Available on PixAi for NSFW on-site "private" generation:
10/04/26: Next version will be a Re-Merge of models with Chenkin v0.5.
Cute Reborn Plus: A merge of Simple Cuteness, Reborn, some LoRas to get close to Animu Plus and a little of Mugen 0.5 too... It works, the quality is ok, but it has problems (complex backgrounds most of the time) due to being a merge of an SD3 model with SDXL. The model tries to adapt to the artstyle of the character, so details and style is not the same for all.
Simple: A simple merge, nothing impresive. Is a test merge of CuteCandyMix, CuteFurryMix, Wai, Janku and my Cyberfix (with Chenkin 0.2 and CyberRealistic v9).
Reborn: An updated, remerged and clean version (no extra LoRas added to the merge) of my Animu Plus version.
Animu Plus Alter: Alternative version, it's a different style than Animu Plus, has more "depth", more details but less "anime coloring" and body can look 2.5d sometimes.
Animu Plus: An upgrade version of my Wai x Janku + NovaMoon-PM ILL XL vZero but now based on NoobAi.
Animu - "2D": A normal anime style on characters but has deep and semirealistic lighting and backgrounds.
You can use Illustrious, SDXL, NAI and some pony LoRas, the quality can vary due to different base training model, so you need to test by yourself.
The model uses booru tag system (1girl/1boy), but you can use descriptive prompt too. NSFW Works
The images i upload use Euler Beta and without using any LoRA, CFG Rescale, Hiresfix, ADetailer or Upscale, so the quality can be improved. Close distance is good; medium is ok and far is ugly. I use Panchovix reForge so the output you can get will be a lot different than mine.
Samplers and Schedulers:
Euler a Normal/Simple/SGM Uniform CFG 3+ for general useEuler Beta/SGM Uniform CFG 3+ for sharper imagesDPM++ 2/3/4M SDE Karras CFG 3+ for more sharpness, details and recommended if you use 3d/"realism" prompts.For Euler/Euler a start with 10steps max 35 steps.For DPM start with 16 steps max 35 steps.Resolution:
Works with all SDXL resolutions, i personally use 1024/1152x1536 and sometimes 1280x1536Recommended base prompts:
(Cute Reborn Plus doesn't need positive quality tags but it can change the image, worse or better, test by yourself)
Positive:
masterpiece, best quality, excellent, high Resolution, aestheticNegative:
worst quality, bad quality, low quality, lowres, low resolution, anatomical nonsense, artistic error, bad anatomy, bad feet, bad hands, extra digits, conjoinedMerge of:
Wai: https://civarchive.com/models/827184?modelVersionId=1761560
Janku: https://civarchive.com/models/1277670?modelVersionId=1864501
MoonMaster: https://civarchive.com/models/1946488/moonmaster-illustrious-model-suite
dims2: https://civarchive.com/models/835655?modelVersionId=1023901
Cyberfix: CyberFixV2
NAI: NoobAI-XL (NAI-XL) - Epsilon-pred 1.1-Version | NoobAI Checkpoint | Civitai
Chenkin: Chenkin Noob XL （CKXL ） - v0.1 | NoobAI Checkpoint | Civitai
atw44qb F2-2 (VAE Animu Plus): XL_VAE_C - F2-2 | Stable Diffusion XL VAE | Civitai
Yiffv63: YiffyMix - v63-NoobXL | NoobAI Checkpoint | Civitai
atw44qb G9 (VAE Animu Plus Alter): XL_VAE_C - G9 | Stable Diffusion XL VAE | Civitai
Description
The base style without any style prompt is the one of all the merged models, looks slightly 2.5D due to Cyberfix and Janku v6.9 but due to being more NAI than Illustrious, it adapts to what you prompt.
Now is a merge of:
Janku V6.9
Wai v16
MoonMaster
MoonToon
Yiffmixv64
YiH xxxtended
Cyberfix (same recipe) made by me using CyberRealistic XL v9 and Chenkin 0.2
XL_VAE_C G9 VAE
Why Illustrious and not NAI? Well, more people know Illustrious, so using it is just to get to more people but is a mix of Illustrious and NAI (and a little of SDXL) like most recent 2D-2.5D models and the CLIP has more NAI than Illustrious.
FAQ
Comments (8)
I used exactly the same prompt and parameters as you did, but the final image I generated doesn't have the same 'grainy/noisy' feeling as your example. Could you please tell me how you achieved that effect?
Edit:
My old answer was wrong, i tested the prompt on ComfyUi and it give the grainy/noisy effect, here you can download the workflow image and test it, for other WebUi i use reForge but i don't know about others, sorry (I haven't changed any system setting to ComfyUi so it uses CPU seed). If even it doesn't show up to you, maybe my GPU is dying and it gives that effect lol
animu plus looked way better than reborn. it has way more detail, a massive difference. hope newer versions will be continuation of animu plus because the difference is not small
Yes, is normal. Animu plus has merged detail LoRas, Reborn is a clean merge for better LoRa and Style compatibility.
@Patuwa what detail loras were they, so i can use them with reborn at same weights and see if the detail goes back to plus
@Patuwa what detail loras please bro
@goeypants400 Sorry, i tried to remember what LoRas i used on the merge but i don't remember all, the ones i remember is CFG Scale Boost, V-Pred Fix (They show as deleted, so you need to find them), Smooth Detailer Booster, Susamix LoRa (i tried to find it but i think it got deleted or they changed the name), Aesthetic Quality Modifier and Anime style mix. I used more (but i don't remember) and some of them was made by me using Pixai but i already delete them some months ago. For weights are low due to on merge the strength is reduced so you need to test them. I don't save the "recipes" of my Merges and since I'm already bored of doing Merges i deleted all, sorry if i can't give you a better answer.
@Patuwa thanks man, i was able to find the first one on civarchive, detail is way better but not as good as plus, and keeps becoming 3d. hope you get inspiration back for some more models eventually, your merges are good