Stability AI announced the release of Stable Video Diffusion, a generative AI video model that can transform static images into video content.
As marketers look for innovative ways to create visual content, we offer a sneak peek into the future of video content creation with generative artificial intelligence (AI).
What Is Stable Video Diffusion?
Stable Video Diffusion is a foundation model in research preview from Stability AI with image-to-video capability.
It was designed to perform tasks like multi-view synthesis from a single image, a capability enhanced by finetuning on multi-view datasets.
Stability AI offers two versions, capable of generating 14 and 25 frames at frame rates ranging from 3 to 30 frames per second.
While the company is enthusiastic about incorporating feedback and updating the models with the latest advancements, it clarified that the model is not intended for real-world or commercial applications at this stage.
Stable Video Diffusion’s code is available on GitHub, and the necessary weights to run it locally can be found on the Hugging Face page. This paper includes detailed technical capabilities of the new model.
Screenshot from Stability AI, November 2023How To Create Video From A Static Image: 3 Free AI Video Generator Demos
The following research demos powered by Stable Video Diffusion offer a glimpse into the future of visual content creation with generative AI.
1. Community Demo For Stable Video Diffusion – Img2Vid – XT On Hugging Face
I used this demo to create a downloadable GIF file from a ChatGPT-generated logo.
Screenshot from Hugging Face, November 2023The resulting AI-generated video:
2. SVD On Replicate
I used this demo to generate a downloadable MP4 file from the same logo.
Screenshot from Replicate, November 2023The resulting AI-generated video:
3. Stable Video Diffusion Playground On Fal
I tried this demo to create a downloadable GIF file.
Screenshot from Fal, November 2023Unfortunately, there were technical difficulties with the logo test. So here is an AI-generated GIF I created a few days ago using this demo and an image generated by DALL-E 3.
I asked @ChatGPTapp to put a lion’s face into a rose I photographed this morning using DALL·E 3 and animated it with @fal_ai_data using Stable Video Diffusion from @StabilityAI 🦁🏵️🔥 pic.twitter.com/S6jfKAEWK9
— Kristi Hines (@kristileilani) November 25, 2023
Waitlist For Text-To-Video Interface
Interested individuals and organizations can sign up for a waitlist to access a new upcoming web experience from Stability AI featuring a text-to-video interface, which will showcase the practical applications of this technology.
The release of Stable Video Diffusion marks a significant step in the evolution of generative AI technology, paving the way for future innovations in marketing and advertising technology.
Featured image: rafapress/Shutterstock
