Anyone trying to make art using AI?

I’ve been playing with stable diffusion over the past few months in my spare time. Mostly trying to get things like blog images. Here are a few of my favorites.

I have all of my diffusions and their prompts on stable-diffusion.waylonwalker.com/

2 Likes

I haven’t spent a lot of time with Stable Diffusion (yet!) but I’d love to know a bit more about your process. for example, what advice do you have about creating prompts that helps to get the images you’re looking for?

I’ve been starting by browsing https://lexica.art/ for the style that I am looking for, and I copy the prompt and swap out the context for my own. The biggest advantage that I see to stable diffusion is that I can run it on my own hardware. I will often make a few feeler prompts trying different words until I get something that is in the ballpark of what I am looking for, then I ask it to make 25-50 with the same prompt. My friend that helped me get started has had better success using higher C values. I don’t exactly know what C does, but its supposed to be how much creativity you allow it to divert from your prompt. Even things like aspect ratio can really change the outcome.

At the current state, I don’t think that you can get exactly what you want if you have something specific in mind, but it’s able to make some really cool artwork very quickly within the ballpark of what you have requested.

I did find the DallE prompt book very helpful as well to start to understand what type of things to describe in the prompt, not only setting and style, but also things like focus and crop. The DALL·E 2 Prompt Book – DALL·Ery GALL·Ery

For me its been mostly replicating cool prompts from others in my own way.

1 Like

this is so helpful, and so much more detailed than what I imagined! no wonder you’re able to get such solid results. and thank you for sharing both your approach and the DALL-E 2 prompt book - I’m not even further behind in my reading list but in the best possible way!

1 Like