doopaNdv

8 months ago

Any chance we'll get Z-Image-Turbo model? Or are you waiting for some decent community finetune (and they in turn wait for full model release cuz training on distilled model is shit) btw reddit login is broken :/

pear-shaped

7 months ago

W update

a

Joul

6 months ago

Hey ! Apologies for the slow update, things have been busy behind the scenes. We're working on a big release for gen-image, and Z-image-turbo is coming very soon, along with a lot more exciting additions

doopaNdv

6 months ago

Hey, so since will be adding new models soon anyway, is there any chance we can get RouWei 0.8 vpred model as well? It's anime sdxl model that's newer than hassaku, it's really good at different artstyles (it has bunch of custom tags for them as well), it's one of the few models that can actually change brightness of generated image. It also has very good (for CLIP based model at least) natural language prompting. And most important for you as a host, it has the same license as hassaku (Fair AI Public License 1.0-SD). If your infra can work with vpred models, then that's better (vpred models have better colors/contrast etc.), but epsilon models are good too. link: civitai.com/models/950531/rouwei

doopaNdv

6 months ago

ah shit, newlines are broken in comments so it's one big block of text now :/ sorry!

..

,,

Joul

5 months ago

I will add Rouwei to the list. Hope to finish the migration this week, and a bunch of new models will be available.