Multitaskvideoreasoning

MultiTaskVideoReasoning is the official training dataset for the research project 'Thinking With Videos: Multimodal Tool-Augmented Reinforcement Learning for Long Video Reasoning'. The dataset was created by author 'zhang9302002' and was last updated on Hugging Face on 2025-08-25. Its structure includes multiple JSON files such as 'actnet.json', 'charades.json', and 'longvideo-reason.json', suggesting it aggregates data from several established video reasoning benchmarks.

Use Cases

• Training multimodal agents for long-horizon video reasoning tasks based on the dataset's stated purpose.
• Benchmarking video question-answering models using aggregated data from sources like 'nextgqa.json' and 'Video-R1-data-image.json'.
• Developing tool-augmented reinforcement learning algorithms for video analysis as described in the associated research project.

Strengths

• Serves as the official training dataset for a specific, cited research project ('Thinking With Videos').
• Aggregates data from multiple established video reasoning sources, as suggested by file names like 'actnet.json' and 'charades.json'.
• Last updated on 2025-08-25, indicating recent maintenance.

Limitations

• Row count, file size, and specific column definitions are unknown, which limits suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• License information is unknown, which may restrict usage.

Provenance

Source

huggingface

Collection Method

Aggregated from multiple video reasoning benchmarks, as suggested by the file structure.

Freshness

Last updated 2025-08-25 03:54:29.