21 upvotes
New foundation model spots deepfakes and physical spoofing alike
What is Venus-DefakerOne2: A Vision-Forensics Foundation Model with Broader Coverage?
Venus-DefakerOne2 is a vision-forensics model that detects digital manipulations and real-world spoofing in images and videos. It provides a single tool for a threat spectrum that usually needs multiple detectors.
Who it is for
Security teams and platforms battling misinformation can use it to flag forged media across formats.
How to try it
Open the model page on Hugging Face