this post was submitted on 27 Mar 2025
10 points (100.0% liked)
Stable Diffusion Art
2005 readers
52 users here now
Share art created with Stable Diffusion
Also see:
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Nice, what generation parameters did you use?
That was the same prompt, Euler 25 steps 3.5 cfg seed 42. Though I'm training a lora right now so it's a training sample from like 1700 steps in
The original used quite a few LoRAs to get that look. OP didn't share a prompt, so it's impressive how close that came out with just the image description.
Mine's more to the realistic side. Also, I'm training on a dataset that is like ultra-sharp just to compare how it can learn differently. Think it's just a little too sharp.
I'd like to see it once you're done with it.
Shoot me a prompt and I'll throw it through the test sample
How about:
A surreal artwork depicting a human skull in profile emerging from a fragmented landscape of blue-tinted architectural debris and twisted branches beside vivid red-purple dahlias against a dramatic background of yellow, orange, and pink gradient clouds.
Yeah, i think it's cooked.
This is the four previous checkpoints. I may have gone too far with the sharpening. Looks like a cool album cover though.
Yeah.
It is a little sharp.