I’m tired of perfect faces with flux! I tought about doing this concept because imperfections are interesting and can be sexy or beautiful. It is trained on various expressions and I was careful to limit any unwanted bias or likeness capture. It is not trained on specific trigger words, but most dataset captions had something like;<ul><li>Gap between front teeth</li><li>Slightly open mouth</li><li>Smiling</li></ul>

Files with SHA256: a90d3031258fde3bf7798f660395fbf3a34ce75814e05c3d4614e67c507c8022