Try size: 768x 1024 or 1024 x 1024

Weights: 0.7~1.0

Clip skip: 2

Training material from Youtube screenshot and AI-generated images

Pictures from ZImageTurbo and DreamshaperXL_v21TurboDPMSDE

Note: The background is designed to recreate the details and atmosphere of the video as closely as possible, but the girl's face is not; you'll see the difference when you compare them. (ILXL and ZImageDeTurbo)

Note: It is not recommended to use the SDXL version, as it is an early and immature version, and it's training on Pixai. (SDXL)

Reference Video: