Trained on SD 1.5, recommend weights 0.7-1.2, checked on deliberate2 and anything4.5, trigger word is (gaben:1.2). Not so good for anime models.

Most of the dataset are screenshots of the three interviews available on YouTube, but there are also some old photos without the beard (add beard to prompt or negatives) and a little bit of drawn memes (sketch, art, parody). He's wearing glasses in all the photos, so sorry about that.

I'm not 100% happy with this LoRA because sometimes the beard looks made up of pieces of hair or random objects appear on the forehead, but that doesn't happen often.