But I have a question about the dataset. The events loaded uses data augmentations such as horizon flip, and random crop to 128x128 when training. But the codes only contain flip, no crop. When training you use 128x128 according to the config files. It is a error If just read the events without crop.
Hi, thanks for the amazing work.
But I have a question about the dataset. The events loaded uses data augmentations such as horizon flip, and random crop to 128x128 when training. But the codes only contain flip, no crop. When training you use 128x128 according to the config files. It is a error If just read the events without crop.
https://github.com/tudelft/ssl_e2vid/blob/84f7c7e59875f134e97c14ec423f396725e04be7/dataloader/h5.py#L216