Create Your Own Images of Radiant Light
Create your own inspiring images of light and hope with our AI image generator. It's free and easy to use!
4 years ago
And gave light to all of us!
Created 4 years ago · 11 comments· 0 likes
Stable Diffusion 1.511.4x
Version 1:
Creation Summary by Vincent
Jesus and the Black Sun: a Dark Art Piece
A surreal image of Jesus silhouetted against a dark sun over a cornfield. The artwork evokes a horror aesthetic reminiscent of Doré, Rutkowski, and Giger.
Created by James Corn on Aug 26, 2022 using the Stable Diffusion 1.5 AI image generator model.
Progress Images
11 Comments
Join the conversation
Rocko42 @Rocko42
Thanks for sharing your prompt. Its a great help for learning.
The same settings and seed created an almost exact replica, not upscaled 12x of course. But I notice that changing the size from low res to thumbnail actually creates a Wildly different image. Is that expected?
I thought thumbnail would simply produce a less detailed lower resolution image. Not quite sure why changing to thumbnail, but keeping all other settings and the random seed, Should really do that.
Do you think the training data for the AI is somehow, like Keyed to certain resolutions? Like it polls an entire body of thumbnail images for thumbnails and low res for low res? That seems sort of janky.
It also implies that the low res and med res image libraries are probably much better to use. Or a pay wall.
Makes me wonder what the med res option would produce. Anyway, very cool image. You hit the jackpot with that seed number. I tried a few others to test it- they werent nearly as cool as yours
2022-08-26T09:41:09.775ZReply
Creation Settings
- Text Prompts
- Shhhh, It will hear us, detailed high quality concept art of Jesus hovering infront of the huge black sun above cornfields, horror Gustave Doré Greg Rutkowski, 16k uhd trending on artstation by Shaddy Safadi ,Ferdinand Knab and H.R. GigerWeight: 1
- Model
- Initial Resolution
Low
- Aspect Ratio
4:3
- Runtime
Short
- Overall Prompt Weight
50%
- Sampling method
K_LMS
