Hi, Andrew.
First, thanks for putting this course and surrounding content together. It’s been super helpful!
I’m mapping out a manga-style webtoon and as one of the steps, I’m trying to get some different poses using ControlNet(canny & openpose), however, no matter what settings/models I use, A1111 doesn’t seem to even acknowledge any ControlNet I apply to my SDXL models (with the exception of InstantID). I really like the SDXL/Lora combo I have so I’m hoping I can get the ControlNets to respond.
All that to say, do I need to have a preprocessor selected when using canny/openpose? Or is there anything that comes to mind that might be causing A1111 to basically ignore my ControlNets?
If there’s any info or screenshots that might help pinpoint this issue, let me know and I’ll post those as well.
Prompt settings:
Checkpoint: AlbedobaseXL_v21
Denoising .75
Clip skip: 2
CFG: 2
Steps: 40
Sampler/Scheduler: DPM++ 2M, Karras
Embeddings: easynegative, badhandv4, By bad artist
ControlNet 0: pre: instant_ID_face_keypoints, control_instant_id_sdxl
ControlNet 1: pre: canny, kohya_controllllite_xl_canny_anime (I have tried weights between .75-1.0)
Lora: MJ Manga, more_details
Thanks again!