Recommended strength and guidance end when chaining multiple controlnets
#7
by
jingno123
- opened
Hi, thanks for the recommended strength and guidance, it's very helpful. Wondering if there's any recommendation when chaining multiple controlnets together? Like if I use depth then canny, should the guidance end be 0.8 and half for depth and half for canny?
Thanks!
oh I saw in the Multi-Inference example code. It should be the strength adds up to about 0.7, but both guidance end will be 0.8?
For Canny, do you have a recommended thresholds?