Aggron LoRA trained on NovelAI with Kohya training scripts using the DAdaptAdam optimizer at 512x512 resolution. Body detail is very inconsistent, expect lots of spikes in places they shouldn't be. May require adding (black skin) to your prompt to get the correct skin color in certain situations. Despite the issues, good results are possible. Should have wide compatibility with model mixes based on NovelAI, which is most of them. Trained on an all female dataset.
Description
Initial release
FAQ
Comments (3)
My second favorite pokemon, LET'S GO!!!!
now all i need is swampert and I'll be set.
Love your Pokemon Loras, you're actually doing some more interesting ones than the million gardevoir ones we have already 🥴
I know it's not very popular but could you please consider doing Tapu Fini in the future?
I took a look at the available dataset for Tapu Fini and I don't think there are enough images to make a good LoRA. I've added them to my 'B' list of Pokemon that I'm going to try after I've plucked all the low hanging fruit ❤️