A work in progress model trained on (currently) 40 images of abstract eyes (gonna add more images and probably try to make custom tags for better eyes). Works best with adetailer or inpainting, putting it in the txt2image prompt will force close-ups and probably bad anatomy. prominent tags in the dataset ' gradient eyes, multicolored eyes)weights 0.8 -1.0May upload a Lora and Glora version.

Files with SHA256: 49891afff40ed3425ec1927e7ffbe6a17fb5dcc48f1214ab7edfda6b1c7d0d95