Nova Animal XL
A realistic furry checkpoint model that can have great details on any type of furs, scales and feathers, which aims to be XL-lized Persika Furry
Rules
You cannot use the generated images for commercial use if it's not edited (or just turning it to black and white)
Advertising is always welcome
Recommended settings
Sampler: Euler a
Steps: 20-40
(Pony)
CFG Scale: 6-7
Prompt: score_9, score_8_up, score_7_up, BREAK, furry, anthro, detailed face eyes and fur, {Animal Type}
Negative Prompt: score_6,score_5,score_4, source_anime,source_cartoon, 2d, art, cartoon, blurry, jpeg artifacts, username, watermark, signature, normal quality, worst quality, large head, low quality, text, error, missing fingers, extra digits, fewer digits, bad eye
(Illustrious)
CFG Scale: 3-6
Prompt: masterpiece, best quality, amazing quality, very aesthetic, high resolution, ultra-detailed, absurdres, scenery, photo, realistic, furry, anthro, {Prompt}, BREAK, depth of field, detailed fluffy fur, photorealistic details
Negative Prompt: smooth skin, human, multiple tails, 2d, art, old, oldest, cartoon, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured, long body, lowres, bad anatomy, bad hands, missing fingers, extra digits, fewer digits, very displeasing, (worst quality, bad quality:1.2), bad anatomy, sketch, jpeg artifacts, signature, watermark, username, simple background, conjoined, bad ai-generated
Description
Changes
Lower contrast for better realism
Style closer to Pony versions
Controlled fur details
NoobAI EPS v1.1 + Illustrious v2.0-stable DARE applied
FAQ
Comments (16)
Why are models so (anthro male) heavily nipple bias now? Has to be some poisoning going on with female nipples/breasts being so aggressively dominant in them now?
With the last few models, I'm having to put a lot of this in the neg. prompt; "wide body, wide hips, big body, female, girl, woman, big ass, wide ass,"
- And that only helps a tiny bit.
What I'm getting at, is that it seems more, and more biased towards women, and big bodies. This is for sure my favorite model, so it's too bad!
Go straight lol
@PolskaPshe Can't. I like dicks too much
@Foxdude lmao
amazing model!! I love the detail and realism! (I think v11.0 is more realistic than v12.0))...
a question: can I use e621 tags in addition to Danbooru? It would be great if I could use them...
Time to make a Furry model! This give me this feeling!
So after testing all models upto 11, I ended up with 11 vs 8 and decided to go V8 for finetuning a character. (likeness is great btw)
I use this only for non furry purposes btw but I like the flexibility of furry models. It came out really great. Great (semi-)realism while still being very flexible.
I found that while Euler A produces nice clean images, Restart is vastly superior for textures and also gives good image quality without producing too much errors. (Euler and DPM++ 2m being a middle ground if you want faster generations)
So I usually use Euler A for initial image and Restart for img2img upscaling (2 steps of x1.4 or 3x 1.26 works really well)
Using Stable_Yogis_Illustrious TI's and the recommended pos/neg prompt and adding just a few detail/lighting tags with a detailer lora seems to work really well,
v12 seems to be much more 2d styled and less photorealistic than v11
I still prefer this version over v11...
The facial anatomy in the previous version is uglier imo 👀
I really like that. When they're too realistic they just look ugly.
good to hear Crody is still keeping a balance, too photoreal and you get what every single photorealistic slopmerge looks like for illustrious and noob lol.
god please bring the magic of nova animal over to Anima lol
After comparing V9, V10, V11 and V12. I feel like overall, V9 is my favorite, V10 is similar while V10 starts to add too much detail where it's not needed. V9 and V10 are both excellent and have a good mix of realism and contrast.
At least when I run the exact same prompt through all 4 checkpoints, V12 is the one that comes out visibly washed out and it has lost a lot of texture detail. I have a Lora I've made, which I'm not sure I want to post online, but if I do I would personally recommend people use V9 and V10 on it.
in your opinion, which one is better in terms of details and realism? V10 or V11?
@egroj992 Honestly. I only use these with Loras I've made myself and judge them through that lens. It's better if you compare the two yourself. I do seem to remember that V11 does too much with the image. Even just a plain bedroom came across as messy, with too much added detail while 9 and 10 kept things mostly clean and aesthetically pleasing.