Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,21 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
datasets:
|
| 4 |
+
- shareAI/DPO-zh-en-emoji
|
| 5 |
+
language:
|
| 6 |
+
- zh
|
| 7 |
+
tags:
|
| 8 |
+
- qa
|
| 9 |
+
pipeline_tag: question-answering
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
## gemma2 中文版
|
| 13 |
+
Github教程仓库地址：https://github.com/CrazyBoyM/LLM-Chinese
|
| 14 |
+
|
| 15 |
+
shareAI-DPO对齐中文版
|
| 16 |
+
- 感谢opencsg的8 * A800 训练机器赞助
|
| 17 |
+
- 9b中文对话版本：https://opencsg.com/models/shareAI/gemma-2-9b-it-Chinese-DPO
|
| 18 |
+
- 2b中文对话版本：https://opencsg.com/models/shareAI/gemma-2-2b-it-Chinese-DPO
|
| 19 |
+
|
| 20 |
+
<img width="800" alt="image" src="https://github.com/user-attachments/assets/565946d2-7d46-4173-8a88-f589e4504c9c">
|
| 21 |
+
截图为2b版本的演示，实测当对话轮次增多后会与9b有一些明显性能差距，但10轮以内对话表现比较好。（测试了英文原版也是这样）
|