I guess I'll occasionally post scraps that aren't so bad that I can't delete them right away. I just need to come up with a hashtag so I don't have to justify myself every time.
#GarbageGem ? #CherryChunks ?
@TarkanSarim This is interesting. I once tried a similar method with AnimateAnyone, but it didn't work well because of the jagged movements. I see a similar problem here, but you have refined it almost to perfection. Nice job!
We throw away thousands of nice images in the process of experimenting with different weights. And I often catch myself thinking how a year ago I would have carefully saved such things for reference or inspiration, but now... they are just like industrial waste. It's so weird.
Digging into instant facial recognition adapters. Quite surprised how a single photo can give a better result than a fully trained LoRa.
#AI#AIArtwork#StableDiffusion#digitalart#Art
Was only planning on taking 20 seconds, but yeah.. ended up rendering a full dance. I hope you don't find it too stretched out.
#StableDiffusion#AIArtwork#digitalart#Art
https://t.co/YU7sC01s1r
I expected to have a small piece of the clip done by this weekend, because the original video was extremely massive. Almost 5000 frames, 3 minutes long and 20 hours of render time at best.
But... I'm not too smart to stop in time, and we're already here, so...
One day to go.
It's a bit rushed but I'll post it today.
I've learned my mistakes from working on the last long video. 2K resolution, more dynamic backgrounds, less glitches.
#StableDiffusion#AIArtwork#digitalart#Art
https://t.co/vkExgOLR6Y