I finally wanted the Lavadress in an XL model. Pictures were learned with 832x1216 pixels
LavadressXL.safetensors
474956_training_data.zip