Stylization by Anton Semenov.

117 images were used to train the model.

the training of the LoRA was on the CyberRealistic 3.1 model.

Training with a resolution of 768x768 images.