<!-- ## **HunyuanVideo** -->

中文阅读

<p align="center"> <img src="https://raw.githubusercontent.com/Tencent-Hunyuan/HunyuanVideo/refs/heads/main/assets/logo.png" height=100> </p>

HunyuanVideo: A Systematic Framework For Large Video Generation Model

<div align="center"> <a href="https://github.com/Tencent-Hunyuan/HunyuanVideo"><img src="https://img.shields.io/static/v1?label=HunyuanVideo Code&message=Github&color=blue"></a> &ensp; <a href="https://aivideo.hunyuan.tencent.com"><img src="https://img.shields.io/static/v1?label=Project%20Page&message=Web&color=green"></a> &ensp; <a href="https://video.hunyuan.tencent.com"><img src="https://img.shields.io/static/v1?label=Playground&message=Web&color=green"></a> </div> <div align="center"> <a href="https://arxiv.org/abs/2412.03603"><img src="https://img.shields.io/static/v1?label=Tech Report&message=Arxiv&color=red"></a> &ensp; <a href="https://aivideo.hunyuan.tencent.com/hunyuanvideo.pdf"><img src="https://img.shields.io/static/v1?label=Tech Report&message=High-Quality Version (~350M)&color=red"></a> </div> <div align="center"> <a href="https://huggingface.co/tencent/HunyuanVideo"><img src="https://img.shields.io/static/v1?label=HunyuanVideo&message=HuggingFace&color=yellow"></a> &ensp; <a href="https://huggingface.co/docs/diffusers/main/api/pipelines/hunyuan_video"><img src="https://img.shields.io/static/v1?label=HunyuanVideo&message=Diffusers&color=yellow"></a> &ensp; <a href="https://huggingface.co/tencent/HunyuanVideo-PromptRewrite"><img src="https://img.shields.io/static/v1?label=HunyuanVideo-PromptRewrite&message=HuggingFace&color=yellow"></a>

</div> <p align="center"> 👋 Join our <a href="assets/WECHAT.md" target="_blank">WeChat</a> and <a href="https://discord.gg/tv7FkG4Nwf" target="_blank">Discord</a> </p> <p align="center">

---

This repo contains PyTorch model definitions, pre-trained weights and inference/sampling code for our paper exploring HunyuanVideo. You can find more visualizations on our project page.

HunyuanVideo: A Systematic Framework For Large Video Generation Model <be>

🔥🔥🔥 News!!

• Nov 21, 2025: 🎉 We release the HunyuanVideo-1.5, a highly efficient and powerful new foundation model.
• May 28, 2025: 💃 We release the HunyuanVideo-Avatar, an audio-driven human animation model based on HunyuanVideo.
• May 09, 2025: 🙆 We release the HunyuanCustom, a multimodal-driven architecture for customized video generation based on HunyuanVideo.
• Mar 06, 2025: 🌅 We release the HunyuanVideo-I2V, an image-to-video model based on HunyuanVideo.
• Jan 13, 2025: 📈 We release the Penguin Video Benchmark.
• Dec 18, 2024: 🏃‍♂️ We release the FP8 model weights of HunyuanVideo to save more GPU memory.
• Dec 17, 2024: 🤗 HunyuanVideo has been integrated into Diffusers.
• Dec 7, 2024: 🚀 We release the parallel inference code for HunyuanVideo powered by xDiT.
• Dec 3, 2024: 👋 We release the inference code and model weights of HunyuanVideo. Download.

🎥 Demo

<div align="center"> <video src="https://github.com/user-attachments/assets/22440764-0d7e-438e-a44d-d0dad1006d3d" width="70%" poster="./assets/video_poster.png"> </video> </div>

🧩 Community Contributions

If you develop/use HunyuanVideo in your projects, welcome to let us know.

