Skip to content

Dataset · unibuc-cs

MAVOS-DD multilingual deepfake benchmark

unibuc-cs/MAVOS-DD

A benchmark for audio-video deepfake detection across multiple languages.

What's inside

Features the open MAVOS-DD benchmark for multilingual audio and video deepfake detection. It contains video data in Arabic, Romanian, English, German, Hindi, and Spanish. The dataset size ranges from 10K to 100K samples and is distributed under the CC BY-NC-SA 4.0 license.

What you can do with it

  • Developing and testing media falsification (deepfake) detection systems
  • Evaluating multimodal model robustness against deepfake attacks
Tasks
Video classification
Languages
Arabic, Romanian, English, German, Hindi, Spanish, Russian
Size
10K–100K rows
Downloads this month
9.9K
Likes
11
Access
on request on Hugging Face

How to load it

from datasets import load_dataset

ds = load_dataset("unibuc-cs/MAVOS-DD")