DeeperML

Meta and University of Waterloo researchers have developed MoCha, an AI system capable of generating complete character animations with synchronized speech and natural movements. Unlike previous models that primarily focused on facial animation, MoCha can render full-body movements from various camera angles. This includes lip synchronization, gestures, and interactions between multiple characters.

Early demonstrations of MoCha feature close-up and semi-close-up shots, showcasing the system's ability to generate upper body movements and gestures that align with spoken dialogue. MoCha utilizes a diffusion transformer model with 30 billion parameters to produce HD video clips approximately five seconds long at 24 frames per second. For scenes with multiple characters, users can define characters once and refer to them with simple tags, streamlining the prompting process.

ImgSrc: the-decoder.com

References :

• THE DECODER: Meta and University of Waterloo researchers have built MoCha, an AI system that generates complete character animations with synchronized speech and natural movements.
• MarkTechPost: Salesforce AI Introduce BingoGuard: An LLM-based Moderation System Designed to Predict both Binary Safety Labels and Severity Levels.
• Salesforce: Salesforce believes that AI agents are only as good as the data they’re able to access, but many organizations are still struggling to bring it all together.

Classification:

• HashTags: #MoCha #SalesforceAI #VideoGeneration
• Company: Salesforce
• Product: MoCha
• Feature: Animation
• Type: AI
• Severity: Informative