This model is trained on 39 art pieces made by: KuroganeNiello It has a monochrome 1bit pixel art style, I didn't take too much effort into training this LoRA and it has issues generating backgrounds, objects &amp; animals. but if there is a big demand for this LoRA I might try and improve this model with more images and more epochs in the future. so far the biggest issue is that it generates uneven pixels, I'm sure I could fix this with enough training however.best settings for this LoRA is:<ul><li>Size: 756x756 to 1024x1024 or higher (pixels become uneven on lower resolutions)</li><li>Steps: 35-70 (pixels become uneven on lower step counts, if your pixels end up uneven, increase step count.)</li><li>CFG scale: 7-14</li><li>LoRA scale: 0.6-0.75 (the lower you will set the scale, the more it will look like traditional art and the more colors will start to appear, but the higher the scale is it will appear much more pixel-y. on scale 0.9 and above it will appear "too" pixel-y)</li><li>Sampling method: DPM++ 2M SDE karras</li><li>Recommended prompt: "kuro_gane, pixel_art, monochrome" (Add these to the start every generation, this isn't "required" but I found it gets the best results)</li></ul>if you are the artist who's style is depicted in this work, please DM me with proper evidence that you are in fact the artist and I will promptly remove this model from CivitAI. I respect artists who do not wish for their work to be used in AI generative learning.Trained using hollowstrawberries's <a target="_blank" rel="ugc" href="https://colab.research.google.com/github/hollowstrawberry/kohya-colab/blob/main/Lora_Trainer.ipynb">LoRA trainer</a>,<a target="_blank" rel="ugc" href="https://colab.research.google.com/github/hollowstrawberry/kohya-colab/blob/main/Lora_Trainer.ipynb"> </a>It runs on google collab!

Files with SHA256: eaa90e849f0f408aa41a3930c4d85eba7407e85a3ffe8fafa547baae66c23b85