• ComfyUI-Kijai (FP8 Inference, V2V and IP2V Generation): ComfyUI-HunyuanVideoWrapper by Kijai

• ComfyUI-Native (Native Support): ComfyUI-HunyuanVideo by ComfyUI Official

• FastVideo (Consistency Distilled Model and Sliding Tile Attention): FastVideo and Sliding Tile Attention by Hao AI Lab

• HunyuanVideo-gguf (GGUF Version and Quantization): HunyuanVideo-gguf by city96

• Enhance-A-Video (Better Generated Video for Free): Enhance-A-Video by NUS-HPC-AI-Lab

• TeaCache (Cache-based Accelerate): TeaCache by Feng Liu

• HunyuanVideoGP (GPU Poor version): HunyuanVideoGP by DeepBeepMeep

• RIFLEx (Video Length Extrapolation): RIFLEx by Tsinghua University

• HunyuanVideo Keyframe Control Lora: hunyuan-video-keyframe-control-lora by dashtoon

• Sparse-VideoGen (Accelerate Video Generation with High Pixel-level Fidelity): Sparse-VideoGen by University of California, Berkeley

• FramePack (Packing Input Frame Context in Next-Frame Prediction Models for Video Generation): FramePack by Lvmin Zhang

• Jenga (Training-Free Efficient Video Generation via Dynamic Token Carving): Jenga by DV Lab

• DCM (Dual-Expert Consistency Model for Efficient and High-Quality Video Generation): DCM by Vchitect

📑 Open-source Plan

• HunyuanVideo (Text-to-Video Model)
  • [x] Inference
  • [x] Checkpoints
  • [x] Multi-gpus Sequence Parallel inference (Faster inference speed on more gpus)
  • [x] Web Demo (Gradio)
  • [x] Diffusers
  • [x] FP8 Quantified weight
  • [x] Penguin Video Benchmark
  • [x] ComfyUI
• HunyuanVideo (Image-to-Video Model)
  • [X] Inference
  • [X] Checkpoints

Contents

• HunyuanVideo: A Systematic Framework For Large Video Generation Model
  • 🎥 Demo
  • 🔥🔥🔥 News!!
  • 🧩 Community Contributions
  • 📑 Open-source Plan
  • Contents
  • Abstract
  • HunyuanVideo Overall Architecture
  • 🎉 HunyuanVideo Key Features
    • Unified Image and Video Generative Architecture
    • MLLM Text Encoder
    • 3D VAE
    • Prompt Rewrite
  • 📈 Comparisons
  • 📜 Requirements
  • 🛠️ Dependencies and Installation
    • Installation Guide for Linux
  • 🧱 Download Pretrained Models
  • 🔑 Single-gpu Inference
    • Using Command Line
    • Run a Gradio Server
    • More Configurations
  • 🚀 Parallel Inference on Multiple GPUs by xDiT
    • Using Command Line
  • 🚀 FP8 Inference
    • Using Command Line
  • 🔗 BibTeX
  • Acknowledgements
  • Star History

---

Abstract

We present HunyuanVideo, a novel open-source video foundation model that exhibits performance in video generation that is comparable to, if not superior to, leading closed-source models. In order to train HunyuanVideo model, we adopt several key technologies for model learning, including data curation, image-video joint model training, and an efficient infrastructure designed to facilitate large-scale model training and inference. Additionally, through an effective strategy for scaling model architecture and dataset, we successfully trained a video generative model with over 13 billion parameters, making it the largest among all open-source models.

We conducted extensive experiments and implemented a series of targeted designs to ensure high visual quality, motion diversity, text-video alignment, and generation stability. According to professional human evaluation results, HunyuanVideo outperforms previous state-of-the-art models, including Runway Gen-3, Luma 1.6, and 3 top-performing Chinese video generative models. By releasing the code and weights of the foundation model and its applications, we aim to bridge the gap between closed-source and open-source video foundation models. This initiative will empower everyone in the community to experiment with their ideas, fostering a more dynamic and vibrant video generation ecosystem.

HunyuanVideo Overall Architecture

HunyuanVideo is trained on a spatial-temporally compressed latent space, which is compressed through a Causal 3D