This model produces high-quality images in a structured 8-step process.
Description
v2 [XL]
Higher fidelity images
Better prompt adherence
Better eyes and limbs
FAQ
Comments (5)
how many faces?
one should be enough for everyone - Bill Gates
with the right settings i get super, super sharp images, for img2img I use easy diffusion 3.5 with a cfg of 1.8 prompt .3 - 30 steps with lcm beta, and that combo gave me the most detail and creativity it could muster. The only problem is, while this is super sharp, its really lacking creativity. Buildings are just straight up and down, and there are some cute animals and it seems to understand concept OK, maybe a 5 out of 10. That's the main issue with this. But I can get in really tight zooming in on detail. While it does do X rated stuff really well, like I can see goosebumps I think, and many other topics, some areas it could really use more creativity.
My current winner is another lcm and that's - SplashedMixdmd v4. If you could combine the creativity of that with the sharpness of yours, that would be very interesting.
I totally agree! Thanks for taking the time to leave such great feedback, it means a lot! The creativity really is lacking. I think I was focussing too much on aesthetics and totally overlooked so much!
Right now its more like a clip art generator, but when i'm expecting a city in the sky and I get like an urban back alley and without a ton of detail its just a womp womp.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.