Aesthetic Booster And Image Enhancer. Also provides a slight dynamic lighting change to each image as well.
Boogu-Image LoRA works with both Base and Turbo!
HIGHDREAM LORA DOES NOT WORK WITH 2604.
NOTE: TURBO LoRA works with BOTH the Z-Image Base and Turbo models!
Z-IMAGE BASE ONLY works on BASE ONLY!
WARNING:
lora strength can be intense. i suggest 0.2 - 0.8, TRUST ME, its going to be enough and this lora is a multipurpose lora. It will handle most aesthetics you want.
.4 - .7 is the sweetspot in my opinion.
drastically improves on seed variation. youll notice details enhance on most images as well.
test it and show me some of your generations!
https://youtu.be/mFjYII4-yCk?si=n6wvk6N78tNPOMRj
Description
works with both Turbo and Base
FAQ
Comments (15)
The "BASE ONLY works on BASE ONLY!" warning is for Z-image only, right, not for Boogu?
What is Boogu? And how does it to manage to have category?
@HugMeIntoFace It's a new image model suite, with its own Edit variant... looks promising, but it's still too new.
@ArtiOtaku That does not answer how they get a category. Who is behind them that they.... nevermind. Thanks for the reply. :)
@HugMeIntoFace wtf
@HugMeIntoFace LOL, I was so wrong, pretty much everyone jumped ship from poor Boogu now that Krea2 is out and Krea2Edit seems to be in the horizon.
@ArtiOtaku Yeah, I am just was not seeing anything in Boogu that would make it interesting enough for the future. But its make courage to go to new things (we did burn ourselfs few times). We will see what Krea2 will offer it seems popular now. (Even the Ideogram seems in the shadow of Krea2) Thanks for reply. Enjoy the AI ride. :)
@HugMeIntoFace LOL, true, I was gonna try ideogram4, but after Krea2 came out, i fully forgot about it.
@ArtiOtaku ideogram is cool forsure, very structured control tbh. but the outputs.... idk man they just seem off in some weird way i cant explain. like you can tell every image is AI without question.
@realrebelai Yeah for me it seems when you push the Ideogram into some kind of composition, the picture looks more like photoshop collage. It is funny, sometimes. :-D
Does this fix the extra limbs/legs/hands issue on Flux 9b (non-base) variant?
Not trained on it but im sure it helps as the dataset has good anatomy captioning 🤷♂️ let me know if you test it for this purpose!
Not trained on it but im sure it helps as the dataset has good anatomy captioning 🤷♂️ let me know if you test it for this purpose!
it looks interesting, would a illust or pony version be possible?
ill get right on it 🫡