ByteDance's New AI Tool Lets Anyone Shine in Videos
ByteDance's MAGREF AI generates multi-subject videos from a single image—no face-swapping, full control over characters & scenes.
"AI Disruption" Publication 6800 Subscriptions 20% Discount Offer Link.
Generate a video with highly consistent subjects based solely on a single reference image, ensuring no face-swapping even with multiple subjects on stage!
ByteDance releases a multi-subject video generation powerhouse — MAGREF (Masked Guidance for Any-Reference Video Generation).
For example, an Einstein-inspired high-speed motorcycle video, with results like this:
Newton indulged in wielding his pen to create a masterpiece:
Provide a reference image of Turing, and it can generate a full-body animated video, maintaining high consistency in facial features and clothing across every frame without gradual deviation:
Based on diverse text prompts, MAGREF achieves “one image, a thousand faces”:
