Good video . Your workflow was a little broken on OpenArt. The Pos and Neg nodes from the Control Net group must connect into the Ksampler in the Output group. Your video was correct. Also - Faceswap is incorrect for the title of this video. This is a Bodypose video. Face swap means you would take the target picture face (indian girl) and put the face onto the Target picture (standing lady with jumper). Everything else in the target picture should remain exactly the same.
Thank you for your comment. In this video, the Target Face is the face we generated using a prompt (instead of loading another photo), and the Source is the face of the Indian girl, which we take and place on the generated face using the prompt.
@@CgTopTips 100% agree with you. That IS what you are doing :) I am only saying that is not really called a 'face swap', as per my comment earlier. I'd be really keen to see how you do a face swap. Take the indian girl's face to replace the face of another picture. You've done some really good work so far. Well done - and thank you
well the object has no attribute error seemed to be due to i neglected to use an sdxl check point (oops) but now i am getting mat1 and mat2 shapes cannot be multiplied (16x2048 and 768x320) error, which i believe may be due to non-matching input image dimensions? should we have a node in here to autoresize input images to keep things smooth?
Thank you, my friend! It is working awesome!
Awesome! DO you think is possible to do this by inputting two images each with diff person and one single image with the two people in it hugging?
Use batch image node
Good video. Thanks
In my opinion, faceID is much better
i keep getting an "X" at the center of the photo. sometimes it is overlayed on top of the portrait and sometimes I only get the "X"
got as far as instant id and it failed in manager so guess that's it .
Good video . Your workflow was a little broken on OpenArt.
The Pos and Neg nodes from the Control Net group must connect into the Ksampler in the Output group. Your video was correct.
Also - Faceswap is incorrect for the title of this video.
This is a Bodypose video. Face swap means you would take the target picture face (indian girl) and put the face onto the Target picture (standing lady with jumper).
Everything else in the target picture should remain exactly the same.
Thank you for your comment. In this video, the Target Face is the face we generated using a prompt (instead of loading another photo), and the Source is the face of the Indian girl, which we take and place on the generated face using the prompt.
@@CgTopTips 100% agree with you. That IS what you are doing :)
I am only saying that is not really called a 'face swap', as per my comment earlier.
I'd be really keen to see how you do a face swap. Take the indian girl's face to replace the face of another picture.
You've done some really good work so far. Well done - and thank you
Getting Error (Getting requirements to build wheel did not run successfully.) When trying to install comfyui-reactor-node.
me, too. Since the new ComfyUi Update reactor does not work anymore.
LivePortrait In ComfyUI - A Hedra AI Alike Talking Avatar In Local PC( tutoriales
at 9:38 when i queue generation the next ksampler errors with 'NoneType' object has no attribute 'shape', anyone with ideas bout possible issue?
well the object has no attribute error seemed to be due to i neglected to use an sdxl check point (oops) but now i am getting mat1 and mat2 shapes cannot be multiplied (16x2048 and 768x320) error, which i believe may be due to non-matching input image dimensions? should we have a node in here to autoresize input images to keep things smooth?
tanx for video . How can I contact you directly?
If you have any questions or problems, please share them with me via comment or email
@@CgTopTips Email?
@@SAMEGAMAN yea, cg.top.tips@gmail.com