Open X-Embodiment
1M+ real-robot trajectories from 20+ institutions.
Streaming Video Dataset Description A consolidated collection of video datasets for streaming video understanding research, including temporal…
Streaming Video Dataset Description A consolidated collection of video datasets for streaming video understanding research, including temporal grounding, video captioning, and video question answering tasks. Directory Structure streaming-cvpr/ ├── ActivityNet-Captions/ │ ├── activity net.v1-2.min.json │ ├── activity net.v1-3.min.json │ ├── missing files/ │ ├── v1-2/ │ └── v1-3/ ├── Charades/ │ └── videos/ ├── CharadesEgo/ │ └── videos/… See the full description on the dataset page:
Source: Hugging Face Hub (interlive/stream-data). Metadata imported from the dataset’s Hub tags.