add README.md for Emollm_Scientist
This commit is contained in:
parent
66b7617f04
commit
0c72c31b4f
94
config/README.md
Normal file
94
config/README.md
Normal file
@ -0,0 +1,94 @@
|
||||
# 微调指南
|
||||
|
||||
|
||||
##数据
|
||||
|
||||
##基座
|
||||
|
||||
##训练方式
|
||||
|
||||
##配置文件
|
||||
|
||||
## 一、基于xtuner的微调🎉🎉🎉🎉🎉
|
||||
|
||||
### 环境准备
|
||||
|
||||
```markdown
|
||||
datasets==2.16.1
|
||||
deepspeed==0.13.1
|
||||
einops==0.7.0
|
||||
flash_attn==2.5.0
|
||||
mmengine==0.10.2
|
||||
openxlab==0.0.34
|
||||
peft==0.7.1
|
||||
sentencepiece==0.1.99
|
||||
torch==2.1.2
|
||||
transformers==4.36.2
|
||||
xtuner==0.1.11
|
||||
```
|
||||
|
||||
也可以一键安装
|
||||
|
||||
```bash
|
||||
cd xtuner_config/
|
||||
pip3 install -r requirements.txt
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 微调
|
||||
|
||||
```bash
|
||||
cd xtuner_config/
|
||||
xtuner train internlm2_7b_chat_qlora_e3.py --deepspeed deepspeed_zero2
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 将得到的 PTH 模型转换为 HuggingFace 模型
|
||||
|
||||
**即:生成 Adapter 文件夹**
|
||||
|
||||
```bash
|
||||
cd xtuner_config/
|
||||
mkdir hf
|
||||
export MKL_SERVICE_FORCE_INTEL=1
|
||||
|
||||
xtuner convert pth_to_hf internlm2_7b_chat_qlora_e3.py ./work_dirs/internlm_chat_7b_qlora_oasst1_e3_copy/epoch_3.pth ./hf
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 将 HuggingFace adapter 合并到大语言模型
|
||||
|
||||
```bash
|
||||
xtuner convert merge ./internlm2-chat-7b ./hf ./merged --max-shard-size 2GB
|
||||
# xtuner convert merge \
|
||||
# ${NAME_OR_PATH_TO_LLM} \
|
||||
# ${NAME_OR_PATH_TO_ADAPTER} \
|
||||
# ${SAVE_PATH} \
|
||||
# --max-shard-size 2GB
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 测试
|
||||
|
||||
```
|
||||
cd demo/
|
||||
python cli_internlm2.py
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 二、基于Transformers的微调🎉🎉🎉🎉🎉
|
||||
|
||||
- 请查看[ChatGLM3-6b lora微调指南](ChatGLM3-6b-ft.md)
|
||||
|
||||
---
|
||||
|
||||
## 其他
|
||||
|
||||
欢迎大家给[xtuner](https://github.com/InternLM/xtuner)和[EmoLLM](https://github.com/aJupyter/EmoLLM)点点star~
|
||||
|
||||
🎉🎉🎉🎉🎉
|
Loading…
Reference in New Issue
Block a user