![](https://lemmy.world/pictrs/image/4209ac05-210c-4de2-9eac-d9f830b1cefc.png)
![](https://lemmy.world/pictrs/image/9c06e075-40ab-45b3-af65-ab50e55d3d60.jpeg)
Hey, I thought your post lent itself well to using img2img stable diffusion to create a more detailed version of your 7yo’s drawing and story.
By far, the best result came when I cranked the de-noising level up, but it’s further removed from the original drawing. I used the story as the prompt to give the model (Dreamshaper XL, a brand new SDXL checkpoint) more information.
Here’s an album including the ones with lower denoising values to line up more with the original sketch: https://imgur.com/a/hH9pYE5
I like video games, goofy stuff, and tech. So lately, it’s been: