I had tried a similar prompt to evaluate Dall-E earlier. So wanted to see how Stable Diffusion worked for it. The difference here is that all these images were created locally on my machine (not in the cloud). It’s getting easier and easier to run advanced AI models with limited technical knowledge thanks to work by people like Divam Gupta whose DiffusionBee is what I used for this.
Here’s a few of them scaled up for a closer look. Which one do you like the most?