Scalsol / mega.pytorch

Memory Enhanced Global-Local Aggregation for Video Object Detection, CVPR2020
Other
565 stars 115 forks source link

Duplicated file names in DET_train_30classes.txt #23

Closed duanzhiihao closed 4 years ago

duanzhiihao commented 4 years ago

It seems that there are duplicated images in datasets/ILSVRC2015/ImageSets/DET_train_30classes.txt, which can be confirmed by the following script.

lines = open('DET_train_30classes.txt', 'r').read().strip().split('\n')
print(len(lines))
# 53639
print(len(set(lines)))
# 53237

Is there a reason that you use duplicated training images?

Scalsol commented 4 years ago

This file is directly copied from the FGFA repo and I did not check whether duplicated training images exist. This would not influence the final performance in my opinion.

duanzhiihao commented 4 years ago

Thanks, I agree.