Although the preview image may not show it, this LoRA is based on Katsuya Terada's LttP artwork. I always like to try to create preview pictures with empty prompts to see what the AI was taught, but of course there is a bias towards female characters. Feel free to experiment and create dark fantasy images. I would love to see your results.
Description
This is my first time trying to train a LoRA for Anima, and I do not have much experience in Anima image generation and prompting yet, so I still hope it will be of use for you guys. Have fun and please share your creations in the gallery as inspiration for others.
FAQ
Comments (2)
The Anima version of the LoRA works great, love the style. I wanted to train one myself but I can't find good enough images for the dataset. Guess I don't have to :D
Trained encoder seem to hurt prompt following a bit, and the LoRA weight seem to work best at 0.8 , might be slightly overtrained.
Thank you so much for your review. Unfortunately I am not really fit when it comes to the technical side of AI. I am good at curating and searching for art. If you have any tips I may try again to see if I get better results. I also have earlier epochs of this training run, so if there is interest I can upload it.