Meta, Facebook's parent corporation, has revealed a new tool that will allow users to make artificial intelligence-generated videos
Image Source: Google
In the newest contribution to the ongoing drive for AI-generated art, Make-a-Video, which is not yet publicly available, will allow users to make films from a text prompt
Image Source: Google
The movies are only five seconds long and have no soundtrack, but they represent a fundamental shift in AI-generated art from static images to video snippets
Image Source: Google
The programme was created by a team of Meta machine learning engineers, who released a white paper on their study findings on the Cornell University website
Image Source: Google
The business has also made available movies created using the tool as well as the written instructions used to make them
Image Source: Google
"This is incredible improvement. Video generation is substantially more difficult than photo generation because, in addition to accurately creating each pixel, the algorithm must forecast how they will change over time "In a statement, Mark Zuckerberg, co-founder of Facebook and CEO of Meta, said
Image Source: Google
"Make-A-Video addresses this by incorporating an unsupervised learning layer that allows the system to comprehend motion in the real environment and apply it to traditional text-to-image production. We want to make this available as a demo in the future. Meanwhile, enjoy the vids."
Image Source: Google