Control image generation by edge maps generated with Edge Drawing Parameter Free (edpf). This is similar to ControlNet Canny but uses a modern algorithm for edge-detection which requires no parameter tuning (Canny was invented in 1986, Edge Drawing in 2012). You can use the gradio demo or edpf.py script to generate edge maps until a pre-processor is implemented in your favorite Stable Diffusion UI.
Feedback is welcome! I'm still improving this model and you can help me by generating simple usecases and discuss the results.
The case for a new edge model
Do you find all these settings for canny confusing and time-consuming?
Do your default edge maps have noise, artefacts and missing edges?
All your images come out a mess like they were drawn by a human?
Well look no further!
With Edge Drawing Parameter Free you can create a masterpiece with a single mouse click!
No tuning! No mistakes! No frustration!
By combining recent advances in deep learning, cloud computing and block chain technology(?) we created the EDPF control net model just for you.
And the best part: it's free! Parameter Free!
If you want to train your own control net see my article Play in Control - ControlNet training setup guide!
Description
Trained on clean image dataset, with non-square images and slightly better captions and batch size=32. Overall much better quality and improved no-prompt inference.
see https://huggingface.co/GeroldMeisinger/control-edgedrawing -> Experiment 6.1 for more information on training
FAQ
Comments (8)
do you have a workflow or method for adding same to comfyUI workflow and get some cool results. I am still not able to figure out how to add this model to the comfyUI-advanced controlnet tab. wait for your feedback
just download a controlnet workflow file for comfy https://comfyanonymous.github.io/ComfyUI_examples/controlnet/ and put the controlnet file in the right folder
ok where is the pre processor for this control net , supposed there's something convert the pictures like examples (lines) ?
you just introduced the control net that will receive drawing-lines ..?
you can use the preprocessor here: https://huggingface.co/spaces/GeroldMeisinger/edpf
it's not yet implemented for any webui
Hey, nice work here! Thanks a lot for sharing this tool. I tried the hugging space and works better than canny in just one click!
One question: will this work on SDXL 1.0 workflow?
thanks for the feedback. it won't work for SDXL because it uses a different architecture. we would have to train the controlnet specificly for SDXL unfortunately it requires at least 24GB VRAM but I only have 12 atm.
This is great, and the pre-processor is amazing for even the other CN models. Someone needs to poke the CN extension people to add this preprocessor
thanks for this!
If someone wants to tackle this, it's actually not that hard, see https://gitlab.com/-/snippets/3601881 it requires a specific opencv version though which might or might not conflict with the other requirements. and a1111 controlnet is quite open for additions I think https://github.com/Mikubill/sd-webui-controlnet/issues
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.