Just sent the final LoRA for my Christmas Advent Calendar (the one for Christmas day) to the trainer as I start writing this. This one is actually Christmas themed, and I actually held off on making it till I was sure we weren't getting another LoRA contest. At Epoch 7 it already looks to be coming along nicely despite the very unclean data.
Taking the chance to ask a question to everyone: My 200th "distinct" LoRA will relatively soon in 2026 since I'm just under 190. I also planned to try some ZIT LoRAs in 2026, both conversions and at least one thing I don't think Illustrious can handle. This brings up the question: What actually counts as a unique LoRA? So far I haven't really had to decide since my only dips into a natural language was one joke (hasty captions to be able to publish an Ultra Lord LoRA seven weeks in a row, which no longer works because the SD3.5 one got killed by retroactive licensing shenanigans and now there's a hole in the publication) and one test. I never counted having made a LoRA for both Pony and Illustrious (or 1.5 the three times I did that) because it's just taking the exact same training data and throwing it into the trainer (there's likely some edge cases where the tag dumps the two are based on differ, but I don't think I ever altered a retrain specifically for that), but tagging is a major part of training, and that would be completely redone when writing natural language captions. Might as well ask the audience what they think.
Pic not really related. It's another Christmas Sweater for the daily challenge, but there's not really anything Christmas about it. Another random image from my history because meat sweater is X rated for some reason.