Stevie Wynne Levine v1.0

This LORA was trained on publicly available photographs of Stevie Wynne Levine. Good high quality training data was very hard to find and there weren't many pictures available, but i believe it turned out quite well despite this. The images were cropped to 1024x1024, the backgrounds were masked out, and the training was done with no captions on the GODDESS of Realism 1.0 checkpoint.

Keywords: None needed
Suggested Weight: 1.0
Suggested Checkpoint: GODDESS of Realism 1.0, Pony Realism 2.1 also works.
Other tips: Use the High Quality positive and negastive embeddings from this page Pony PDXL Negative Embeddings.I use tile upscaling to refine the face and eyes.

描述:

训练词语: