It would be great help to post your own creations or to like the pictures, it helps to generate lora faster. Tipping buzz would be appreciated but not required.
This lora aims to imitate the style of Kakeru.
Weight: 1.0
Trigger words: none
Useful prompts: Monster girl, mermaid, lamia, harpy, winged arms, ogre, tail, goat feet, bird feet
Comic, Grayscale, Monochrome- also produce nice results
This was my first time trying to make a style lora. While I feel the end result is not too bad, it can certainly be improved on. Better prompting and cleaning up the training data should be able to produce better results. Trained in nearly 700 images caused the prompting to be basic and the lora to be under trained. This is definitely one I want to redo and take my time prompting and cleaning up the training images while I save up the buzz to properly run it again.
Description
FAQ
Comments (5)
Thanks for making this style lora, I love this artist so much! Is it possible to make a illustrious or NoobAI version?
I have been working on that for like 8 months. Currently have a dataset of around 3k images. I honestly work on it a little then put it off. Between cleaning the images and tagging them I make progress here and there. The only problem is that max training Images is 999, so Im thinking of curating a general lora then 1 lora for each respective series.
@sector171 To put in context, this was the second lora I ever made, and the actual reason I learned how to train.
You should read this. https://civitai.com/articles/22109/the-25-days-of-christmas
Who knows? if i'm motivated good things could happen on the 25.
@PinGOrio glad to hear this lora is still in progress, thanks for your effort.
@sector171 Merry Christmas!