Alibaba Group HoldingsWednesday's AI and Cloud Computing Unit Released the latest open source, WAN2.2-S2V tool artificial intelligence A model that generates expressive film-quality character videos from static images and audio clips.
The new model forms part of Alibaba Cloud's WAN2.2 FamilyLast month, the company promoted it as the AI industry's first open source, large-scale video generation model, and incorporated the so-called expert (MOE) architecture. Hangzhou– The post is owned by the base Alibaba.
The WAN2.2-S2V model, powered by advanced audio-driven animation technology, “provides realistic character performance, from natural dialogue to musical performance, seamlessly handling multiple characters within a scene,” Alibaba Cloud said Wednesday.
WAN2.2-S2V can be used by professional content creators to “capture accurate visual representations tailored to specific storytelling and design requirements.” He added that the enhancement stems from a large audiovisual dataset of models tailored to film and television production scenarios.
The latest WAN2.2 variant reflects the way Chinese AI companies continue to narrow the gap with their US peers through an open source approach.
WAN2.2-S2V is now available for download from online developer platforms Hugging my face and githuband from the Alibaba Cloud ModelScope Open source community.
