//=time() ?>
@DuncanTiminey k_euler_ancestral is generally my default favourite. With automatic you can plot an x/y matrix to compare samplers at different steps with seed and all other settings the same.
@AndrewJHoldun @CoffeeVectors @GanWeaving Cool. This is another one I made at the time. I liked it but the styles bled into each other a bit so it didn't demonstrate what I wanted.
Here's another example blending styles:
1.Andy Goldsworthy
2.Van Gogh
3.Andy Goldsworthy and Van Gogh
4.andy goldsworthy :2 AND van gogh
I needed to add weight to Goldsworthy as Van Gogh was a stronger concept.
@eerac Thanks. I found @johnlindquist's secret stash of prompts https://t.co/BsVB4zg2mS and have been testing some out.
This is using SD 1.5 - to be honest I think I prefer the original.
I suspect if I use my concept it will overpower whole scene. Will try it though.
@deforum_art @mubertapp This was the first attempt and yep, not responsive enough and the prompt needed tweaking. It always takes a few tries but it's worth persevering. I'm glad I reworked it as the second one was so much better.
Interesting. I've been playing around with #StableDiffusion guided by Clip H and most of the time there's an improvement in coherence.
Both images used same prompt, seed and settings with SD 1.5 but second uses Clip H.
More settings to learn.
@Mikerhinos Lol. There you go and yep, I mostly look like Trevor!
Out of curiosity (and vanity) I fine-tuned #stablediffusion on 4 photos of myself and it's hilarious😆. The images are uncanny but with the "Tom"ishness turned up. Big nose bigger, big chin bigger etc.
Can imagine training from a metahuman to use as a consistent character.
An experiment with video_init from Little Otik by Jan Švankmajer. In this scene Otik's parents begin to be concerned about his appetites. Prompt: "An aesthetic ultra detailed film still, artwork by studio ghibli and pixar, makoto shinkai, pixv, 8k" #AIart #stablediffusion