The character must do somersaults (backflip). This is my first test-LORA on movement for WAN2.2 I2V (image to video), in most cases it works, but in 10% of images (depending on the scene and the environment of the character) it does not work correctly. The keyword "performs the fl1p" is, for example, "men performs the fl1p".
Description
The character must do somersaults (backflip). This is my first test-LORA on movement for WAN2.2 I2V (image to video), in most cases it works, but in 10% of images (depending on the scene and the environment of the character) it does not work correctly. The keyword "performs the fl1p" is, for example, "men performs the fl1p".