Image augmentation for training

[dongyun-kim-arch]

I would like to know how to augment my training data in data-loader.
I am trying to img2img translation but due to the limitation of dataset, augmentations such as random crop, zoom in and out, etc. are necessary. I am familiar with augmentation using Compose method, but I couldn’t figure out where I should write code. It would be great if someone could show simple augmentation example as well. :)

[danieltanhx]

try to modify the Dataset class just like in the link->https://github.com/Hramchenko/simplified_pix2pixHD/blob/main/simplified_pix2pixHD.ipynb

[zhangxiaojuan66]

你好，邮件已收到，祝你万事如意，生活愉快！