EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·29d ago·5 min read
Apple wants to train AI on your private personal data
TL;DR
Apple has introduced its third generation of Foundation Models (AFM), designed to enhance user experiences while prioritizing privacy. These models, developed in collaboration with Google, include both on-device and server-based options.
✦ Why It Matters
Explore how to integrate Apple’s Foundation Models into your applications for enhanced user experiences.
Key Takeaways
- Utilize AFM 3 Core Advanced for improved on-device AI capabilities.
- Leverage AFM 3 Cloud for enhanced server-side processing without compromising privacy.
- Incorporate new image editing tools powered by ADM 3 Cloud in applications.
- Stay updated on privacy practices to ensure user data protection.
How It Works
AFM 3 Core Advanced employs a sparsely activated architecture, activating only a portion of its parameters based on the task, which optimizes memory usage and processing speed. This allows the model to handle complex requests efficiently by loading weights incrementally, rather than requiring all parameters to be active simultaneously.