Why not use torchvision.transforms

sanghyun-son / EDSR-PyTorch

PyTorch version of the paper 'Enhanced Deep Residual Networks for Single Image Super-Resolution' (CVPRW 2017)

MIT License

2.41k stars 666 forks source link

Why not use torchvision.transforms #345

Closed xlxingRun closed 1 year ago

xlxingRun commented 1 year ago

Author use a lot of custom processing when handling data, so why not just call library functions?

xlxingRun commented 1 year ago

Perhaps custom data augment is more flexible, but I don't think that ShiftMean layer is set in model.py, which is related to the dataset, not the network model.