Nadia is yet another of my entirely fictional people, trained on prompted images with no lora or embeddings involved.
She's raised well on 25 vectors involving 53 prompted images in 2500 steps with a learning rate of 0.002