search

kuangliu / pytorch-cifar

95.47% on CIFAR10 with PyTorch
MIT License
5.94k stars 2.14k forks source link

Question regarding train transforms used #130

Open ssgosh opened 3 years ago

ssgosh commented 3 years ago

Hello,

I am wondering why only random crop and horizontal flip were used for data augmentation, and not something like RandomAffine as well? Does using RandomAffine not have any additional gains?

https://github.com/kuangliu/pytorch-cifar/blob/49b7aa97b0c12fe0d4054e670403a16b6b834ddd/main.py#L32