tensorrt-llm

Optimize LLM inference with NVIDIA TensorRT-LLM. Use when running trtllm-build, converting HF checkpoints, serving with FP8/INT4, or maximizing GPU throughput.