V1 is trained on 60 images and captioned with Civit Ai's auto caption feature. Settings left at default.V2 is trained on more images but without captions. May be better or worse.V3 turned out really good so far. Would love to see what you make with it.Prompting: "Highly detailed traditional artwork by Ayami Kojima of..." Feedback is welcome.

Files with SHA256: 9b4244968dae52fe8772956305b5af0c63d22b41fe5a06cf49cc3f5a23f957a3