TensorRT-LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and build TensorRT engines that contain state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT-LLM also contains components to create Python and C++ runtimes that execute those TensorRT engines.
I'm testing kv reuse feature
Everything works fine until i try to use offloading to host mem
I enable offloading by these lines
But when I tried to use the server I encountered an error
I'm using version from 30 april and the GptManager backend