About CMD dataset

[liziming5353]

1. The original video is mkv format. But your code use the format of image-frames. So do we need to preprocess the video first?
2. Is the caption.json the subtitle of CMD dataset? I find that the content in caption.json does not match the videos.

[KerolosAtef]

Hello @liziming5353 sorry for late replay.
I have upgraded the CMD's data loader to support .mp4 videos (you can edit it to .mkv in your case in get_item function). Additionally, I have removed the caption.json file and adapted the code to handle subtitles.vtt files similar to the webvid and videochatgpt data loaders.
check these updates and let me know if you have any other questions