ROMA Proactive: Multimodal Streaming Video Interaction Data

A subset of the dataset introduced in the paper 'ROMA: Real-time Omni-Multimodal Assistant with Interactive Streaming Understanding'. This dataset is designed to train multimodal models for streaming video understanding, focusing on proactive interaction tasks. It was authored by EurekaTian and last updated on the Hugging Face platform in January 2026.

Use Cases

• Training models for real-time video understanding based on the described streaming focus.
• Developing proactive interaction capabilities for AI assistants based on the dataset's stated purpose.
• Benchmarking multimodal models on tasks involving continuous video and audio streams.

Strengths

• Dataset is explicitly designed for a specific research area: training models for streaming video understanding.
• Associated with a named research paper ('ROMA: Real-time Omni-Multimodal Assistant with Interactive Streaming Understanding'), providing academic context.

Limitations

• Description metadata is limited; actual data quality, structure, and content require manual inspection after download.
• Column-level documentation, file formats, and dataset size are unknown, complicating suitability assessment.

Provenance

Source

Hugging Face dataset authored by EurekaTian.

Collection Method

Likely collected or generated as part of the research for the associated 'ROMA' paper.

Freshness

Last updated 2026-01-19 02:42:22; freshness should be verified.