GPU Docker image for stable baselines
2.0K
This image has been installed the stable-baselines dependency packages. You just have to mount the stable-baselines source code and the gym environment, then you can start your training and evaluation.
You are able to find the images in “Tags” above, one of them has a MPI tag, and the others doesn’t have.
If you need to use GAIL, DDPG, TRPO, and PPO1 algorithms, choose the version with the MPI tag. Otherwise, use the latest version.
Detailed description: install stable-baselines package
git clone https://github.com/hill-a/stable-baselines.git
Hypothesis:
stable-baselines project path: /homes/user/stable-baselines/ gym environment path: /home/user/env/
Command:
docker run --gpus all -it --rm --network host --ipc=host \ --mount src=/homes/user/stable-baselines/stable_baselines/,target=/root/code/stable_baselines/,type=bind \ --mount src=/home/user/env/,target=/root/code/env/,type=bind \ ntutselab/stable-baselines-gpu:latest \ bash -c "cd /root/code/env/ && python train.py"
project directory: /homes/user/stable-baselines/ source code directory: (Please mount this directory) /homes/user/stable-baselines/stable_baselines/
Script:
If it is complicated for you to use these commands, please refer to here to run by scripts.
Content type
Image
Digest
Size
1.6 GB
Last updated
over 6 years ago
docker pull ntutselab/stable-baselines-gpu