Purana LoRa

V3.0

1. We have added images to the dataset, which allows for diverse character expressions.

2. We have completed more training and can now depict character features such as clothing and accessories in detail.

3. We have improved the tags to avoid duplication. The tags that require manual input will be listed below.

This is not a trigger word.

>> 1girl, highres, masterpiece, halo, white hair, hairband, grey eyes, <<