To which resolution do you usually crop your images when training LoRA models?
Poll options
What's this?