Hangzhou, May 15: Alibaba has announced the release of Wan 2.1-VACE (Video All-in-one Creation and Editing), a powerful open-source model designed to revolutionize video generation and editing by merging multiple video processing tasks into one unified platform.
Part of the Wan2.1 series, VACE is the first open-source model of its kind to offer an all-in-one solution for video creation. It supports multi-modal input—including text, images, and video—and provides a broad suite of editing features. These include image and frame referencing, selective video modification, video repainting, motion and pose control, spatio-temporal extension, and content-aware video boundary expansion.
With these capabilities, users can animate static images, generate videos featuring interacting characters based on image samples, and replace or modify objects without affecting surrounding areas. The model also allows for precise motion control and recolourisation, as well as the ability to expand a vertical image into a horizontal video.
Wan2.1-VACE features a unified interface called the Video Condition Unit (VCU), which facilitates streamlined input processing. It also incorporates a Context Adapter structure to manage temporal and spatial data, supporting a wide variety of video synthesis tasks.
Ideal for short-form content, marketing, film post-production, and educational training, VACE is available in two versions—a 14-billion-parameter and a 1.3-billion-parameter model. Both can be downloaded for free via Hugging Face, GitHub, and Alibaba Cloud’s ModelScope.
The launch follows Alibaba’s recent open-sourcing of four Wan2.1 models in February and another video generation model in April. These models have collectively surpassed 3.3 million downloads, reflecting strong interest from global developers and creators.
Media & PR: editor@dailystraits.com. Copyright 2021–Present DailyStraits.com. All rights reserved.