Chinese-Alpaca-2-13B-GGUF

This repository contains the GGUF-v3 models (llama.cpp compatible) for Chinese-Alpaca-2-13B.

Performance

Metric: PPL, lower is better

Quant	original	imatrix (-im)
Q2_K	13.7636 +/- 0.19446	20.6803 +/- 0.31594
Q3_K	9.5388 +/- 0.13078	9.1016 +/- 0.12565
Q4_0	9.1694 +/- 0.12668	-
Q4_K	8.6633 +/- 0.11957	8.6377 +/- 0.11932
Q5_0	8.6745 +/- 0.12020	-
Q5_K	8.5161 +/- 0.11796	8.5210 +/- 0.11803
Q6_K	8.4943 +/- 0.11759	8.5011 +/- 0.11775
Q8_0	8.4595 +/- 0.11718	-
F16	8.4550 +/- 0.11713	-

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-2-13b

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