NO NSFW! NO PORN! NO SEXY PHOTO!
After download this model, try to share with us about your finding.
Again, not for commercial use and she is not a existing person.
Please use ChilloutMix and based on SD1.5
768,768 image
Use DPM++ 2M Karras or DPM++ SDE Karras
Use 18 sampling steps
Strength: 0.8
Sampling step: 20
Hires. fix: On
Hires step: 17
Denoising: 0.7
Description
FAQ
Comments (4)
very high productivity, can you also share the number of training images you used
Depends on how many photo I can gather, hard to tell. Normally around 40~50 image, 30~40 for face and 10~20 for half body.
It would be nice to see the model in at least 6 photo positions (sitting, standing, running position), and at least with 3 face expressions... normal face... normal "Smile"... and a big happy smile (Smiling face expression showing teeth:1.1).
This way we will know how she will look with diferent faces and body positions.
I like your LORA very much, I also collect them, but I have to test them using checkpoint Realistic-Vision with my own special Prompts just to know how the models look like.
Thank you.
Lora looks fine, although it has a tendency to look extremely young or have some sort of greywashed effect for some reason (abit luck dependent to get her to look like ikuta). Looking forward to a future update to the lora!