This is a highly experimental style LoRA made with one image.
This model does not produce stable results, so download and use it with that knowledge.
Fingers and general ??????? is quite broken and the model is generally overfit.
Training Diary Article (lots of images)
Usage
Use the trigger word: "KawaiisStyle" along with the LoRA to activate it, but you also need to describe the character.
Example prompt: kawaii style 1boy, solo, beard, a pirate standing on a mighty ship at sea. A parrot on his shoulder<lora:KawaiisStyle:1>
Or if you are testing the non-mask version:
Example prompt: kawaiis 1boy, solo, beard, a pirate standing on a mighty ship at sea. A parrot on his shoulder<lora:Kawaiis:1>
You will want to add an environmental description, as the models are trained without backgrounds.
Note:
I may be using the following extensions in the reference images: FreeU or ADetailer / Face Editor. I may also be using a specific Hires Checkpoint (swaps to another model when doing the hires steps). If your results don't match mine exactly, this may explain why.
Tips and tricks
Below are some tips and tricks for how to get even more out of the model.
You can always add some extra keywords to enhance the style. Like:
cute
chibi
2DLike it?
Description
This model is trained on a single image dataset, using network rank 16 and 2000 steps of training. The published version used step 2000. It's definitely overfit!
Please read more about the training diary over here: Training Diary Article (lots of images)
⚡⚡⚡ Please support the training of this model by unlocking it with some buzz. Any buzz donated will go towards future experimentation and knowledge sharing ❤️ ⚡⚡⚡