search

jfzhang95 / pytorch-video-recognition

PyTorch implemented C3D, R3D, R2Plus1D models for video activity recognition.
MIT License
1.18k stars 250 forks source link

Why data enhancement on test datasets,not on train datasets? #39

Open momohuangsha opened 4 years ago

momohuangsha commented 4 years ago

https://github.com/jfzhang95/pytorch-video-recognition/blob/ca37de9f69a961f22a821c157e9ccf47a601904d/dataloaders/dataset.py#L81

DongdongMeng commented 4 years ago

Yes, i have the same question.