Recently, Step Yuexingchen and Geely Automobile Group jointly announced a major technological breakthrough - two Step Step Series multimodal models, namely Step-Video-T2V video generation model and Step-Audio voice model. This collaboration marks another leap in the fields of video generation and voice processing, providing developers with powerful tool support.
Among them, the Step-Video-T2V video generation model is leading the world with its 30 billion parameters and excellent performance. This model can directly generate high-quality videos with 204 frames and 540P resolution, ensuring the information density and consistency of the generated content. The evaluation results show that Step-Video-T2V performs excellently in command compliance, motion smoothness, physical rationality and aesthetics, significantly surpassing the existing open source video model.
At present, these two models are now available on Yuewen App, and developers can experience it for free and provide valuable suggestions. Step-Video-T2V video generation model demonstrates excellent generation ability in complex movements, beautiful characters and visual imagination. It can accurately understand the instructions and help video creators achieve creative presentation efficiently. Whether it is elegant ballet, intense karate confrontation, or tense badminton games and high-speed flip diving, Step-Video-T2V can generate real and physically consistent pictures.
In addition, the model also supports a variety of lens movement modes and scene types to generate visual effects of large-scale mirror movement. The generated characters are more realistic and vivid, with rich details and natural expressions, providing more possibilities for video creation.
Developers can obtain more technical details and resources through the following links:
GitHub: https://github.com/stepfun-ai/Step-Audio
Hugging Face: https://huggingface.co/collections/stepfun-ai/step-audio-67b33accf45735bb21131b0b
Technical report: https://github.com/stepfun-ai/Step-Audio/blob/main/assets/Step-Audio.pdf