HeyGem.ai's Open-Source Digital Human Model Garners 1,000 Stars in Just 3 Days
2025-03-11 / Read about 0 minute
Author:小编   

HeyGem.ai, a digital human model developed by Silicon-based Intelligence, has swiftly amassed over 1,000 Stars on GitHub within just three days of its open-source release. This innovative model is capable of cloning an image and voice within 30 seconds by simply inputting "a 1-second video or 1 photo," and generating a 4K ultra-high-definition video in just 60 seconds. Even in intricate scenarios, it maintains a perfect 100% lip-sync accuracy and outputs a diverse range of facial expressions and movements, achieving a performance that is unparalleled in the industry.

From a technological standpoint, HeyGem.ai boasts an impressive inference speed of 1:0.5, with its ultra-high-definition 4K output at 32 frames per second surpassing even cinematic standards. This latest open-source model is poised to revolutionize the widespread application of digital humans across various sectors, including education, film and television, live streaming, and beyond, making them more accessible to the general public.