Upload 2 files
#4
by
ka1tovo
- opened
- README.md +35 -35
- distill.png +2 -2
README.md
CHANGED
|
@@ -1,35 +1,35 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: mit
|
| 3 |
-
---
|
| 4 |
-
# internlm2.5_7b_distill
|
| 5 |
-
## 架构图
|
| 6 |
-
<div align="center">
|
| 7 |
-
<img src="distill.png" width="800"/>
|
| 8 |
-
</div>
|
| 9 |
-
|
| 10 |
-
## 基座模型
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
## 数据集
|
| 14 |
-
### 数据集组成
|
| 15 |
-
5k条精选通用领域含思维链数据(https://huggingface.co/datasets/Congliu/Chinese-DeepSeek-R1-Distill-data-110k-SFT)+3k条含思维链心理辅导对话(https://huggingface.co/datasets/CAS-SIAT-XinHai/CPsyCoun)
|
| 16 |
-
|
| 17 |
-
1. 110k的数据怎么挑选10k条?:基于score
|
| 18 |
-
2. 3k条心理辅导对话:原始对话(alpaca,含history字段,不含思维链,来自于CPsyCounD),使用deepseek r1【通过设置max_token实现低成本获取思维链】给output增加思维链,获得含思维链的数据集
|
| 19 |
-
|
| 20 |
-
### 训练方式
|
| 21 |
-
qlora,变长注意力,数据拼接
|
| 22 |
-
|
| 23 |
-
## 下载模型
|
| 24 |
-
```bash
|
| 25 |
-
git lfs install
|
| 26 |
-
git clone https://huggingface.co/juneup/internlm2.5_7b_distill
|
| 27 |
-
```
|
| 28 |
-
若不想克隆大型文件:
|
| 29 |
-
```bash
|
| 30 |
-
GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/juneup/internlm2.5_7b_distill
|
| 31 |
-
```
|
| 32 |
-
|
| 33 |
-
### 在Ollama下载
|
| 34 |
-
```bash
|
| 35 |
-
ollama run Juneup/internlm2.5_7b_distill:q4_k_m
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
---
|
| 4 |
+
# internlm2.5_7b_distill
|
| 5 |
+
## 架构图
|
| 6 |
+
<div align="center">
|
| 7 |
+
<img src="distill.png" width="800"/>
|
| 8 |
+
</div>
|
| 9 |
+
|
| 10 |
+
## 基座模型
|
| 11 |
+
https://huggingface.co/internlm/internlm2_5-7b-chat
|
| 12 |
+
|
| 13 |
+
## 数据集
|
| 14 |
+
### 数据集组成
|
| 15 |
+
5k条精选通用领域含思维链数据(https://huggingface.co/datasets/Congliu/Chinese-DeepSeek-R1-Distill-data-110k-SFT)+3k条含思维链心理辅导对话(https://huggingface.co/datasets/CAS-SIAT-XinHai/CPsyCoun)
|
| 16 |
+
|
| 17 |
+
1. 110k的数据怎么挑选10k条?:基于score
|
| 18 |
+
2. 3k条心理辅导对话:原始对话(alpaca,含history字段,不含思维链,来自于CPsyCounD),使用deepseek r1【通过设置max_token实现低成本获取思维链】给output增加思维链,获得含思维链的数据集
|
| 19 |
+
|
| 20 |
+
### 训练方式
|
| 21 |
+
qlora,变长注意力,数据拼接
|
| 22 |
+
|
| 23 |
+
## 下载模型
|
| 24 |
+
```bash
|
| 25 |
+
git lfs install
|
| 26 |
+
git clone https://huggingface.co/juneup/internlm2.5_7b_distill
|
| 27 |
+
```
|
| 28 |
+
若不想克隆大型文件:
|
| 29 |
+
```bash
|
| 30 |
+
GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/juneup/internlm2.5_7b_distill
|
| 31 |
+
```
|
| 32 |
+
|
| 33 |
+
### 在Ollama下载
|
| 34 |
+
```bash
|
| 35 |
+
ollama run Juneup/internlm2.5_7b_distill:q4_k_m
|
distill.png
CHANGED
|
Git LFS Details
|
|
Git LFS Details
|