Def skim the top rated civitai models with a common workflow for it you find online if you use comfyui. Also, if you want the best quality, check out hiresfix and skim through the shared pictures to find out your ideal settings range for the sampler and stuff.
I can recommend SDXL workflows & models. Turbo is very good but it’s still a bit unstable. Expect it to get more stable top rated models in like 2-3 months.
The issue isn’t with generating good pictures in general. I’ve been doing this a while and have gone through a fair number of models.
The issue is getting it to generate human-animal hybrids that look good. I tend to get a lot of things similar to OP’s “first attempt” or creatures that are more on the animal side.
So that’s why I was hoping OP could give more specific tips about going from where they started to where they ended up.
If I were going to try for this, I would first try YiffyMix for the first pass, then see if I can get good results using a photorealistic model on a second pass with Hires fix or img2img. If that doesn’t do it, I bet that with regional prompter could make it work. I’m not as good with that though, so I usually only resort to it if I can’t get what I need another way.
e: I’m talking local stable diffusion here, usually A1111
That’s a good idea. I’m not looking for photo real necessarily, but not quite as furry-lover either lol.
When I was looking at YiffyMix I found a “WereAnimal” LoRa and started playing with that. Initial results seemed good then started going downhill. I’ve started using ComfyUI since getting back into this and I’m still learning my way around the workflows.
Got any tips? I’ve been trying to make a couple of human-animal hybrids and get a lot of crap.
Yiffymix?
Thanks, a little more… Anime?.. then I was aiming for, but I’ll give it a shot.
Def skim the top rated civitai models with a common workflow for it you find online if you use comfyui. Also, if you want the best quality, check out hiresfix and skim through the shared pictures to find out your ideal settings range for the sampler and stuff.
I can recommend SDXL workflows & models. Turbo is very good but it’s still a bit unstable. Expect it to get more stable top rated models in like 2-3 months.
The issue isn’t with generating good pictures in general. I’ve been doing this a while and have gone through a fair number of models.
The issue is getting it to generate human-animal hybrids that look good. I tend to get a lot of things similar to OP’s “first attempt” or creatures that are more on the animal side.
So that’s why I was hoping OP could give more specific tips about going from where they started to where they ended up.
First image was done by deepai, and the second one with Bing ai. Prompt is simply “coyote businessmen”.
Haha, okay, thanks!
If I were going to try for this, I would first try YiffyMix for the first pass, then see if I can get good results using a photorealistic model on a second pass with Hires fix or img2img. If that doesn’t do it, I bet that with regional prompter could make it work. I’m not as good with that though, so I usually only resort to it if I can’t get what I need another way.
e: I’m talking local stable diffusion here, usually A1111
That’s a good idea. I’m not looking for photo real necessarily, but not quite as furry-lover either lol.
When I was looking at YiffyMix I found a “WereAnimal” LoRa and started playing with that. Initial results seemed good then started going downhill. I’ve started using ComfyUI since getting back into this and I’m still learning my way around the workflows.
I’ve gotten good results on civitai - combine,choose your settings and go. It’s very resourceful.