An experimental build of Zimage Turbo tuned for deeper moods, darker scenes, and smoother handling of challenging low-light imagery. This version pushes contrast, vibe, and atmosphere while keeping the core model’s strengths intact. Still a WIP — try it out and share your feedback so the next evolution hits even harder.
For best results, use the same workflow I used for the preview images — it’s faster, sharper, and pulls out way more detail. If you need help, support, or just wanna vibe with the community, feel free to join my Discord.
Get the GGUF versions Q4 (3.3 GB), Q5 (4.1 GB), Fp16 and ZIT Pro Workflows from my Patreon
Description
FAQ
Comments (15)
its not for ComfyUI right?
It works on Comfy and Forge Neo.
@Stable_Yogi sorry to bother you but I'll be damned if I could've make it work in Forge Neo, no problems in Comfy (so Text Encoders, VAE and all the support should be correct, still: RuntimeError: mat1 and mat2 shapes cannot be multiplied (3072x64 and 256x960)
I'm definitely mssing something :(
@mzaina783 Please use chatgpt for troubleshooting, give it full terminal log, it could take few minutes.
@Stable_Yogi first of all thanks for your time, then, sure I will, I suspect the issue is more on Neo Forge, lately has weird behaviour even with Kontext which was running perfectly, sorry if I looked like "yet another ranting dude". Btw your models ROCK! Ciao.
Wow! This brings zit to another level! I was about to give up with z image and then I found your model. Thanks a lot!
U don't like porn master zit ?
I'm sorry for this noob question - will ZIT models work on SD WebUI?
If you mean Automatic1111, nope, it doesn't.
@Dernet yes I meant that, thank you for responding 🙌
@bhookhabudha Use Forge Neo, it works.
可以在fogre上使用，有个中国人一直在跟新forge版本的，兼容所有主流模型，但他说的是中文，我不知道英文社区有没有。
Is this trained or a merge? Cause it feel like a merge when I use loras...
Lora's trained and Merged.
"Trained and merged" is not "Model Trained" wtf.
This is just a sloppy model merge :`D