# GVCLab/PersonaLive

[CVPR 2026] PersonaLive! : Expressive Portrait Image Animation for Live Streaming

Repository: https://github.com/GVCLab/PersonaLive
Canonical: https://ross.abutalabs.com/products/personalive
Homepage: https://arxiv.org/abs/2512.11253
Language: Python
License: Apache-2.0
License Family: permissive
Topics: cvpr, cvpr2026, talking-head, video-generation
Last push: 2026-05-15T07:54:56+00:00

## Health v2 (maintenance only)
Score: 53/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 82, release rhythm 35, longevity 20
- inputs: {"age_days": 281, "days_push": 110, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3552, forks 502 (observed 2026-08-28T04:08:09.600494+00:00)

## What it is
PersonaLive is a diffusion-based framework for real-time, streamable portrait image animation, generating infinite-length expressive talking-head videos from a single reference image. It combines hybrid implicit facial signals, appearance distillation, and autoregressive micro-chunk streaming to achieve 7-22x speedups over prior diffusion portrait animation models.

## Use cases
- animate a portrait photo into a talking-head video
- generate real-time avatar animation for live streaming
- create long talking videos from a single image on 12GB VRAM
- drive facial expressions and head motion from audio or keypoints
- integrate portrait animation into ComfyUI workflows
- research streaming diffusion video generation

## When to choose
- you need real-time or low-latency talking-head generation for streaming
- you want infinite-length portrait animation with limited GPU memory
- you need state-of-the-art expressive portrait animation quality
- you want a CVPR-published model with pretrained weights and training code

## When to avoid
- you need production-grade licensed software for commercial use (academic-only disclaimer)
- you have no GPU or limited VRAM below 12GB
- you need non-portrait video generation
- you require guaranteed support or stability for mission-critical systems

## Facets
- artifact type: library
- maturity: active
- function: video-processing, computer-vision, deep-learning, machine-learning, image-processing
- domain: computer-vision, deep-learning, artificial-intelligence, media
- platform: python, cross-platform
- tags: talking-head, portrait-animation, live-streaming, diffusion-models, real-time, cvpr2026, video-generation, streaming-inference, video, gpu, linux

## Member repositories
- GVCLab/PersonaLive (main) score 53

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:09.600494+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:34:19.204926+00:00, confidence not recorded.
  - readme: https://github.com/GVCLab/PersonaLive (fetched 2026-08-28T04:08:09.600494+00:00, sha 8eb4d23cfdf7)
  - homepage: https://arxiv.org/abs/2512.11253 (fetched 2026-08-29T09:28:16.343649+00:00, sha e50d764184f6)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T09:28:16.352945+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T09:28:16.358167+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T09:28:16.360642+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T09:28:16.355717+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.