Make sure this is what you used: github.com/Fannovel16/comfyui_controlnet_aux. There are many other options today. Just search for ControlNet preprocessors.
Thank you for your video. Finally, someone mentioned that xl should use a dedicated model. It has always bothered me about the errors during operation.
THX - Super Usefull for Newbie like me... I was super annoyed about the lack of controll what is Generated inside the Program. This Guide gave me some hope for better results in my future AI Pic. Generations 👍
Search "preprocessors" in the "Install custom nodes" section. There you should see the nodes with the preprocessors. I explain it here: ruclips.net/video/5yNyJTcPYao/видео.htmlsi=7u_shneFkPZUB2Vh&t=390
Thank you for this!! Amazing video. Have you done anything with applying controlnets to a specific part of an existing image before? That is what I am trying to figure out right now. Seems like best way would be through a sort of masking workflow mixed with inpaint models I'd assume, but I haven't found anything about doing this on SDXL yet.
The Load Image node has "Mask Draw" mode on it (Just right click and its there in the menu). This allows you to choose specific part of the image that you can use as a latent mask for a sampler. This means that the sampler will only render the image on the masked area.
At 17:10 ApplyControlNet node has Conditioning unplugged on right side, a few sec later it's plugged. Where did you plug it in? Thanks in advance.
It should be plugged to the output of the prompt text. Its in the video on 16:53
Oh.. Sorry, you ment the output lol. It goes into the KSampler positive input :) I didn't notice i missed that in the video. Thanks :)
You're really the ones that is really good to teach complex things in detail about ComfyUI, great job!!
Thanks for the feedback. :)
Very useful, so much detailed information and well explained, thank you!
Glad it was helpful! Thanks :)
very helpful videos
I installed the node indicated but the controlnet preprocessor shown @7:50 is not there when I search.
Make sure this is what you used: github.com/Fannovel16/comfyui_controlnet_aux.
There are many other options today. Just search for ControlNet preprocessors.
Thank you for your video. Finally, someone mentioned that xl should use a dedicated model. It has always bothered me about the errors during operation.
Glad it was helpful!
Great tutorial, thanks a million.
For some reason, ControlNet Processors node is missing ( 7:45 ), any idea why?
Make sure you installed the custom nodes. Check out the video, it explains how to install it.
THX - Super Usefull for Newbie like me... I was super annoyed about the lack of controll what is Generated inside the Program. This Guide gave me some hope for better results in my future AI Pic. Generations 👍
@The AI Art -> BTW - The "ControlNet Prerprocessor" i dont have found... But i got the Idea how it should work.
Search "preprocessors" in the "Install custom nodes" section. There you should see the nodes with the preprocessors. I explain it here: ruclips.net/video/5yNyJTcPYao/видео.htmlsi=7u_shneFkPZUB2Vh&t=390
Thank you for this!! Amazing video.
Have you done anything with applying controlnets to a specific part of an existing image before? That is what I am trying to figure out right now. Seems like best way would be through a sort of masking workflow mixed with inpaint models I'd assume, but I haven't found anything about doing this on SDXL yet.
The Load Image node has "Mask Draw" mode on it (Just right click and its there in the menu). This allows you to choose specific part of the image that you can use as a latent mask for a sampler. This means that the sampler will only render the image on the masked area.
You can not trademark or copyright a pose. You can take a picture of how someone is posing and use it however you wish.
Yep. I wonder what are the usage rights when it comes to depth maps or lines.
How can I apply scribble controlnet to SDXL? Can't seem to find a model for this, only for 1.5
Try using the Sketch or Canny models. They might be able to produce nice images with it.
👋
Hi :) Thank you for the consistency ;)
@@the-ai-art
You're doing the AI community a service.......so yes, I'm supporting your efforts with likes and comments on your content. 👍