Llama-3-Chinese-8B-GGUF

This repository contains Llama-3-Chinese-8B-GGUF (llama.cpp/ollama/tgw, etc. compatible), which is the quantized version of Llama-3-Chinese-8B.

Note: this is a foundation model, which is not suitable for conversation, QA, etc.

Further details (performance, usage, etc.) should refer to GitHub project page: https://github.com/ymcui/Chinese-LLaMA-Alpaca-3

Performance

Metric: PPL, lower is better

Note: Old models have been removed due to its inferior performance.

Quant	Size	PPL (old model)	👍🏻 PPL (new model)
Q2_K	2.96 GB	17.7212 +/- 0.59814	11.8595 +/- 0.20061
Q3_K	3.74 GB	8.6303 +/- 0.28481	5.7559 +/- 0.09152
Q4_0	4.34 GB	8.2513 +/- 0.27102	5.5495 +/- 0.08832
Q4_K	4.58 GB	7.8897 +/- 0.25830	5.3126 +/- 0.08500
Q5_0	5.21 GB	7.7975 +/- 0.25639	5.2222 +/- 0.08317
Q5_K	5.34 GB	7.7062 +/- 0.25218	5.1813 +/- 0.08264
Q6_K	6.14 GB	7.6600 +/- 0.25043	5.1481 +/- 0.08205
Q8_0	7.95 GB	7.6512 +/- 0.25064	5.1350 +/- 0.08190
F16	14.97 GB	7.6389 +/- 0.25001	5.1302 +/- 0.08184

For full model, please see: https://huggingface.co/hfl/llama-3-chinese-8b
For LoRA-only model, please see: https://huggingface.co/hfl/llama-3-chinese-8b-lora
If you have questions/issues regarding this model, please submit an issue through https://github.com/ymcui/Chinese-LLaMA-Alpaca-3