Please post the first picture directly here in the Model Gallery. This is the best way to support creators :)
Hello ♥
for whatever reason you want to show me appreciation, you can: ❤️Ko-Fi❤️
_________________
This Lora is a collaboration with the great @bo4rsworld who had the idea for this Lora.
Please check the 2nd Lora from this collaboration
Edit August 16th
Added Flux version. U can use this workflow
Please use at least one of the training words in your prompt, in other case there is a chance to get no effect.
___________________
This Lora generates images that remind you of Once Human.
It gives you that unique color style and creates cool gaming scenes
Trigger
Once Human
Training words
gaming, outdoors, giant, monster, spider
Hope you get a lot of beautiful results ♥
Description
v0.1
FAQ
Comments (5)
this is so incredibly sick!
Thanks for this project, I'm sooooo excited <3
i wonder how big the image model has to be to generate every single frame for an actual neurally generated game..
I bet a 40B image diffusion model could be enough if it was trained on images "that contained tables of frames". That way every render could yield 30 frames each. And thereby it could inference gameplay by just using previous image, since it contains the memory based on 30 earlier frames..
With a good frame-interpolation i bet you could crank it up to much higher "frame tables" per image, maybe even 300 frames @ 256x256 per frame, in a 4K image.. Imagine that a neurally generated game , in 256x256 pixel resolution but still a fully neural game :D Literally the matrix lmao.
Positive ai :D
Nice one
very fing good