This is another test version, so please keep in mind that I’m currently experimenting with the model. There’s no need to focus too heavily on its flaws yet, they’ll be addressed as I continue refining it. For now, I mainly want to see how the model reacts across different situations.

This is designed as a general-purpose LoRA that can work with a wide variety of subjects and styles. Most of the classic settings in NSFW should work well.

For the LoRA strength, I recommend staying between 0.50 and 0.80. Within that range, the model should retain its intended behavior and overall quality. Going above 0.80 may begin to overpower the base model and reduce image quality.

Most standard models should work with it. You may simply need to experiment with the number of steps and slightly adjust the settings depending on the model and workflow you’re using.

This is still an early test version. It was trained on a broad and varied image dataset because I wanted to evaluate how it responds to different prompts, styles, characters, and situations. I’ll continue refining the dataset and improving the LoRA based on the results.

Feel free to experiment with it and ask questions. Feedback is welcome, but please remember that this version is primarily intended for testing and will be improved soon.

https://t.me/ccmb1ne
https://t.me/+BrNISE72oLYwNDMx