Thank you for sharing this great repo. I really like it.
I am trying to train my own controlnet model. The model works well after 15k iterations, but results worsen after 30k iterations. This happens in many different condition types (edge, depth). Could you tell me why this happens? Is there a way to fix it?
geroldmeisinger
commented
1 year ago
Dear Lvmin,
Thank you for sharing this great repo. I really like it.
I am trying to train my own controlnet model. The model works well after 15k iterations, but results worsen after 30k iterations. This happens in many different condition types (edge, depth). Could you tell me why this happens? Is there a way to fix it?
Thank you for your help.
Best Wishes,
Zongze