//=time() ?>
Scientists:
- We've spent years to solve the task of reconstructing and translating texts in any language on Earth, even the dead ones. Whew, we can finally get some rest!
Text2image models:
- Hold my beer
@Zugani_nft @zippy731 @autodesk Not yet, the one there is turbo, without smoothing. If you need turbosmoothed version, I'd suggest waiting a bit.
@remi_molettee Outstanding! Finally, someone is putting this tech to some good use 😎
I'm working on turbo-smoothed version so that you can only diffuse every nth frame, and it looks much smoother, like this one here.
It has consistency check nearly disabled here (for a very subtle fade in)
I'm absolutely in love with how @DaviddelosSan breathtaking timelapse from @NatGeo 's film festival looks with DD+Warp
Take a moment and appreciate the small things.
Turbo Warp fork of @Somnai_dreams #discodiffusion
Easter Bunnies and their Eggs are so outdated! Meet Easter Dragon Eggs! (and some hatchlings)
#discodiffusion turbo settings by @zippy #aiart
The 3rd comic faces model version is in early access RN!
#discodiffusion #aiart #portrait
More on it here https://t.co/XitCofrINu
Public model version:
https://t.co/wP41LcLnz9
Can't say no to all those cyberpunk haircuts! #Discodiffusion warp custom faces model
Okay, I've got the weirdest idea. What if we do an #nft of a pre-trained generative model's weights?
With all its derivatives having to point back to the model's #nft
It's an endless collection by itself.
#discodiffusion