Multi-output using U-Net

[tamaraalshekhli]

Hi there, my input is one image and the out put is 2 masks one for road and the other one for centerline segmentation , for the (trainGenerator) I defined 3 generators, image_generator, mask_generator1 and mask_generator2, now I dont know how to train my Unet using these generators, is there anyone tried multi-labels? any help or advice will be appreciated