Skip to content

Code Server Docker image for PyTorch with python development on the browser. Includes CUDA!

License

Notifications You must be signed in to change notification settings

works-on-my-machine/pytorch-code-server

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

29 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PyTorch Code Server with CUDA

Code Server Docker image for PyTorch with python development on the browser. Contains:

  • CUDA 11.1
  • Python 3.8.3
  • PyTorch 1.81.1
  • Code Server 3.10.0 (VSCode 1.56.0)

Requirements

Quickstart

$ docker run --privileged --rm -it --init \
  --gpus=all \
  --ipc=host \
  --user="$(id -u):$(id -g)" \
  --volume="$PWD:/projects" \
  -p 8443:8443 \
  ghcr.io/works-on-my-machine/pytorch-code-server:1.10.0

After running above command open localhost:8443 in your browser. Find your password under ~/.config/code-server/config.yaml

$ docker exec -it <your_container_name> /bin/bash
$ cat ~/.config/code-server/config.yaml

Login with your password. For a working example look at example project folder. Contains recommended extensions and settings. Example project also contains noVNC support. Checkout docker-compose.yml file for more information.

Notes

- Due to a possible bug in ms-python.python extension. After installing it revert it to a previous version by selecting "Install Another Version". 2020.10.332292344 worked for me.

TODO

  • Better documentation
  • NO-VNC support for visualization
  • Docker compose support