Kenmac opened this issue on Jul 03, 2023 ยท 18 posts
Kenmac posted Mon, 03 July 2023 at 11:44 AM
I just thought I'd talk about and show you some examples of the new Stable Diffusion XL 0.9. First things first, you can try it out at https://clipdrop.co/stable-diffusion You can use it for free but you only get four or five images, however if you sign in with your e-mail you can generate as many images as you want. They've included new styles with it and I have to say this version is a pretty big leap forward. The image quality and coherency is greatly improved. Sure it'll still make the odd mistake once in a while but it doesn't happen quite as often. For being a beta version I'm already pretty impressed. The final version will be coming out later this month. I'll be listing some of the prompts and styles that I've used to make these images. I didn't use all the styles but I did use most of them. BTW no processing was used with these images. The only thing I did was to reduce the resolution from 2048X2048 to 1024X1024.
Prompt: Spiderman sitting on top of building looking down.
Style: Comic Book

Prompt: Superman flying over metropoiis at dusk.
Style: Comic Book

Prompt: man and woman laughing while sitting at restaurant table with food
Style: Photographic

Prompt: Same as above but when you use the analog film style it changes them to your aunt and uncle from 1973 :o)
Style: Analog Film

Prompt: punk girl standing outside building
Style: Neon Punk

Prompt: man and woman walking hand in hand on the beach
Style: No Style

Prompt: smiling anime girl
Style: Anime

Prompt: man waving hello to a friend outside camera range
Style: Line Art

Prompt: sailing ship on rough waves
Style: Cinematic

Prompt: furry lion in the woods
Style: Fantasy Art

Prompt: tiger chasing a bird in the woods
Style: Origami

This next one shows that, for the most part, text will work with the images. It may take a couple of tries but it works.
Prompt: anthromorphic mouse holding up a sign reading "cheese please!"
Style: Fantasy Art

These next ones, to me, came out exceptionally well.
Prompt: maine coon cat lounging on black leather sofa
Style: Photographic

Now, I know there are people who feel that A.I. generated images are "soulless" but my answer to them are these two images. (I chose two because I like the way both of them turned out.)
Prompt for both images: woman holding her newborn baby
Style: Photographic


I'd also like to point you to a couple of A.I. oriented videos from YouTube. First there's MattVidPro who covers everything A.I. and he seems to really like SD XL 0.9 and he gives it a fair comparison to Midjourney. https://www.youtube.com/watch?v=9VUBlCv6Z8g
Then there's AItrepeneur who gives it a very good review: https://www.youtube.com/watch?v=XR5BMIYaxQ8
That's it. Hope you liked the images and the YouTube videos and maybe you'll give this a try.
I realize there's a lot of images here but I just wanted to show some of the variety you can get with the new Stable Diffusiion XL.