As with all my loras, there's no custom tags aside from his name, prompting his in-game clothing through common sense should be enough. Due to extremely poor eye quality in the limited dataset, not specifying eye color might make eyes smear less - mostly untested but it seems to have worked for me in the past.Hard to tell which version's best but I opted to leave 1e4 as main because it seems to go better with heavy stylized models.I've uploaded the "new_1700" one because I had success using it in the past - theres a chance it's still the best one but I don't trust the old settings and I don't exactly feel like spending my entire day testing all of this further.As per usual, tell me if you need any other version uploaded or if you got some insight on which version works best.

Files with SHA256: 5c801a39496196af623396c18f0832cdb7bf9cd4a49d508f7bfb2637fc371500