Follow Us

Meta Launches Gen-AI Models for Generating Video and Editing 

Share on facebook
Share on twitter
Share on linkedin

Share

Meta Launches Gen-AI Models for Generating Video and Editing 
Share on facebook
Share on twitter
Share on linkedin

Meta, the leading social media giant company that turned towards the metaverse and is now bullish on artificial intelligence, recently launched two AI models. These models are meant to be deployed for its social media platforms Facebook and Instagram. 

In an announcement made on November 16, Meta noted the two generative AI models are: Emu Video and Emu Edit. Both the models will perform namesake tasks where the former is meant to generate video clips and the latter will be used for image editing. 

Meta already has an Emu model and Emu Video is going to leverage the same model to generate video clips based on text and image inputs given to it. On the other hand, the AI-based editing tool, Emu edit will focus on image manipulation and ensures precise image editing. 

The social media company noted in the blog post that both the generative AI models are suitable for creators, animators, and artists. It also highlighted the training model for both video and editing AI tools. 

According to Meta, Emu video is trained in two separate phases where dividing the process into different chunks makes it easier. These two steps include: generating the image based on text prompts and then generating video from both the texts and the generated image. 

With the diffusion model, rather than deploying tricky training procedures, it became easier to train on diffusion models. Meta disclosed that the Emu video used 2 diffusion models and generated four-second-long videos at 16 frames per second (FPS). 

On the other hand, the AI-based image editing tool Emu edit mainly focuses on image manipulation, adding or removing background, coloring and enhancing the geometry, etc. The company stated to keep “precision” in focus while training the AI model. 

Meta noted in the blog post that rather than just producing a ‘believable’ image, the AI editing model should take care of “precisely altering” the pixels that are relevant to the given prompt for editing requests. 

“For instance, when adding the text ‘Aloha!’ to a baseball cap, the cap itself should remain unchanged” it added. 

Meta Advancing Towards Emerging AI Tech

Almost every other tech company is hoarding towards artificial intelligence and Meta would not want to stay behind. The deployment of gen AI tools to perform various tasks on its social media platform is the result of the same efforts. 

Earlier in September this year, Meta announced its own AI chatbot. Dubbed the “Meta AI Chatbot”, the AI assistant was seen as a contender to the popular chatbot from OpenAI, ChatGPT. The AI model could be deployed to Facebook, Instagram, and Whatsapp-like social media platforms and is also planned to be added to mixed reality devices of the company down the road. 

Leave a Reply

Your email address will not be published. Required fields are marked *

Download our App for getting faster updates at your fingertips.

en_badge_web_generic.b07819ff-300x116-1

We Recommend

Top Rated Cryptocurrency Exchange

-
00:00
00:00
Update Required Flash plugin
-
00:00
00:00