Designed to introduce more Asian facial features. Recommended for use with other realistic-style LoRAs, as it was trained on lower-resolution images.

Trigger: raw photo
or Trigger: retouched photo

Trigger words are not strictly separated due to overlapping visual features in the dataset.