Hi EgoExo4D team,
Since downloading all takes all at once consumes a large amount of storage, I am thinking about downloading a subset of takes according to take scenarios.
I realized that there is a trick for Ego4D dataset which we can filter using the visualization tool and then use the .txt file to download filtered videos. However, this trick doesn’t help in the new EgoExo4D case. The generated .txt file no longer contains downloadable take_uids but take_name instead. Although I guess we can still map the take_name to the take_uids through annotations, it would be better if we can generate .txt with take_uid directly.
Thank you for your help and hope to hear from you soon!