I really liked the characters of Betelgeuse and Minami River in deep-sea movies, so I trained this Lora with about a hundred or so 4K movie screenshots.
This model performs well on close-range faces, and the trigger keyword is shenxiu nanhe.
In addition, the weight of this model is optimally set to 0.7-1.
I hope you like this model, if you have any questions, please feedback me in the comment area.
Also, due to uncertainties, I do not recommend using this model for commercial operations, so please consider how you should use this model as appropriate.