Skip to content

Commit

Permalink
Update README.md
Browse files Browse the repository at this point in the history
  • Loading branch information
binary-husky committed Jun 18, 2023
1 parent 8d7ee97 commit 29c6bfb
Showing 1 changed file with 7 additions and 2 deletions.
9 changes: 7 additions & 2 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -113,11 +113,16 @@ conda activate gptac_venv # 激活anaconda环境
python -m pip install -r requirements.txt # 这个步骤和pip安装一样的步骤
```

P.S. 如果需要依赖Latex的插件功能,请见Wiki

<details><summary>如果需要支持清华ChatGLM/复旦MOSS作为后端,请点击展开此处</summary>
<p>


<details><summary>如果需要依赖Latex的插件功能(如Arxiv文献翻译),请点击展开此处</summary>
<p>
请见[Wiki](https://github.com/binary-husky/gpt_academic/wiki/%E5%A6%82%E4%BD%95%E4%BD%BF%E7%94%A8Latex%E7%9B%B8%E5%85%B3%E6%8F%92%E4%BB%B6%EF%BC%88arxiv%E6%96%87%E7%AB%A0%E7%BF%BB%E8%AF%91%EF%BC%89)
</p>
</details>

【可选步骤】如果需要支持清华ChatGLM/复旦MOSS作为后端,需要额外安装更多依赖(前提条件:熟悉Python + 用过Pytorch + 电脑配置够强):
```sh
# 【可选步骤I】支持清华ChatGLM。清华ChatGLM备注:如果遇到"Call ChatGLM fail 不能正常加载ChatGLM的参数" 错误,参考如下: 1:以上默认安装的为torch+cpu版,使用cuda需要卸载torch重新安装torch+cuda; 2:如因本机配置不够无法加载模型,可以修改request_llm/bridge_chatglm.py中的模型精度, 将 AutoTokenizer.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True) 都修改为 AutoTokenizer.from_pretrained("THUDM/chatglm-6b-int4", trust_remote_code=True)
Expand Down

0 comments on commit 29c6bfb

Please sign in to comment.