brnguyen2 merged 5 commits into

NVIDIA:mainNVIDIA/TensorRT-LLM:mainfrom

brnguyen2:feat/spec-dec-perf-metricsbrnguyen2/TensorRT-LLM:feat/spec-dec-perf-metricsCopy head branch name to clipboard