You can put the workflows in the ./workflows/
folder.
To run the workflow:
python run.py --service=<comfyui_version_name> --workflow=<workflow_path> --port=8188
You can choose one <comfyui_version_name>
to bring up:
- comfyui
- comfyui_ipadapter_plus
<workflow_path>
can be such as: ./workflows/ipadapter_portrait.json
The program will download all necessary models on the first run, and will store them in ./models
.
After you see messages like:
To see the GUI go to: http://0.0.0.0:8188
Open your browser and navigate to http://localhost:8188/
.
If you don't see your workflow rightaway, click Load Default
.
And you are good to go.
./input
, ./output
, ./models
, and ./workflows
will be mounted to ComfyUI, so you can use the images inside them.
To build image:
python build.py ./ComfyUI
It'll create an image liusida:comfyui
locally.
The image was created as it was. It might need to be updated.