NB: This is mainly just a test LoRA to see what furry baras in Anima might look like by training in some basic concepts that aren't in the current Anima model. It's obviously not a replacement to illustrious nor is it meant to be taken seriously. Just a fun little toy for those interested.
Trained on a 1.5k image dataset mostly focused on furry bara artwork, but was surprised to see the LoRA is still capable of furry concepts beyond the dataset, plus its nsfw capable too which is neat! Super happy with how this came out. It might not be illustrious levels good but I'm shocked at how consistent it is.
I recommend using at 1 strength and at 1024px using a mix of captioning and natural text prompting. Check the sample pics for a quick workflow I put together. Some sample pics I tried to push the res past 1024 but I wouldn't recommend it unless you're happy with losing a lot of consistency. You'll also see I had a little too much fun with text generation, another feature of Anima, although it doesn't get it right every time.
Lastly a huge thanks to Vennin for their help. I highly recommend checking their Anima LoRA also. It has a much larger and wider dataset than mine, plus he's actively working on it. Will link it below.