Code Availability and Model Weights
The code for Stable Video Diffusion has been made available on Stability AI's GitHub repository. Additionally, the weights required to run the model locally can be accessed via their Hugging Face page.
Stable Video Diffusion is the first foundational model for generative video based on the image model Stable Diffusion released by Stability AI. It's an advanced generative AI video model, representing a significant step towards creating various types of models for everyone. Stable Video Diffusion can adapt to various downstream tasks, including multi-view synthesis from a single image and fine-tuning on multi-view datasets. It is released in two image-to-video model forms, capable of generating 14 to 25 frames at customizable frame rates between 3 to 30 frames per second. Stable Video Diffusion is part of Stability AI's diverse open-source model series, covering image, language, audio, 3D, and code, demonstrating Stability AI's commitment to enhancing human intelligence.
The code for Stable Video Diffusion has been made available on Stability AI's GitHub repository. Additionally, the weights required to run the model locally can be accessed via their Hugging Face page.
The video model is adaptable to a range of downstream tasks, including multi-view synthesis from a single image with fine-tuning on multi-view datasets. Stability AI plans to develop various models that build upon and extend this foundational model.
A new web experience featuring a Text-To-Video interface is being developed. This tool demonstrates the practical applications of Stable Video Diffusion in sectors like Advertising, Education, Entertainment, and more.
Stable Video Diffusion is available in two image-to-video model forms, capable of generating 14 and 25 frames at customizable frame rates between 3 and 30 frames per second.
The model is in the research preview stage and is not intended for real-world or commercial applications yet. Insights and feedback on safety and quality are crucial for refining the model for its eventual release.
It is a significant addition to Stability AI’s diverse range of open-source models, which span image, language, audio, 3D, and code, showcasing their commitment to amplifying human intelligence.