Datasets:

---

sakshamsingh1

/

av_robust_data

like 2

This PR adds a comprehensive dataset card for AVROBUSTBENCH.

It includes:

• The paper title and a link to the Hugging Face paper page.
• Links to the GitHub repository and the YouTube demo.
• The abstract of the paper.
• Detailed information about the four constituent datasets: AUDIOSET-2C, VGGSOUND-2C, KINETICS-2C, and EPICKITCHENS-2C.
• A "Sample Usage" section with code snippets directly from the GitHub README, demonstrating how to set up the environment, prepare data, load the dataset, and create corrupted videos.
• Appropriate task_categories (audio-classification, video-classification), language (en), and descriptive tags (audio-visual, robustness, benchmark, corruption) in the metadata for better discoverability.