Niji Noob
NoobAI XL v-pred model (requires different settings than usual to get the most of it, see below)
Merged by dumping web UI model from memory to disk
My contribution: I've balanced the painterly look with coherence and used some of my own LoRA's for adjustment.
Settings:
Steps: 14, Sampler: DPM++ SDE CFG++, Schedule type: SGM Uniform, CFG scale: 1.7, Size: 832x1216Tool: Panchovix/stable-diffusion-webui-reForge
Tips:
This is a v-pred model so CFG needs to be very low depending on the sampler/scheduler or you will fry your gens.
Lower CFG (e.g. 1.1) will produce less saturated results. I like my gens saturated so I use 1.7.
You really only need 15 steps at most, less for hiresfix.
reForge has the best memory management of all the SD webUI's so I can output 1664 x 2432 no worries on a GTX 3090 (using 2x on hiresfix).
Use NoobAI artist tags if you don't like / want to adjust the default look.
Non v-pred (as in normal) LoRAs works fine (tbh even better in most cases).
If you get tired of circles in backgrounds the solution is simple: put "tunnel" in your neg prompt.
For more Niji-like output use simple prompts and let the model steer your gens
Want to merge your own similar model, check out the LoRA's used by exlight: AIBooru
Also check out LyiaEngine for CivitAI backups of the above: LyliaEngine (Narontaka)
Description
FAQ
Comments (5)
Sorry for spamming your gallery, but I just really love your checkpoint!
It follows prompts so well, and the style is beautiful, it has this rough, sandy texture that I absolutely adore. I promise I’ll stop soon….....maybe.
I know I already left a comment here, but I just had to write another one!
After a few days (well, exactly 18 days) of playing around with this checkpoint, I can honestly say it’s the only one that truly clicks with me. The style feels perfect if you’re looking for something a bit... arty, a bit rough and raw, like, I don’t even know how to describe it properly, but somehow it makes the images feel less AI-ish (I know, it sounds crazy, but it really does!).
The second thing I love most about this checkpoint is its creativity. Sometimes I just throw in random words, and I’m honestly surprised by the results. I don’t think I’ve generated a single bad image with this one.
Of course, it’s not 100% perfect, wide-angle or top-down shots can mess up the hands sometimes. But honestly? I don’t even mind. That imperfection feels like part of its charm and fits its unique style perfectly.
I just have to say no matter how many Illustrious models or LoRAs I test, I always end up coming back to this checkpoint. Even without a specific LoRA for it, I think it doesn’t need one at all, it completely holds its own.
Thank you to the OP for letting me spam your gallery!
I seriously recommend this checkpoint to everyone. Trust me, it’s absolutely worth playing around with.
Here I am (again), alone in the gallery, spamming image after image because I just can’t get enough. I feel like I’m the only one shouting into the void about how great this checkpoint is. I honestly cannot fathom why this checkpoint isn't more widely known and used!
I swear, this model reads prompts like a mind reader, it just gets me. To be honest, this is the most perfect checkpoint I’ve ever used.
I’ve tried many models, but I always end up coming back to Niji Noob. I guess first love really never fades.
this checkpoint has very unique features and the pictures it produces have a great texture，Many other checkpoints don't have such a feature. Every time I produce an image, I want to try this one again. It's really great！！！