cvdfoundation / kinetics-dataset

716 stars 92 forks source link

When decoding some of the kinetics400 videos, "moov atom not found" raised #3

Closed irvingzhang0512 closed 2 years ago

irvingzhang0512 commented 3 years ago

Hi, Thanks for your dataset! It's very helpful!

I have successfully download kinetics400 dataset and annotations and have two questions.

  1. Some of the videos in kinetics400 train set are corrupted. Try to decode video with cv2.VideoCapture("/path/to/video"), "moov atom not found" raised. The corrupted video file list is here test.txt

  2. There are 240k videos in kinetics400 train set. However, it seems that not all of them(about 238k-239k) are in kinetics400 train annotations

kinetics-cvdf commented 3 years ago

Hi Irving, thanks for your message. We will try to replace these corrupted videos in the training set. Regarding your second point, some of the videos in the training set were not publicly available anymore from youtube at download time, so they are not included.

irvingzhang0512 commented 3 years ago

Regarding your second point, some of the videos in the training set were not publicly available anymore from youtube at download time, so they are not included.

Sorry for my poor english.

~There are 246k samples in training annotation file, 240k video files in tar. However, There are 1-2k video files are not included in training annotation. I may double check this and generate a filelist~

It seems to be a bug of my own codes, all video files are included in annotation file

kinetics-cvdf commented 2 years ago

Added a tar file containing around 1300+ replacement videos (around just 80 missing now, which may be hard to get, so closing this).