Data augmentation for limited data

Hi all,
I want to train a large CNN with limited amount of data.
So I need some random cropping and horizontal flipping to train this net.
How can I get the data augmentation by the torchvision.transforms ?
Could you please show me some examples?


you can look at the imagenet example to get yourself started. Here are the relevant lines:


Will the composed transformer just give one output or every transform give a output image?Just the last one can be the data augmentation method.

1 Like