Datasets:

---

sunwhw

/

InstructionBench

like 0

This pull request significantly enhances the InstructionBench dataset card by:

• Adding the video-text-to-text task category to the metadata, which greatly improves the dataset's discoverability on the Hugging Face Hub.
• Including relevant tags such as video-qa, instructional-videos, and benchmark for better categorization.
• Linking the dataset to its associated paper: InstructionBench: An Instructional Video Understanding Benchmark.
• Expanding the description with key information from the paper's abstract, detailing the benchmark's purpose, methodology, and scale, providing a richer context for users.
• Refactoring the content for better readability, including a clear title and a dedicated "Dataset Structure" section.