第四课：XTuner InternLM-Chat 个人小助手认知微调实践

1、安装环境，想冒个险，都用默认最新的版本试下，这样的话，出现问题后在解决问题的过程中加深下记忆。

conda create --name xtuner-env python=3.10 -y
conda activate xtuner-env

conda install pytorch==2.1.1 torchvision==0.16.1 torchaudio==2.1.1 pytorch-cuda=12.1 -c pytorch -c nvidia

cd ~
git clone https://github.com/InternLM/xtuner.git
cd xtuner
pip install -e '.[all]'
pip install -U xtuner

2、开干

mkdir ~/ft-oasst1 && cd ~/ft-oasst1
xtuner list-cfg

等的好慢啊，出结果了。

cd ~/ft-oasst1
xtuner copy-cfg internlm_chat_7b_qlora_oasst1_e3 .

cp -r /share/temp/model_repos/internlm-chat-7b ~/ft-oasst1/

cd ~/ft-oasst1
cp -r /share/temp/datasets/openassistant-guanaco .

xtuner train ./internlm_chat_7b_qlora_oasst1_e3_copy.py --deepspeed deepspeed_zero2 

cp -r /share/temp/model_repos/internlm-chat-7b /root/personal_assistant/model/Shanghai_AI_Laboratory

xtuner train /root/personal_assistant/config/internlm_chat_7b_qlora_oasst1_e3_copy.py --deepspeed deepspeed_zero2

ssh -CNg -L 6006:127.0.0.1:6006 root@ssh.intern-ai.org.cn -p 33693
streamlit run /root/personal_assistant/code/InternLM/web_demo.py --server.address 127.0.0.1 --server.port 6006

 user_avator = "/root/personal_assistant/code/InternLM/doc/imgs/user.png"

robot_avator = "/root/personal_assistant/code/InternLM/doc/imgs/robot.png"

ssh -CNg -L 127.0.0.1:6006:0.0.0.0:6006 -o "StrictHostKeyChecking no" root@ssh.intern-ai.org.cn -p 33693

后面那三句，至关重要啊，