File size: 870 Bytes
47bc728 d45a7b9 47bc728 cc7fe76 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 |
---
license: apache-2.0
language:
- zh
- en
---
# Chinese-Mixtral-Instruct-GGUF
This repository contains the GGUF-v3 models (llama.cpp compatible) for **Chinese-Mixtral-Instruct** (chat/instruction model).
## Performance
Metric: PPL, lower is better
| Quant | PPL |
| ----- | ---- |
| IQ2_XXS | - |
| IQ2_XS | - |
| Q2_K | - |
| Q3_K | - |
| Q4_0 | - |
| Q4_K | - |
| Q5_0 | - |
| Q5_K | - |
| Q6_K | - |
| Q8_0 | - |
| F16 | x |
Due to the file size limitation, for F16 model, please use `cat` command to concatenate all parts into a single file. **You must concatenate these parts in order.**
## Others
For Hugging Face version, please see: https://huggingface.co/hfl/chinese-mixtral-instruct
Please refer to [https://github.com/ymcui/Chinese-Mixtral/](https://github.com/ymcui/Chinese-Mixtral/) for more details. |