Makes images in a monochrome/greyscale style with hatching textures.
How to use
While the dataset was very small this style LoRA came out rather versatile so you should be able to prompt most things. Approx half of the images had Chainsaw Man content so anything from that should work particularly well. For more tips see the sample images and uploaded zip of tags.
Recommended strength is anything from 0.5 to 1.0 depending on what you prompt. The non-human stuff requiring higher values. For samples I recommend DIMM or DPM++ 2M Karras since those best recreate the hand drawn style.
MISC
Trained on NAI at 512px with a dataset of 25 images.
All images are raw txt2img upscaled without any inpainting or helper embeddings.
The training was done on a small subset of the artists images, as such the output of this LoRA is not an accurate representation neither in quality or style of the artist Epeulu. Go support them if you want the good stuff.