# vita-epfl/Stable-Video-Infinity

[ICLR 26 Oral] Stable Video Infinity: Infinite-Length Video Generation with Error Recycling

Repository: https://github.com/vita-epfl/Stable-Video-Infinity
Canonical: https://ross.abutalabs.com/products/stable-video-infinity
Homepage: https://stable-video-infinity.github.io/homepage/
Language: Python
License: MIT
License Family: permissive
Topics: long-video-generation, end-to-end-filming, audio-driven-talking-face, dance-generation, video-diffusion-transformers
Last push: 2026-06-03T07:30:44+00:00

## Health v2 (maintenance only)
Score: 55/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 85, release rhythm 35, longevity 23
- inputs: {"age_days": 328, "days_push": 91, "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 2556, forks 227 (observed 2026-08-28T04:07:00.759111+00:00)

## What it is
Stable Video Infinity (SVI) is a research codebase for infinite-length video generation using video diffusion transformers with an error-recycling training approach. It supports streaming text-to-video short-film generation, audio-driven talking faces, and dance generation.

## Use cases
- generate infinitely long videos from a stream of text prompts
- create short films with per-clip storyline control for vlogs or TikTok
- generate audio-driven talking-face videos
- generate dance videos
- run long-form video diffusion research experiments

## When to avoid
- you need a production-ready commercial video generation service (academic use only)
- you lack a GPU or cannot run large diffusion transformer models locally
- you need simple one-click video editing rather than generative modeling

## Facets
- artifact type: library
- maturity: active
- function: video-processing, machine-learning, deep-learning, llm-inference
- domain: artificial-intelligence, deep-learning, media
- platform: python
- tags: video-generation, diffusion-transformers, long-video, text-to-video, error-recycling, talking-face, dance-generation, research, iclr-2026, video, gpu, linux

## Member repositories
- vita-epfl/Stable-Video-Infinity (main) score 55

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:00.759111+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 2026-08-30T02:23:40.191575+00:00, confidence not recorded.
  - readme: https://github.com/vita-epfl/Stable-Video-Infinity (fetched 2026-08-28T04:07:00.759111+00:00, sha 0ccfcdc55a30)
  - homepage: https://stable-video-infinity.github.io/homepage/ (fetched 2026-08-29T10:06:21.071658+00:00, sha 514558be98ad)
- Data as of 2026-08-30T08:39:29.467469+00:00.