makes people (and some objects) look like in early 2000s music videos.
Also good on lower values to give characters in a creation a more african-american or latin look since the LoRA was almost exclusively trained on such
Description
more intense training on same data