*Sup*
Not gonna paste some elaborate AI-generated description here full of bullshit about how cutting-edge this model is or try to steal your buzz.
(but if you wanna donate some that's dope, it does help with all the character trainings I've got in the works.)
Update 7/22- Krea2 -
Import my showcase images for a workflow I've been working on. Tried making it easy to figure out, message me if you have any issues with it.
(also the arrows don't line up until you generate one image with it because the preview should pop in at the bottom, but after that they should point to things correctly.)
Works best for me at high res - 1440x1920 is my go-to.
Sampler - Euler
Scheduler - Beta
CFG - 1
Steps - 9
My goal was simple: Character LoRA compatibility.
Honestly, I'm mostly uploading this because it's annoying seeing so many models posted for stupid high costs. So if you make something cool with it, I'd appreciate you posting your generations and helping get it some exposure so I can compete with the giant models milking people for buzz with new bunk versions every week.
Z-Turbo specific below:
_______________________
It generates NSFW content pretty damn well with the right prompting.
I'm honestly pretty happy with where it's at, and I don't want to slowly destroy the LoRA compatibility just to keep pumping out new versions to make buzz. But we'll see.
ANYWAYS,
Highly recommend trying the two other Text Encoders on CivitAI. I'll add them to the suggested resources. They can make a surprisingly big difference sometimes. Some of my examples use them, but most were generated with the default Qwen 3_4b.
THM_UNC_Z-image_TE
Z_Image_TE_SoLordZ
Description
Krea2.
FAQ
Comments (31)
You use my Lora thks mate! Your model work fine with it, good jobs
oh yeah man, all of yours are really solid!! had to showcase some
please consider uploading either an nf4 version, or the bf16 version so we can convert ourselves.
bf16 already uploading (taking forever though) and I'm looking into how to convert to other forms also!
@sickollie you weren't lying, still not there😭, the hype is real
@Neon_signs haha yeah I got to like 80% and then the download failed and cancelled for some reason. Waiting a bit to restart it but I'll have them uploaded tonight for sure. I've got BF16, mxfp8, and nvfp4 ready to upload.
@sickollie damn that makes sense, make sure to focus on that bf16 please
@Neon_signs finally got the BF16 uploaded!
Your Zit model was the best so I am really happy you are doing one for Krea2 as well - looking forward to testing it! Thank you!
I hope you'll consider also doing a Krea2 version of you character LoRAs!
Hey thank you! Also I am, plus many new ones I never got the chance to release on Z-IT.
have several retraining right now and a few almost done testing.
@sickollie Thanks a lot looking forward to it!
Krea2: Super clean results, loving this! Works perfectly with character LORAs, honestly better than the native model in a lot of cases. Would love a convrot version but this is still pretty fast. You fricken rock!
My fav ZIT model by far ! Can't wait for the int8 on Krea2 to play with it !
that should be in the download variants section now!
If you createBF16, any chance you can upload to HF too? I find civitai downloads crap out for files over 20GB, usually at 95% (repeatedly, and don't resume!). Thanks so much for making Krea2 version, your ZIT model was superb.
Thanks for BF16. I got it downloaded. No that anyone will see this, but I tried setting firefox (about:config) setting 'network.http.http3.enable' to 'false' as I wondered if QUIC/http3 might expose some odd issue. Lo and behold - one complete BF16 of your model. Delighted, thanks.
@dikaxa8718570 oh awesome! I just woke up and was planning to go make a HF page to upload it haha. I may still do that, it took me a couple tries to get the BF16 uploaded because it would fail around 80%.
i love it, works well with my own face lora, even better than the og.
tested with krea2 int8 on my rtx 4060 ti,
4m 17s.
Even though I'm skeptical about downloading any models other than the base model (because I believe 99% of the people who share models don't really know what they're doing they just throw stuff together without any real knowledge, creating a huge mess, adding biases that then become difficult or even impossible to remove later, especially when other people merge those models with others. You completely lose track of everything, it's always chaos, and you end up with "fried" models that generate images covered in spots, the classic sign of overfitting) but what was I saying? Oh yeah.
You're the only person who actually convinces me to download a model that's not the base one. You're one of the very few I can count on one hand, one of the few who genuinely has good taste, both aesthetically and when it comes to girls. There's way too much garbage out there with these obese pigs that look like water heaters. At least you have taste.
Thanks. I'll try this today.
Couldn't agree more.
First comment I've ever left here. I loved your ZiT model, and your Krea version is even better. It produces beautiful images.
I've been messing with merging models the past few days to some moderate success because no model really did what I want right out of the gate. I really like Kreamania 3's grit/grain and lighting, but it's such a censored model that it requires bypassing which I hate using. Even merging it with others tends to retain that censorship. This model, however, I've loved right out of the gate. I still want just a bit of that grain and darker lighting, but that's something so low-strength loras can add. This model is truly great and absolutely works with my character loras. Great job :)
Wake up babe, Sickollie posted a Krea2 Checkpoint!
AWESOME! The best model in Z and now the best model on Krea2...thanks Ollie!
Completely insane model. Kudos.
love this model
Real good.Very impressive bro.
I keep getting abstract images when I attempt to run it. I can't figure out what's causing this, I have the right models downloaded. It's strange.
Hey I'm not sure! I have only used the fp8 version myself, which version are you trying to use? If you drop one of my showcase images into Comfyui it will import my workflow with all of my working settings with the fp8 version.