Chinese-Alpaca-2-7B-GGUF

This repository contains the GGUF-v3 models (llama.cpp compatible) for Chinese-Alpaca-2-7B.

Performance

Metric: PPL, lower is better

Quant	original	imatrix (-im)
Q2_K	10.3705 +/- 0.14109	11.7442 +/- 0.16034
Q3_K	8.8478 +/- 0.12085	8.7104 +/- 0.11925
Q4_0	8.6418 +/- 0.11876	-
Q4_K	8.3294 +/- 0.11396	8.3034 +/- 0.11391
Q5_0	8.3320 +/- 0.11411	-
Q5_K	8.2361 +/- 0.11298	8.2136 +/- 0.11281
Q6_K	8.1956 +/- 0.11259	8.1852 +/- 0.11246
Q8_0	8.1784 +/- 0.11232	-
F16	8.1799 +/- 0.11243	-

The model with -im suffix is generated with important matrix, which has generally better performance (not always though).

Others

For Hugging Face version, please see: https://huggingface.co/hfl/chinese-alpaca-2-7b

Please refer to https://github.com/ymcui/Chinese-LLaMA-Alpaca-2/ for more details.