v1.0 : LoRa trained on 100 images of the upper half of actress Elizabeth Olsen. Best results with weight between 0.7 and 1.1 for models like Dreamshaper and other models with semi-realism.
v1.1 : Retrained on the 100 images but with a resolution of 1024x1024. Works best for generations with 896 or less pixels.