Stream Data: A Consolidated Collection for Streaming Video Understanding Research

A consolidated collection of video datasets for streaming video understanding research, including temporal grounding, video captioning, and video question answering tasks. The dataset is hosted by the author 'interlive' on Hugging Face and was last updated on 2026-02-22 06:23:27.

Use Cases

• Temporal grounding research based on the consolidated video datasets mentioned in the description
• Video captioning model training based on the ActivityNet-Captions and Charades datasets referenced in the directory structure
• Video question answering system development based on the dataset's stated purpose for video QA tasks

Strengths

• Consolidates multiple established video datasets (ActivityNet-Captions, Charades, CharadesEgo, DiDeMo) into a single collection
• Explicitly targets three core video understanding research tasks: temporal grounding, video captioning, and video question answering
• Last updated on 2026-02-22 06:23:27, indicating recent maintenance

Limitations

• Row count, file formats, and total size are unknown, which may limit suitability assessment
• Column-level documentation is absent; field semantics must be inferred after download
• License information is unknown, which may restrict usage

Provenance

Source

Hugging Face (author: interlive)

Collection Method

Consolidation of existing video datasets.

Time Range

null

Freshness

Last updated 2026-02-22 06:23:27

Geography

null