[feat] Docker Support

[soulteary]

Hi maintainer @lllyasviel 🐱

There are too many feedbacks about running problems due to environmental reasons in the project, maybe there should be a stable Docker image(s)?

If necessary, I can provide some basic templates, including traditional gpu or coreml env.

// Played this project for a few days, it is still very interesting. Hope the project is getting better and better. 🍻

[belyak]

+1

[tqcq]

I look forward to this feature.

[SamSamhuns]

I've set up a PR for this at #256.

Did an example run with the scribble generation. You guys can test it out.

[Utopiah]

I've set up a PR for this at #256.

Thanks for that, any rough indication on minimum or recommended specifications e.g min 2080 or 8G VRAM?

Also could "after models have been downloaded to ControlNet/models." become another script? Eventually if no model detected then download automatically on a an optionally mounted directory? (to avoid redownload)

[SamSamhuns]

8G VRAM was not enough for me. Needed like 10GBs for running the scribble-to-image script.

When the docker builds, the models are not copied directly into the image since these models are so big.

They are volume-mounted with docker run if you check the run_docker.sh script.

docker run \
      -ti --rm \
      -p "0.0.0.0:$port:7860" \
      -v "$PWD/models:/home/user1/models" \
      --gpus "device=0" \
      --name "$def_cont_name" \
      controlnet:latest