docker中使用GPU

github项目https://github.com/chatchat-space/Langchain-Chatchat使用docker容器启动时使用以下命令

docker run -d --gpus all -p 80:8501 registry.cn-beijing.aliyuncs.com/chatchat/chatchat:0.2.7

如果docker gpu没有安装好将报错如下

docker: Error response from daemon: could not select device driver "" with capabilities: [[gpu]]. AFTER installing nvidia-docker2

解决方案参考stackoverflow

Configure the repository:

curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg
&& curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list |
sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' |
sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list
&&
sudo apt-get update

Install the NVIDIA Container Toolkit packages:

sudo apt-get install -y nvidia-container-toolkit

Configure the container runtime by using the nvidia-ctk command:

sudo nvidia-ctk runtime configure --runtime=docker

Restart the Docker daemon:

sudo systemctl restart docker

上述解决方案的前提是gpu驱动已经安装成功，查看gpu驱动是否安装成功可以参考这篇文章、安装NVIDIA显卡、安装NVIDIA显卡

nvidia-smi # 确认有显卡以后输入下面命令，以检查之前是否安装了驱动。
ubuntu-drivers devices # 如果发现没有安装过驱动，则可再输入下面命令，以查看显卡型号以及推荐的驱动版本
sudo ubuntu-drivers autoinstall # 安装驱动

nvidia-docker

https://hub.docker.com/r/windj007/jupyter-keras-tools

物理机显卡驱动安装

ubuntu 查看是否存在显卡

lspci | grep -i vga

docker run --gpus all -p 8089:8089 -p 8001:8000 --rm -it -v /data/llm:/data/llm huggingface/transformers-pytorch-gpu /data/llm/ChatGLM-6B/run.sh
docker run -d --gpus all -p 8089:8089 -p 8001:8000 --rm -it -v /data/llm:/data/llm huggingface/transformers-pytorch-gpu jupyter-lab --port=8089 --NotebookApp.token='' --allow-root --ip=0.0.0.0