Skip to content

Dataset · ControlNet

Large-scale English audiovisual deepfake dataset

ControlNet/AV-Deepfake1M

Millions of English videos for training and evaluating deepfake detection systems.

What's inside

It contains English video files created with LLM assistance for deepfake research. The data volume ranges from 1 to 10 million records. The dataset is designed for video classification tasks, as well as detecting and localizing fake audiovisual content.

What you can do with it

  • Training classifiers for recognizing video and audio deepfakes
  • Testing algorithms for localizing fake segments in clips
  • Comparing modern synthetic content detection methods
Tasks
Video classification
Languages
English
Size
1M–10M rows
Downloads this month
4.2K
Likes
29
Access
on request on Hugging Face

How to load it

from datasets import load_dataset

ds = load_dataset("ControlNet/AV-Deepfake1M")