Home Artists Posts Import Register

Content

Thank you all for your patience this past month, got to visit the folks for the holiday, then got caught up with some other irl stuff, but now we're back!

Housekeeping:
I've finished my testing messing around with the cyberpunk model, and am finally moving onto training the hyper model. By the end of it, you should be able to not only pull out hyper content, but specify it *by size* - yes, there are different sizes of hyper- because it's a catch-all term for "Anything outside the plausibility", thats the minimum size where the range starts, and it then it basically just goes on forever. That's the plan, at least, but it's taking a lot of tooling around with the code and trial and error testing to get it behaving just right

Wheres the 3d content?
A side project I'm working on is training a model to replace the render step in Daz3d rendering. Ie: you feed it a 3d viewport, tell it what you want- and it spits out a photoreal image, and is able to be reiterated upon and edited with natural language- all in the 5-10 sec timeframe of a diffusion model instead of the 10 hr per timeframe of iray render engine. The tools are all there, and it's able to be done today, just gotta go through all the nitty gritty of coding it up a custom trainer, custom inference, etc and bringing it to life. All my 3d energy is going into this, because should I be able to bring it to fruition.. man it'd be such a game changer

Files

Comments

No comments found for this post.