LoRA trained for high likeness and versatility. It was trained on a curated dataset of 60 images. Same dataset as V1, duplicate training completed by mistake, not sure which/if one is better.