Docker-Caffe-Jupyter

Install docker

https://www.docker.com/

Install Nvidia docker

If you use GPU, please install Nvidia docker
https://github.com/NVIDIA/nvidia-docker

Download caffe image

CPU version

docker run -ti bvlc/caffe:cpu caffe --version>

GPU version

nvidia-docker run -ti bvlc/caffe:gpu caffe --version
After above operations,you should see these images:
docker images

  REPOSITORY          TAG                 IMAGE ID            CREATED             SIZE
  nvidia/cuda         latest              1b0262301072        4 days ago          1.94GB
  bvlc/caffe          gpu                 ba28bcb1294c        8 weeks ago         3.38GB

Enter container

docker run -p 5555:8888 -i -t imagename /bin/bash # assign a port like 5555 or 7777

Install jupyter notebook

pip install jupyter notebook

Configure

wget https://raw.githubusercontent.com/waszy110/Docker-Caffe-Jupyter/master/jupyter_notebook_config.py -P /root/.jupyter/

Run jupyter notebook

jupyter notebook --allow-root

Open jupyter notebook

http://localhost:5555/ #password is 000

Build a new image include caffe and jupyter

check running container id

docker ps

Build new image

docker commit containerid newimagename

Run new image

docker run -p 5555:8888 -i -t newimagename /bin/bash

Some simple operations

Download image

docker pull imagename

Run a container

docker run -it imagename

Stop container

docker stop containerid 

Remove container

docker rm containerid

Remove image

docker rmi imageid

Check image id

docker images