UniVBench: A Benchmark Dataset for Video Editing Tasks

UniVBench is a benchmark dataset for video editing tasks, organized into numbered folders representing specific test cases. The dataset was created by JianhuiWei and was last updated on March 26, -2026. The full structure and description are available on its Hugging Face dataset page.

Use Cases

• Benchmarking video generation models based on the structured test cases
• Evaluating video editing algorithm performance based on the task-specific folders
• Training multimodal AI systems for video manipulation based on the benchmark data

Strengths

• Dataset is structured into specific, numbered test cases for video editing
• Last update timestamp is precisely recorded as 2026-03-26 09:06:14

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download
• Column-level documentation is absent; field semantics must be inferred after download
• Row count, file formats, and license are unknown, which may limit suitability assessment

Provenance

Source

Hugging Face dataset created by JianhuiWei

Collection Method

Likely compiled for benchmarking video editing tasks; specific method is not described.

Time Range

null

Freshness

Last updated 2026-03-26 09:06:14; freshness should be verified

Geography

null