1 dataset found

Groups: Video Action Localization Organizations: No Organization Formats: JSON

Filter Results

  • CrossTask

    The CrossTask dataset contains 2,750 instructional videos, annotated for 133 keystep labels spanning 18 tasks.