hfl/chinese-llama-alpaca-2-awq

This folder contains pre-computed search results for Chinese-LLaMA-2 and Chinese-Alpaca-2 models, which is used to generate AWQ (Activation-aware Weight Quantization) models.

WARNING: These models MUST BE used with the original weights.

For usage, see:

• AWQ official github page: https://github.com/mit-han-lab/llm-awq
• llama.cpp github page: https://github.com/ggerganov/llama.cpp/tree/master/awq-py