juneup ka1tovo commited on
Commit
cd41098
·
verified ·
1 Parent(s): 6941d75

Upload 2 files (#2)

Browse files

- Upload 2 files (24408cd0ce19b33be106d7663f117a0ffc802555)


Co-authored-by: kuroba <[email protected]>

Files changed (3) hide show
  1. .gitattributes +1 -0
  2. README.md +32 -40
  3. distill.png +3 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ distill.png filter=lfs diff=lfs merge=lfs -text
README.md CHANGED
@@ -1,40 +1,32 @@
1
- ---
2
- license: mit
3
- ---
4
- ## 数据集
5
-
6
- ### 通用领域含思维链数据集:
7
- ```bash
8
- git lfs install
9
- git clone https://huggingface.co/datasets/Congliu/Chinese-DeepSeek-R1-Distill-data-110k-SFT
10
- ```
11
- 若不想克隆大型文件:
12
- ```bash
13
- GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/datasets/Congliu/Chinese-DeepSeek-R1-Distill-data-110k-SFT
14
- ```
15
- ### 含思维链心理辅导对话数据集:
16
- ```bash
17
- git lfs install
18
- git clone https://huggingface.co/datasets/CAS-SIAT-XinHai/CPsyCoun
19
- ```
20
- 若不想克隆大型文件:
21
- ```bash
22
- GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/datasets/CAS-SIAT-XinHai/CPsyCoun
23
- ```
24
-
25
- ## 模型
26
-
27
- ### 下载
28
- ```bash
29
- git lfs install
30
- git clone https://huggingface.co/juneup/internlm2.5_7b_distill
31
- ```
32
- **若不想克隆大型文件:
33
- ```bash
34
- GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/juneup/internlm2.5_7b_distill
35
- ```
36
-
37
- ### 在Ollama下载
38
- ```bash
39
- ollama run Juneup/internlm2.5_7b_distill:q4_k_m
40
- ```
 
1
+ # internlm2.5_7b_distill
2
+ ## 架构图
3
+ <div align="center">
4
+ <img src="distill.png" width="800"/>
5
+ </div>
6
+
7
+ ## 基座模型
8
+ https://huggingface.co/Slipstream-Max/Emollm-InternLM2.5-7B-chat-GGUF-fp16
9
+
10
+ ## 数据集
11
+ ### 数据集组成
12
+ 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)
13
+
14
+ 1. 110k的数据怎么挑选10k条?:基于score
15
+ 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
16
+
17
+ ### 训练方式
18
+ qlora,变长注意力,数据拼接
19
+
20
+ ## 下载模型
21
+ ```bash
22
+ git lfs install
23
+ git clone https://huggingface.co/juneup/internlm2.5_7b_distill
24
+ ```
25
+ 若不想克隆大型文件:
26
+ ```bash
27
+ GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/juneup/internlm2.5_7b_distill
28
+ ```
29
+
30
+ ### 在Ollama下载
31
+ ```bash
32
+ ollama run Juneup/internlm2.5_7b_distill:q4_k_m
 
 
 
 
 
 
 
 
distill.png ADDED

Git LFS Details

  • SHA256: 7dce2fa42bf884305c8a8f4a8807954888987b5896d331a95f52c16656f5bf30
  • Pointer size: 131 Bytes
  • Size of remote file: 177 kB