| --- |
| datasets: |
| - vita-video-gen/svi-benchmark |
| language: |
| - en |
| tags: |
| - video generation |
| pipeline_tag: image-to-video |
| library_name: diffusers |
| license: mit |
| project_page: https://stable-video-infinity.github.io/homepage/ |
| papers: |
| - title: 'Stable Video Infinity: Infinite-Length Video Generation with Error Recycling' |
| authors: |
| - Wuyang Li |
| - Wentao Pan |
| - Po-Chien Luan |
| - Yang Gao |
| - Alexandre Alahi |
| url: https://huggingface.co/papers/2510.09212 |
| conference: arXiv preprint, 2025 |
| --- |
| |
| <div align="center"> |
|
|
| <h1>Stable Video Infinity: Infinite-Length Video Generation with Error Recycling<h1> |
|
|
| <p align="center"> |
| <a href="https://huggingface.co/papers/2510.09212"> <img src="https://img.shields.io/badge/Paper-HuggingFace-red?logo=huggingface&logoColor=yellow" alt="Paper on Hugging Face"/> </a> |
| <a href="https://stable-video-infinity.github.io/homepage/"> <img src="https://img.shields.io/badge/Project-Page-green" alt="Project Page"/> </a> |
| <a href="https://github.com/vita-epfl/Stable-Video-Infinity"> <img src="https://img.shields.io/badge/SVI-GitHub-black?logo=github&logoColor=white" alt="SVI on GitHub"/> </a> |
| <a href="https://huggingface.co/datasets/vita-video-gen/svi-benchmark"> <img src="https://img.shields.io/badge/SVI_Dataset-Hugging%20Face-orange?logo=huggingface&logoColor=yellow" alt="SVI Dataset"/> </a> |
| <a href="https://huggingface.co/vita-video-gen/svi-model"> <img src="https://img.shields.io/badge/SVI_models-Hugging%20Face-FFCC00?logo=huggingface&logoColor=yellow" alt="SVI Models"/> </a> </p> </div> |
|
|
| ## 🎯 About This Repository |
| **Stable-Video-Infinity(SVI)** is able to generate ANY-length videos with high temporal consistency, plausible scene transitions, and controllable streaming storylines in ANY domains. |
| This repository contains the model weights of SVI Family. |
|
|
| ## 🌟 Key Highlights |
| - **OpenSVI**: Everything is open-sourced: training & evaluation scripts, datasets, and more. |
| - **Infinite Length**: No inherent limit on video duration; generate arbitrarily long stories (see the 10‑minute “Tom and Jerry” demo). |
| - **Versatile**: Supports diverse in-the-wild generation tasks: multi-scene short films, single‑scene animations, skeleton-/audio-conditioned generation, cartoons, and more. |
| - **Efficient**: Only LoRA adapters are tuned, requiring very little training data: anyone can make their own SVI easily. |
| ## 📦 Resources |
| | **Model** | **Task** | **Input** | **Output** | **Hugging Face Link** | **Comments** | |
| |-------|------|-------|--------|-------------------|------------------| |
| | **ALL** | Infinite possibility | Image + X | X video | [🤗 Folder](https://huggingface.co/vita-video-gen/svi-model/tree/main/version-1.0) |Family bucket! I want to play with all! | |
| | **SVI-Shot** | Single-scene generation | Image + Text prompt | Long video | [🤗 Model](https://huggingface.co/vita-video-gen/svi-model/resolve/main/version-1.0/svi-shot.safetensors?download=true) | Generate consistent long video with 1 text prompt. (This will never drift) | |
| | **SVI-Film** | Multi-scene generation | Image + Text prompt stream | Film-style video | [🤗 Model](https://huggingface.co/vita-video-gen/svi-model/resolve/main/version-1.0/svi-film.safetensors?download=true) | Generate creative long video with 1 text prompt stream (5 second per text). | |
| | **SVI-Film (Transition)** | Multi-scene generation | Image + Text prompt stream | Film-style video | [🤗 Model](https://huggingface.co/vita-video-gen/svi-model/resolve/main/version-1.0/svi-film-transitions.safetensors?download=true) |Generate creative long video with 1 text prompt stream. (More scene transitions due to the training data) | |
| | **SVI-Tom&Jerry** | Cartoon animation | Image | Cartoon video | [🤗 Model](https://huggingface.co/vita-video-gen/svi-model/resolve/main/version-1.0/svi-tom.safetensors?download=true) | Generate creative long cartoon videos with 1 text prompt stream (This will never drift in our 20 min test)| |
| | **SVI-Talk** | Talking head | Image + Audio | Talking video | [🤗 Model](https://huggingface.co/vita-video-gen/svi-model/resolve/main/version-1.0/svi-talk.safetensors?download=true) |Generate long videos with audio-conditioned human speaking | |
| | **SVI-Dance** | Dancing animation | Image + Skeleton | Dance video | [🤗 Model](https://huggingface.co/vita-video-gen/svi-model/resolve/main/version-1.0/svi-dance.safetensors?download=true) | Generate long videos with skeleton-conditioned human dancing | |
|
|
| Note: If you want to play with T2V, you can directly use SVI with an image generated by any T2I model! |
|
|
| ## 📝 Citation |
| If you find our work helpful for your research, please consider citing our paper. Thank you so much! |
|
|
| ```bibtex |
| @article{li2025stable, |
| title={Stable Video Infinity: Infinite-Length Video Generation with Error Recycling}, |
| author={Wuyang Li and Wentao Pan and Po-Chien Luan and Yang Gao and Alexandre Alahi}, |
| journal={arXiv preprint arXiv: arXiv:2510.09212}, |
| year={2025}, |
| url={https://huggingface.co/papers/2510.09212}, |
| } |
| ``` |
