tprompt-llm

pip install ctranslate2
pip install --upgrade huggingface_hub
pip install accelerate

Manually do this: OpenNMT/CTranslate2#1501

ct2-transformers-converter --model mistralai/Mistral-7B-v0.1 --quantization int8 --output_dir ./models/ctranslate2 --low_cpu_mem_usage