Created a LoRA from heterochromia images from gelbooru.
Most likely will release a better version in the future since I learned how to do more than one trigger word after training the model.
Support me: https://ko-fi.com/elevatedone
Recommended weights: 0.6
Use "<color> and <color> eyes" for different eye colors (replace <color> with the color you want; works 75% of the time)