The Illustrious version of a realistic merge designed for generating good results from NSFW concept LoRAs and transformations. I personally use this model to test out my IL LoRAs. I haven't been able to put much time into the merges yet - I'm mostly posting this just so my LoRA generations have a base model to show.
Generating
I found the best results with forge to be:
txt2img:
864x1152
cfg: 8
sampler: euler a
scheduler: beta
steps: 30
img2img:
sampler: DPM++ SDE
scheduler: beta
steps: 30
resize by 1.5x
denoise 0.15
cfg 6
Script: SD Upscale -> ESRGAN 4x+ @ scale factor 2
With this flow, img2img is absolutely necessary to get a good result. Do not skip it!
Tips
Using "anthro", "furry", or "scalie" will make the results less human.
Specifying a hair color, facial features, or expressions (like "ahegao" or "smiling") will make the results more human.
Description
FAQ
Comments (5)
WOW. Illustriousversion of transformix! thank you sir for always. May I ask which trainer do you use for checkpoint and loras? and I'm curious if you have a plan to make donation sites such as patreon or fansly.
Aw man. It's been way too long since seeing any new stuff from yah. Nice to find your model in a new Illustrious flavor.
Amazing—just generating normally gives such a wide variety of outputs. And they’re all realistic and beautiful!
Thank you for sharing this.
Ironicly, i've actually found this checkpoint to be pretty bad when paired with transformation based Loras.
Are there any plans to improve compatibility with Loras?
this is one of my top 3 model