In the fast-paced world of AI, whereas text-to-image fashions have been quickly advancing, text-to-video AI instruments have lagged behind. There are just a few video generator instruments out there, however solely Runway’s Gen2 and Pika Labs have managed to provide actually compelling outcomes.
Today, Meta AI introduced their very own model of an AI video generator known as Emu Video, and it seems to be superb.
What is Emu Video?
Emu Video, an extension of the Emu mannequin recognized for picture technology, brings an revolutionary method to text-to-video technology. It leverages diffusion fashions in a method that’s each easy and extremely efficient.
The mannequin is educated on the biggest dataset of its variety—10 million synthesized samples with an enter picture, job description, and goal output picture. This makes it the biggest dataset of its variety up to now.
Here are some examples:
