wuxiaotiankevin/dockers

A collection of useful dockers.

Docker

This repo holds a collection of useful dockers.

Docker Basics

Install Docker

See https://docs.docker.com/.

Basics

An image is an executable package with everything to run the service.

An container is a runtime instance of an image.

Useful Commands

Containers

List running containers:

docker ps

Images

Build image:

• cd to the directory with Dockerfile
• docker build -t DOCKER_NAME .

List images:

docker image ls

Delete image:

docker image rm [-f] IMAGE

Delete untagged images:

docker rmi --force $(docker images --filter "dangling=true" -q --no-trunc)

Data Science

Start with Jupyter Docker Stack's Scipy Notebook

Install:

docker pull jupyter/scipy-notebook

Run:

docker run --rm -p 8888:8888 -v ~/projects:/home/jovyan/work/ jupyter/scipy-notebook

Deep Learning

There are many deep learning ready docker image available on GitHub. The following two are tested and ready to run.

Modern Deep Learning

Setup:

https://hub.docker.com/r/waleedka/modern-deep-learning/

or use my copy of Dockerfile

Run with command line: docker run -it -p 8888:8888 -p 6006:6006 -v ~/:/host modern-deep-learning

Run with Jupyter:

docker run -it -p 8888:8888 -p 6006:6006 -v ~/projects:/host modern-deep-learning jupyter notebook --ip=0.0.0.0 --allow-root /host

Deepo

CPU

Setup:

https://github.com/ufoym/deepo#CPU

Run:

docker run -it -p 8888:8888 -v $PWD:/root --ipc=host ufoym/deepo:all-py36-jupyter-cpu jupyter notebook --no-browser --ip=0.0.0.0 --allow-root --NotebookApp.token= --notebook-dir='/root/'

GPU

Setup:

https://github.com/ufoym/deepo#Jupyter

Run:

docker run --runtime=nvidia -it -p 8888:8888 -v $PWD:/root --ipc=host ufoym/deepo:all-jupyter-py36 jupyter notebook --no-browser --ip=0.0.0.0 --allow-root --NotebookApp.token= --notebook-dir='/root/'

Build my own

tf-object-detection

Setup:

docker build -t tf-object-detection .

Run:

docker run -it -p 8888:8888 -p 6006:6006 -v ~/:/host tf-object-detection jupyter notebook --ip=0.0.0.0 --allow-root /host

tensorflow-gpu-jupyter

Setup:

cd ./tf-gpu-jupyter docker build -t dl-gpu-jupyter .

Run:

docker run --runtime=nvidia -it -p 8888:8888 -v $PWD:/root --ipc=host dl-gpu-jupyter jupyter notebook --no-browser --ip=0.0.0.0 --allow-root --NotebookApp.token= --notebook-dir='/root/'

tensorflow-gpu

Setup:

cd ./tf-gpu docker build -t dl-gpu .

Run:

docker run --runtime=nvidia -it -v $PWD:/tmp -w /tmp dl-gpu bash

tensorflow-cpu

Setup:

cd ./tf-cpu docker build -t dl-cpu-jupyter .

Run:

docker run --runtime=nvidia -it -p 8888:8888 -v $PWD:/root --ipc=host dl-cpu-jupyter jupyter notebook --no-browser --ip=0.0.0.0 --allow-root --NotebookApp.token= --notebook-dir='/root/'

Investment

Build image:

• cd to ./investment/
• docker build -t invest .

Run:

docker run --rm -p 8888:8888 -v ~/projects:/home/jovyan/work/ invest