Google’s Imagen AI text-to-video generator is coming soon to their Test Kitchen app for beta testers. Sound familiar? That’s because it debuted months ago for images, but now, the model has been updated to use Phenaki, a model for generating super-resolution videos from text.
Researchers managed to scale up the Imagen system as a high definition text-to-video model including design decisions such as the choice of fully-convolutional temporal and spatial super-resolution models and the choice of the v-parameterization of diffusion models. They discovered that Imagen Video is not only capable of generating high fidelity videos, but also features a high degree of controllability and world knowledge. Still interested? You can get AI Test Kitchen for iOS devicesd and on Google Play now.
- LIGHT AND PORTABLE: With its compact 8.7” screen, slim design and sturdy metal frame, Galaxy Tab A7 Lite tablet is perfectly sized for entertainment...
- LASTING PROTECTION: This device features an upgraded metal frame that helps protect against everyday hiccups; It's made to be durable, so your tablet...
- FASTER PERFORMANCE: Galaxy Tab A7 Lite has faster speed than previous models for smooth streaming with minimal interruptions; Also packs more...
Visualization of Imagen. Imagen uses a large frozen T5-XXL encoder to encode the input text into embeddings. A conditional diffusion model maps the text embedding into a 64×64 image. Imagen further utilizes text-conditional super-resolution diffusion models to upsample the image 64×64→256×256 and 256×256→1024×1024,” said Google.
