PaintsChainer-Docker

Docker container for PaintsChainer

How to use this image

If you want to run this on Windows and have no familiarity with Docker, etc you can find Complete setup instructions for Windows (including Docker) on the wiki.

Processing via CUDA on GPU

Ensure you have nvidia-docker installed for GPU passthrough to containers. To run with the default GPU:

$ nvidia-docker run --name paintschainer --rm -p 8000:8000 liamjones/paintschainer-docker

If you have multiple GPUs and want to specify an alternate one you can specify GPU number via an environment variable:

$ nvidia-docker run --name paintschainer --rm -p 8000:8000 -e PAINTSCHAINER_GPU=1 liamjones/paintschainer-docker

GPU numbers can be verified by running:

$ docker run --rm liamjones/paintschainer-docker nvidia-smi

Processing via CPU

This will be slower than GPU processing but has some advantages;

• Doesn't require an NVIDIA card
• Doesn't require nvidia-docker
• Can potentially process larger images (assuming you have more RAM than VRAM)

$ docker run --name paintschainer --rm -p 8000:8000 -e PAINTSCHAINER_GPU=-1 liamjones/paintschainer-docker

Access the web interface

After bringing up the container, the web interface should be available at http://localhost:8000/

Notes

This image is currently only recommended for local use because directory browsing is available with the python server so all uploaded images are essentially public