# fudan-generative-vision/hallo3

[CVPR 2025] Hallo3: Highly Dynamic and Realistic Portrait Image Animation with Video Diffusion Transformer

Repository: https://github.com/fudan-generative-vision/hallo3
Canonical: https://ross.abutalabs.com/products/hallo3
Language: Python
License: MIT
License Family: permissive
Last push: 2025-03-13T08:34:43+00:00

## Health v2 (maintenance only)
Score: 26/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 11, release rhythm 35, longevity 46
- inputs: {"age_days": 645, "days_push": 538, "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 1401, forks 181 (observed 2026-08-28T04:04:37.272615+00:00)

## What it is
Hallo3 is a research model from Fudan University that animates a single portrait image into a highly dynamic and realistic talking-head video using a video diffusion transformer, driven by audio. It is the CVPR 2025 release in the Hallo series with pretrained weights on Hugging Face.

## Use cases
- animate a portrait photo into a talking video
- generate talking head videos from audio
- create lip-synced avatar videos from a single image
- audio-driven portrait animation research
- make a digital human video from one photo
- video diffusion model for face animation

## When to choose
- you need state-of-the-art audio-driven portrait animation from a single image
- you want expressive, dynamic head and body motion beyond simple lip-sync
- you are doing research on video diffusion transformers

## When to avoid
- you need real-time or low-latency animation
- you lack a GPU or cannot run heavy diffusion inference
- you need a production-ready API or GUI rather than a research codebase

## Facets
- artifact type: library
- maturity: active
- function: video-processing, deep-learning, image-processing, machine-learning
- domain: computer-vision, deep-learning, artificial-intelligence
- platform: python
- tags: talking-head, portrait-animation, diffusion-transformer, video-generation, audio-driven, research, cvpr-2025, video, gpu, linux

## Member repositories
- fudan-generative-vision/hallo3 (main) score 26

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:37.272615+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 2026-08-30T04:39:02.403481+00:00, confidence not recorded.
  - readme: https://github.com/fudan-generative-vision/hallo3 (fetched 2026-08-28T04:04:37.272615+00:00, sha 87aa0afdf2df)
- Data as of 2026-08-30T08:39:29.467469+00:00.