Stable Video Diffusion
video
open-source
Stable Video Diffusion (2023) is Stability AI's text-to-video model based on Stable Diffusion. It generates short video ...
Version: 1.0
Released: 2y 2m ago on 09/01/2023
Architecture
- parameters: 0
- context_length: 0
- training_data: Fine-tuned from image diffusion to handle video frames
- inference: video diffusion
Capabilities
- Text-to-video: generates multi-second video clips from a text prompt
- Supports customizable frame rates (30 fps) and outputs at 25 fps.
- Generates short video (2-4s) in under 2 minutes.
Benchmarks
- VideoSynth: N/A
Safety
- Capable of realistic video content
- follow content policies.
Deployment
- regions: private
- hosting: Offered under research license, Stability API
- integrations: integration via Stability API and cloud GPU.
Tags
video-generationdiffusionopen-source