OpenAI has released the long-awaited neural network model for video generation


The developer says it is taking a number of security steps before making Sora available to the general public.

OpenAI makes another revolution and releases the neural network Sora, which generates high-quality videos based on text description. The company published information about the development on its official website. There you can also see successful examples of generation.

Sora Neural Network

“We imagine Sora – our model. It can create videos up to one minute long, maintaining visual quality and matching the user’s request,” the developer’s website says.

Currently, Sora is available in closed beta to a very small number of users, but soon it will be available for everyone.

Read also: Improving photos on a smartphone

The model has a deep understanding of language, which allows her to accurately interpret requests and create convincing characters that express vivid emotions. Sora can also create multiple shots in a single generated video that accurately preserves the characters and visual style.

The developer says it is taking a number of security steps before making Sora available to the general public. We work with a team of “red” testers – experts in areas such as misinformation, bias, and hateful content. They will test the model for strength.

In case you have found a mistake in the text, please send a message to the author by selecting the mistake and pressing Ctrl-Enter.



Source link