LLM inference in C/C++
--extra-index-url https://download.pytorch.org/whl/cpu torch torchvision; platform_machine != "s390x" transformers huggingface-hub accelerate sentence-transformers