Meta presented AI models for photo and video processing
Meta Corporation announced Emu Video and Emu Edit, generative AI tools for editing and creating content.
Both neural networks, based on the Emu language model, are still at the testing stage. According to the statement, the products are already demonstrating potential benefits for artists, animators and other creative professionals.
Emu Video is capable of generating video clips based on entered text and attached images in 512x512 format at 16 frames per second.
The neural network was trained using a “factorized” approach, dividing the process into two stages, which allowed the tool to respond to different inputs.
“First, [Emu Video] creates images based on the text prompt, and then generates a video based on that and the text. This “factorized” or split generation approach allows us to efficiently train models for videos,” explained Meta
Emu Edit allows you to remove or add a background to a photo, perform color and geometry conversions, and supports local and global editing.
The AI is trained on a data set of 10 million samples, each of which has an input image and a description of the task, as well as a target result.
“While Emu Video, Emu Edit and similar new technologies certainly cannot replace professional artists, they will help people express themselves in new ways: from an art director coming up with a new concept or a video editor bringing their latest video to life, to their best friend , who shares a unique birthday greeting,” the company emphasized.
Meta previously presented a set of AI tools in the form of a voice assistant, a neural network with different personalities, smart glasses and a sticker generator.
Let us recall that in August the corporation announced the AudioCraft neural network, which creates sounds and music based on text descriptions.
#Binance #MetaverseInvestment #etf #ETH #NFT
$MATIC $ETH $BNB
