This model is trained to create better anthro art. Because, this was trained on 993 images using the dev training model. Using sd 1.5-2.0 checkpoint.diff models may differ when using my lora model with those model types. Feel free to sell any images you create with this model.
Description
FAQ
Comments (2)
really nice one, it will be one of my favourite flux source :D
Glade you like it. :) The model might not be as flexible as other character lora models. Since, I can't train models on my intel optiplex 3010. But, if youd ever need help tagging or prompting images for characters. I'd be happy to help out. I find manually adding 30 or tags per images. Does help train a better model with what's in the image. Since, ai tagging misses so~ much image detailed. Like, hair types, hair color, how layered fur might look. Making a distinction between body/ chest hair/ fur. Since, I don't think civit ai auto tagger can tell the difference between things. To make detailed training data. Even, the captions can be iffy at times XD
I could talk about, tagging stuff for a good while lol