2025-03-20 14:32:50.AIbase.16.5k
Motion Controllable! Step-Video-TI2V, an Open-Source Text-to-Video Model
Shanghai Step Star Intelligent Technology Co., Ltd. has announced the open-source release of its latest text-to-video model, Step-Video-TI2V. Trained on a 30B parameter Step-Video-T2V model, it generates 102-frame, 5-second videos at 540P resolution. Key features include controllable motion amplitude and camera movement, performing particularly well in generating anime-style videos.
