GPTQ-for-Llama-docker

Sophisticated docker builds for parent projects oobabooga/GPTQ-for-Llama and qwopqwop200/GPTQ-for-Llama

Easy setup. Compatible. Tweakable. Scaleable.

WIP

source alias.gptq-for-llama

gptq-for-llama convert_llama_weights_to_hf.py --input_dir /path/to/downloaded/llama/weights --model_size 7B --output_dir ./llama-hf