Beta101: Reuse the refined training data from Pony
============================================
Warning! Beta102 and Beta89 v13 was created as part of the research on AdamW 32 bits, and not intended for regular usage.
Please add the following arguments in Automatic1111 WebUI webui-user.bat in order to enable 32 bits.
Feel free to create the picture if interested, however, the file is huge.
============================================
Thanks for the interest in this LoRA creation. This project has been conducted since March 2023. The objective of this project is to create an ideal training data set for LoRA.
This LoRA has been created and tested scientifically from March 2023. The face of the researcher was taken as reference in order to create the training set. 30 photos were taken purposely for AI training only. These 30 photos contain different angles of face and full body to fulfill the AI training requirement.
LoRA has been tested internally since March 2023, and the version is updated based on the outputs and feedback. Thanks to the members from the following groups for generating the pictures and discussing the result, so that this LoRA could be improved greatly before released:
AI乐园: https://t.me/+1XT8FjX5CUlhNDNl
AI绘图魔法群: https://t.me/+jPp0_tAk-UY5NDhh
AI练丹交流
奇思妙想会
AI绘画
lora交流
AIGenerateMagic
兔狲的SDXL百老汇
After reaching this stage, there is no further element to be added. Therefore this LoRA is released in public so that more justification from the output, and more feedback perhaps.
Description
For best result on face, try LoRA weight: 0.5 ~ 0.7.
The nipple piercing and tattoo are used to justify the training steps, if the LoRA weight is more than 0.8, it would display them in detail, however, the face might show over trained.
This LoRA produces nude photo naturally, you have to assign clothing manually and reduce the Weight.