Video Question Answering Dataset for Multimodal Reasoning

Egotextvqa is a multimodal dataset for video question answering tasks. The dataset was created by ShengZhou97 and was last updated in April 2025. It contains video and text data, focusing on reasoning tasks that require understanding both visual and language information.

Use Cases

• Train models to answer questions about video content using associated text annotations.
• Benchmark multimodal reasoning systems on tasks requiring joint understanding of visual sequences and language.
• Develop video-language models that can process and reason over features extracted from video frames and textual queries.

Strengths

• Dataset is hosted on HuggingFace, ensuring accessibility and integration with popular ML tools.
• Focuses on the specific and challenging task of video question answering.

Limitations

• Unknown dataset size, row count, and video duration limits hinder assessment of scale.
• Potential geographic or content bias is unknown due to unspecified data collection sources.

Provenance

Source

HuggingFace, uploaded by author ShengZhou97.

Collection Method

null

Time Range

null

Freshness

Last updated in April 2025.

Geography

Region tag suggests 'us', but specific coverage is unknown.