EnvyBetterHands_LoCon_Alpha-3

(Note: This advice is for RevAnimated 1.2. YMMV with other models.)

I lowered the learning rate a bit and added a bunch of new images to the dataset and have had another jump in quality. In particular, it seems to be getting better at rendering hands that are holding things or interacting with objects, which have never really worked before.

This model is getting stronger, and I'm finding that it tends to overcook things at strength 1.0 (at least in RevAnimated), but you can fix that by using the dynamic thresholding extension and setting it to mimic a low CFG value (like 6 or so). It overcooks things a bit, but you need the strength set to 1.0 for it to really work well. You can work around this by reducing the CFG value to 5 or 6 or so. I've had good luck with enabling the dynamic thresholding extension and setting it to mimic CFG 5, and then I can set my CFG value to 9 or 10 and things come out fine.

I tried using it with another LoRA and got some pretty strange results, so YMMV there as well. Right now I'm just trying to get it to work consistently in a simple use case.

Oddly, I think it's regressed a bit on hands in neutral positions, but it's noticeably better at more complicated interactions, such as holding objects (which is why I have so many pictures of blacksmiths and librarians in the example images).

Keep your prompts simple and it tends to do better.

With RevAnimated, I tend to get 1 or 2 usable images out of every 8, with a bunch of other ones that are pretty close and can probably be fixed with inpainting.

Civitai link: https://civitai.com/models/47085?modelVersionId=53396