1.安裝nvidia-docker
$ wget -P /tmp https://github.com/NVIDIA/nvidia-docker/releases/download/v1.0.1/nvidia-docker_1.0.1-1_amd64.deb
$ sudo dpkg -i /tmp/nvidia-docker*.deb
無網(wǎng)絡(luò)安裝可能就麻煩一點(diǎn)
下載現(xiàn)成的包
nvidia-docker-1.0.1-1.x86_64.deb
2. pull tensorflow鏡像
$ nvidia-docker pull tensorflow/tensorflow:latest-gpu
3. 運(yùn)行
sudo service nvidia-docker start #一定要開啟
sudo nvidia-docker run -it -p 8888:8888 tensorflow/tensorflow:latest-gpu 或者
sudo nvidia-docker run -it -p 8888:8888 tensorflow/tensorflow:latest-gpu /bin/bash
nvidia-docker run -it -v /home:/home -p 8883:8883 tensorflow/tensorflow:latest-gpu /bin/bash
$ sudo nvidia-docker exec -it pp_tf /bin/bash #
$ systemctl start nvidia-docker #開啟nvidia-docker
$ systemctl status nvidia-docker #查看nvidia-docker狀態(tài)
$ sudo nvidia-docker run --rm nvidia/cuda nvidia-smi
在這個(gè)基礎(chǔ)上安裝其他包萝喘，再進(jìn)行commit
**4. docker save **
sudo docker save 840bf96be71f -o /dir/pp_tf_gpu_v1.tar #將鏡像保存 也可以用鏡像名進(jìn)行保存
5. docker load
docker load < / dir/pp_tf_gpu_v1.tar #可以重新加載鏡像了
sudo docker tag 840bf96be71f pp/tf:v1 #設(shè)置tag
