ganchengguang
commited on
Commit
·
1a7e7d2
1
Parent(s):
e256d64
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,20 @@
|
|
1 |
---
|
2 |
license: mit
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: mit
|
3 |
---
|
4 |
+
|
5 |
+
This model is traned with [guanaco](https://huggingface.co/datasets/JosephusCheung/GuanacoDataset) dataset.
|
6 |
+
Improved performance in Chinese and Japanese.
|
7 |
+
Use the QLoRA to fine-tune the vanilla LLaMA2-7B.
|
8 |
+
|
9 |
+
|
10 |
+
|
11 |
+
### Recommend Generation parameters:
|
12 |
+
* temperature: 0.5~0.7
|
13 |
+
* top p: 0.65~1.0
|
14 |
+
* top k: 30~50
|
15 |
+
* repeat penalty: 1.03~1.17
|
16 |
+
|
17 |
+
|
18 |
+
|
19 |
+
|
20 |
+
Contribute by Yokohama Nationaly University Mori Lab.
|