This Lora is trained with highres images of angel and ancient greek fashion:

• V1: without any caption. It is a test to see how it can work.

• V2: captionned with tags

• V3: (stay tuned for this one) will be captionned with natural language