Skip to content
FastVideo
Esc
↑↓navigate↵open⌘Jpreview

Welcome to FastVideo

Run, train, and accelerate video generation models with FastVideo.

hao-ai-lab/FastVideo | Trendshift

FastVideo is an inference and post-training framework for video diffusion models. One pipeline covers data preprocessing, fine-tuning, distillation, and inference. Attention backends, parallelism, and quantization are separate modules, so you can add one without changing model code.

Start with a FastVideo modelFastH3 8-Step V2Configure a recipe →

Key Features

FastVideo has the following features:

Projects and research

Explore work using FastVideo’s code or the attention research behind it.

  • SGLang: SGLang’s diffusion inference functionality is based on a fork of FastVideo on Sept. 24, 2025.
  • DanceGRPO: A unified framework to adapt Group Relative Policy Optimization (GRPO) to visual generation paradigms. Code based on FastVideo.
  • SRPO: A method to directly align the full diffusion trajectory with fine-grained human preference. Code based on FastVideo.
  • DCM: Dual-expert consistency model for efficient and high-quality video generation. Code based on FastVideo.
  • HY-WorldPlay: An action-conditioned world model model trained using FastVideo framework.
  • Hunyuan Video 1.5: A leading lightweight video generation model, where they proposed SSTA based on Sliding Tile Attention.
  • Kandinsky-5.0: A family of diffusion models for video & image generation, where their NABLA attention includes a Sliding Tile Attention branch.
  • LongCat Video: A foundational video generation model with 13.6B parameters with block-sparse attention similar to Video Sparse Attention.