The goal of this LoRA is to depict full penile insertion together with the deformation of the surrounding flesh caused by close physical contact.
In other words, it is designed to reproduce effects such as ass ripple, ass press, and pussy press resulting from balls deep.
Initial release notes(v0.0.1)
For v0.0.1, I trained the LoRA on 60 images in total:
50 images showing full insertion to the base from various angles
10 cropped images focusing on the deformation of the surrounding flesh
The overall result feels somewhat like a jack-of-all-trades, master of none.
I compared the results with no LoRA, this LoRA, and LangdonAlger's balls deep LoRA. For full insertion to the base and completely concealing the penis inside the vagina, LangdonAlger's balls deep still appears to perform better.
For the deformation of the surrounding flesh, the Anima base model can sometimes produce interesting results. However, FT models often fail to reproduce the deformation. This is likely due to the insufficient number of cropped close-up images.
For the next version, I plan to significantly increase the proportion of cropped close-up images in the training dataset. Based on what I learned from developing my previous two LoRAs, increasing the proportion of close-up images should help the LoRA learn the deformation details more effectively.
As of v0.0.1, this LoRA still needs to be used alongside LoRAs specialized in localized deformation, such as an ass ripple LoRA.
Description
Changes from v0.0.1
1. Training Dataset
Increased the total number of training images from 60 to 80, including an increase in cropped close-up images from 10 to 21.
Images containing two or more panels were cropped to retain only the relevant panel.
Replaced several training images that showed insufficient deformation of flesh in close contact with new images.
2. Captioning Adjustments
Standardized captioning by assigning
pussy pressto images showing deformation of the female genital area, andass pressto images showing deformation of the buttocks or thighs.Reserved
ass ripplefor images with more pronounced deformation. Images with excessively exaggeratedass ripplewere not included in the training dataset.
3. Training Configuration
v0.0.1: LoRA, dim16/alpha16, AdamW8bit, UNet LR 1e-04, constant scheduler, 1,100 steps, 60 training images, 1024² resolution. 66MB.
v0.1.0: LoKr, dim32/alpha32, CAME, UNet LR 3e-05, constant scheduler, 1,000 steps, 80 training images, 512² resolution. 3MB.
Results and Recommendations
These changes make flesh deformation easier to reproduce. However, due to hardware limitations, the training resolution was reduced from 1024² to 512², which may result in a loss of detail in generated images.
For this reason, a LoRA weight of 0.8 or lower is recommended to help preserve image quality.
LoKr, dim32/alpha32, CAME, UNet LR 3e-05 constant 1,000steps, 80 images.