AB
AB
AiBoss

All Categories

project

PersonaLive - An open-source real-time AI face-swapping live streaming tool developed by the University of Macau and other institutions.

PersonaLive is an open-source real-time digital human live streaming tool developed by the University of Macau and GVC Labs. The tool can run on a single consumer-grade graphics card with 12GB of VRAM, enabling unlimited portrait animation generation and supporting low-latency interaction.

What is PersonaLive?

PersonaLive is an open-source real-time digital human live streaming tool developed by the University of Macau and GVC Labs. The tool can run on a single consumer-grade graphics card with 12GB of VRAM, enabling unlimited portrait animation generation and supporting low-latency interaction. By optimizing memory management and using TensorRT acceleration, PersonaLive solves the problems of high hardware requirements, memory overflow, and audio-visual desynchronization inherent in traditional digital human live streaming. PersonaLive easily enables AI-powered live commerce and virtual avatars to participate in events, bringing new possibilities to the field of digital human live streaming.

The main functions of PersonaLive

  • Low hardware requirementsConsumer-grade graphics cards that only require a single 12GB of video memory (such as the RTX 3060 and 4070) significantly lower the entry barrier.
  • Generate unlimited durationIt supports continuous generation of video streams, making it suitable for long-duration live streaming and solving the problems of memory explosion and timing inconsistency in traditional models.
  • Low latency and real-time interactionBy optimizing memory management and inference acceleration, it supports TensorRT acceleration, enabling audio-visual synchronization and real-time interaction.
  • Eco-friendlySupports ComfyUI and RTX 50 series graphics cards, keeping up with hardware and community trends.
  • WebUI Visual OperationIt provides a simple and easy-to-use web interface, allowing users to upload photos, adjust parameters, and generate animations in real time.

The technical principles of PersonaLive

  • Streaming Diffusion FrameworkBased on the diffusion model, it achieves the generation of continuous video streams by optimizing memory management and inference acceleration, rather than the traditional segmented video generation.
  • Video memory optimizationThe tool employs efficient video memory management technology to ensure stable operation over long periods of time with limited video memory and to avoid video memory overflow.
  • Timing moduleIt introduces a special timing module that supports the generation of video streams of unlimited duration, making it suitable for long-duration live streaming scenarios.
  • TensorRT AccelerationIt supports converting models to TensorRT format, improving inference speed by about 2 times and significantly reducing latency.

PersonaLive project address

  • GitHub repositoryhttps://github.com/GVCLab/PersonaLive
  • HuggingFace model libraryhttps://huggingface.co/huaichang/PersonaLive
  • arXiv technical paper: https://arxiv.org/pdf/2512.11253

Application scenarios of PersonaLive

  • AI virtual anchorIt is used for e-commerce live streaming, news broadcasting and entertainment live streaming to achieve low-cost, high-quality 24-hour uninterrupted virtual character live streaming.
  • Virtual meetings and remote collaborationProvide virtual avatars in video conferencing and remote teaching to protect privacy and enhance the interactive experience.
  • Content creation and entertainmentTools can help create short videos, movies, games, and other content, enriching the formats and improving production efficiency.
  • Customer service and supportIt supports the creation of virtual customer service and intelligent assistants, enabling 24-hour online service and enhancing customer experience.
  • Education and TrainingIt generates virtual teachers and training simulation scenarios, providing personalized and efficient teaching and training services.
Back