While I'm between working on some other models, this one is for fun and for a learning experience. What happens when training a FLUX model on so few images, so little variation, and there's a non-human feature on the character?The images of this model are from screenshots of the music video for the song Anoana by Heilung. The <a rel="ugc" href="https://www.youtube.com/watch?v=SVbc_Fwbt50">video </a>is on YouTube and a description about the lyrics on <a rel="ugc" href="https://genius.com/Heilung-anoana-lyrics">Genius</a>.Overall, I think it turned out well. Usually, it will generate the character without the antlers unless they are prompted.

Files with SHA256: 453bfc18d2ddcf245abd47879a7cb90a08e90ebe9546c29f5eb787b6398554b0