search

xinntao / EDVR

Winning Solution in NTIRE19 Challenges on Video Restoration and Enhancement (CVPR19 Workshops) - Video Restoration with Enhanced Deformable Convolutional Networks. EDVR has been merged into BasicSR and this repo is a mirror of BasicSR.
https://github.com/xinntao/BasicSR
1.48k stars 320 forks source link

There maybe a bug for REDS_dataset.py to remove testing clips #125

Closed DengpanFu closed 4 years ago

DengpanFu commented 4 years ago

In data preparation, original REDS validation set's clips were renamed starting from 240, but in codes/data/REDS_dataset.py lin58 self.paths_GT = [ v for v in self.paths_GT if v.split('_')[0] not in ['000', '011', '015', '020'] ] ['000', '011', '015', '020'] were filtered out for testing, but i think the real ones to filter out should be ['240', '251', '255', '260']. Am I understanding it wrong?@xinntao

ckkelvinchan commented 4 years ago

We take out the four clips in the training set (['000', '011', '015', '020']) as the test set. All videos in the validation set are used in training.