Meta company from the system Artificial intelligence to convert text to video It revealed that it can start making a video by receiving a simple descriptive text.
Meta unveiled its new AI system that can convert text to video. This system is similar to Dall-E, but when you give it text, instead of turning your description into an image, it creates a video for you.
Read more: Image creation from text with DALL-E artificial intelligence (DALL-E) is coming soon
Details of the artificial intelligence system for converting text to video
Mark Zuckerberg published a video of himself drawing a teddy bear, from this app that apparently Make-A-Video named, unveiled. The video also features a horse drinking water, a flying saucer flying near the ground, an artist with a canvas, and a flying superhero dog. All of these videos are created with simple text descriptions.
“The performance of this system is amazing. Creating a video is much more difficult than an image, because in addition to trying to accurately reproduce each pixel, the system must predict how it will change over time. “Make-A-Video does this by adding supervised layers, allowing the system to understand movements in the real world and use these learnings in conventional text-to-image algorithms.”
Meta’s manager says the company will share a full version of the system in the future, but it’s unclear when users will be able to try out the AI. But overall, the videos seem to be short and high-quality, made with the help of the company’s recent advances in reproductive technologies. Meta said in a statement about its new AI:
“Make-A-Video uses public data sets to provide a higher level of transparency in research. “We freely conduct reproductive AI studies and their results to receive feedback from users using AI methods.”
What do you think about Meta company’s text-to-video artificial intelligence project? Share your thoughts with us technology news Follow from Tikrato.