On March 20th, Step-Star Intelligence Technology Co., Ltd. announced the open-sourcing of its latest image-to-video model, Step-Video-TI2V. Trained with an impressive 30 billion parameters, this model excels at generating 5-second videos comprising 102 frames at a resolution of 540P. Key highlights of Step-Video-TI2V include adjustable motion amplitude, controllable camera movements, and inherent capabilities for generating special effects. Furthermore, the model has been optimized for Huawei's Ascend computing platform and is now accessible to the ModelArts community. The release of Step-Video-TI2V introduces a new era of flexibility and efficiency for video creators.
