A Flux-dev version of the lora. It uses the same training images as the SDXL version.
Trained on 19 images, mostly artwork by Lorraine.
