Meta Platforms Inc., the tech giant formerly known as Facebook, has made a leap in the world of artificial intelligence (AI) with the unveiling of two pioneering AI models: Emu Video and Emu Edit. These innovations, announced on November 16, 2023, are set to revolutionize the way we create and edit videos and images.
Emu Video is an advanced AI model capable of generating video clips based on text and image inputs. This technology leverages Meta’s previous Emu model and is built upon diffusion models. The Emu Video tool offers a straightforward method for text-to-video generation, opening up possibilities for content creators, marketers, and social media enthusiasts alike.
Emu Video uses just two diffusion models to generate four-second long videos at a resolution of 512×512 pixels and 16 frames per second. The tool’s effectiveness is highlighted by its preference among users over previous methods, with 96% of respondents favoring Emu Video over older tools based on quality, and 85% citing its faithfulness to text prompts.
Meanwhile, Emu Edit is an AI tool that allows users to alter images based on text inputs. This model is capable of free-form editing through text instructions, offering a new level of precision and control in image editing. The video demonstration for Emu Edit, shared by the company, showcases its potential to redefine the realm of image manipulation.
Unlike many generative AI models that tend to over-modify or underperform in various editing tasks, Emu Edit is designed to follow instructions precisely, ensuring that only the pixels relevant to the edit request are altered. This precision is made possible by a training dataset comprising 10 million synthesized samples, each including an input image, a task description, and the targeted output image.
The integration of these tools into Facebook and Instagram, as suggested by some reports, could be a game-changer in the social media landscape. It could empower users to create more engaging and personalized content, potentially driving user engagement and interaction on these platforms to new heights.
Yann LeCun, a renowned computer scientist and a key figure in the AI field, praised the developments, highlighting the capabilities of both Emu Video and Emu Edit for video generation and image animation.
