Stability AI Reveals State-of-the-Art AI Video Model – Stable Video Diffusion
Stability AI has launched its latest effort in the field of artificial intelligence with the introduction of “Stable Video Diffusion,” an AI model designed to generate videos by animating existing images.
Following the success of Stability’s prior text-to-image model, “Stable Diffusion,” this new endeavor marks a significant milestone in the company’s mission to democratize generative AI video models.
“In a major step forward for us, we are pleased to present this cutting-edge generative AI video model for research preview, marking a significant stride in our journey to create models for all types,” the company announced in a press statement.
Stability AI describes the Stable Video Diffusion as a state-of-the-art generative AI video model, expanding its capabilities across different modalities based on the existing Stable Diffusion text-to-image model. The company has made the code for Stable Video Diffusion readily available on its GitHub repository for interested users, along with comprehensive technical details in a dedicated research paper.
The model is equipped with two image-to-video models capable of generating 14 and 25 frames. Notably, the models offer customizable frame rates between 3 and 30 frames per second, providing users with creative flexibility.
While the company is excited about the potential applications of Stable Video Diffusion, it emphasizes that, at this stage, the model is not intended for real-world or commercial applications. Stability AI actively seeks user feedback to enhance the model’s safety and quality aspects.
Stability AI’s exploration of generative AI models extends to 3D model creation with the introduction of “Stable 3D,” targeting graphic designers, digital artists, and game developers to simplify the process of generating concept-quality textured 3D objects.
These latest developments from Stability AI highlight the company’s commitment to advancing the capabilities of AI in the realm of video and 3D model generation.
ⓒ 2023 TECHTIMES.com All rights reserved. Do not reproduce without permission.
