Update README.md
Browse files
README.md
CHANGED
|
@@ -11,6 +11,24 @@ This repository contains GGUF-v3 version (llama.cpp compatible) of **Chinese-Alp
|
|
| 11 |
|
| 12 |
**The optimal context length is 1K for this model. Specify `-c 1024` when using with llama.cpp**
|
| 13 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
For full model in HuggingFace format, please see: https://huggingface.co/hfl/chinese-alpaca-2-1.3b-rlhf
|
| 15 |
|
| 16 |
Please refer to [https://github.com/ymcui/Chinese-LLaMA-Alpaca-2/](https://github.com/ymcui/Chinese-LLaMA-Alpaca-2/) for more details.
|
|
|
|
| 11 |
|
| 12 |
**The optimal context length is 1K for this model. Specify `-c 1024` when using with llama.cpp**
|
| 13 |
|
| 14 |
+
|
| 15 |
+
## Performance
|
| 16 |
+
|
| 17 |
+
Metric: PPL, lower is better
|
| 18 |
+
|
| 19 |
+
| Quant | original | imatrix (`-im`) |
|
| 20 |
+
|-----|------|------|
|
| 21 |
+
| Q2_K | 20.1066 +/- 0.29236 | 18.8209 +/- 0.27561 |
|
| 22 |
+
| Q3_K | 16.9214 +/- 0.26133 | 16.5729 +/- 0.25706 |
|
| 23 |
+
| Q4_K | 16.1579 +/- 0.25064 | 15.7746 +/- 0.24476 |
|
| 24 |
+
| Q5_K | 15.3198 +/- 0.23627 | 15.4791 +/- 0.23959 |
|
| 25 |
+
| Q6_K | 15.3718 +/- 0.23764 | 15.2572 +/- 0.23549 |
|
| 26 |
+
|
| 27 |
+
*The model with `-im` suffix is generated with important matrix, which has generally better performance (not always though).*
|
| 28 |
+
|
| 29 |
+
|
| 30 |
+
## Others
|
| 31 |
+
|
| 32 |
For full model in HuggingFace format, please see: https://huggingface.co/hfl/chinese-alpaca-2-1.3b-rlhf
|
| 33 |
|
| 34 |
Please refer to [https://github.com/ymcui/Chinese-LLaMA-Alpaca-2/](https://github.com/ymcui/Chinese-LLaMA-Alpaca-2/) for more details.
|