Description

Dockerfile respository for different tasks.

Prerequisites

• Docker: follow this guide to install it.
• dockerx: follow this guide to install it.

List of templates

If you click on any template below, you will see a guide on how to use it.

• Coding:

  • pyenv: development environment with CUDA + pyenv

  • pycharm: development environment with CUDA + pyenv + PyCharm

  • vscode: development environment with CUDA + pyenv + vscode

  • pytorch: development environment with CUDA + pyenv + PyTorch

  • pytorch-with-charm: development environment with CUDA + pyenv + PyTorch + PyCharm

  • pytorch-with-vscode: development environment with CUDA + pyenv + PyTorch + vscode

  • pytorch-with-charm-mmf: development environment with CUDA + pyenv + PyTorch + PyCharm + MMF

How to use any of the templates

1. Clone this repository:

   $ git clone https://github.com/luiscarlosgph/docker-templates.git
   $ cd docker-templates
   

2. Build the image you want to deploy:

   $ cd <image_dir>
   $ chmod +x build.sh
   $ ./build.sh <your_favourite_image_name>
   

   For example, for the luiscarlosgph/pycharm:latest image:

   $ cd pycharm
   $ chmod +x build.sh
   $ ./build.sh luiscarlosgph/pycharm:latest
   

   You do not need to follow the luiscarlosgph/pycharm:latest syntax, you can simply write whatever as name of the image.

3. Launch container:

   $ python -m dockerx.run --name <container_name> --image <image_name> --nvidia 1 --command 'sleep infinity'
   

   For example, for the luiscarlosgph/pycharm:latest image:

   $ python -m dockerx.run --name wild_turin --image luiscarlosgph/pycharm:latest --nvidia 1 --command 'sleep infinity'
   

   dockerx also allows you to mount volumes and set environment variables using the parameters --volume and --env. More info here.

4. Get a terminal of the container:

   $ docker exec -it <container_name> <terminal_path>
   

   For example, for the container named wild_turin:

   docker exec -it wild_turin /bin/bash 
   

   More info on the parameters of docker exec here.