https://github.com/geyang/ml-research-containers

A list of community maintained docker images for deep reinforcement learning research
https://github.com/geyang/ml-research-containers

Last synced: about 2 months ago
JSON representation

A list of community maintained docker images for deep reinforcement learning research

• Host: GitHub
• URL: https://github.com/geyang/ml-research-containers
• Owner: geyang
• Created: 2022-01-14T21:03:35.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2024-10-31T05:14:37.000Z (2 months ago)
• Last Synced: 2024-10-31T06:18:39.197Z (2 months ago)
• Language: Dockerfile
• Size: 67.4 KB
• Stars: 3
• Watchers: 3
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# RL Docker Images

A list of community-maintained docker images for deep reinforcement learning research.

Our experience has shown that docker and singularity-based workflow works really well, to the extent that you only need to install your packages once, in a docker image, and you can deploy it directly via singularity anywhere.

## Why Upload Singularity Images

In some of the HPC systems, the file system is locked, causing images that are built within to fail.

```bash

mkdir -p /state/partition1/user/$USER  

export TMPDIR=/state/partition1/user/$USER  

singularity pull --docker-login docker://alpine 

this works fine. And then the following line breaks:

singularity exec --nv alpine_latest.sif bash -c 'echo "hey"'

```

Actual behavior

```

INFO:    Could not find any nv files on this host!

WARNING: passwd file doesn't exist in container, not updating

WARNING: group file doesn't exist in container, not updating

```

## Considerations for Building Singulairity Images

- https://stackoverflow.com/questions/61143522/file-ownership-and-permissions-in-singularity-containers

## Model-free Deep Reinforcement Learning

This image includes mujoco-py, gym, and DeepMind control. The Singularity image is built on Canon.