Description: A fan-made LoRA capturing the style of Manaka Laala from her 3D Live performances.
What's New:
Updated with the Krea2 version!
Recommended Prompt/Tags: manaka laala, hairbow, ahoge
P.S. You might wonder why I'm not making a version for Anima, since it's a great model. First, Anima can already handle this character pretty well even without a LoRA. Second, since it's an NSFW-easy model, I personally feel it doesn't really fit her pure/innocent vibe, so I don't plan to support it. Thanks for understanding!
Description
My first Flux version Lora.
FAQ
Comments (7)
Hey, this is amazing, I am also trying to train a LORA for dragon ball characters but am not able to get good results, may you share your training methods and training data, that would be a huge help
Thank you for your comment. Due to various reasons, I can't share the training data, but generally speaking, it's a pair of character images with the background removed and caption files. I used the AI Toolkit by Ostris for training. Most of the training settings are the same as in its example files, and it does a very good job.
I look forward to seeing your great work.
@i_need_maidrobo I can understand but if possible may I ask what type of language you use to caption? an example you can provide?
@krigeta For example,
"laala, closed eyes, gloves, white background, open mouth, bow, dress, very long hair, twintails, choker, simple background,”
More sophisticated methods need to be explored further.
@i_need_maidrobo, do you use the booru tagging captioning?
@krigeta Yes, I don't think danbooru tagging is the best but at least it helps.
kashikoma
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.