Microsoft introduces Phi Silica: a 3.3 billion-parameter AI model that transforms efficiency and performance in personal computing

ai models have become essential for improving computing efficiency, productivity, and user experiences. The development of small language models (SLM) is a key approach, enabling more efficient processing on personal computing devices.

The problem the researchers are addressing is the high computational demand of ai models, which often require substantial power and resources, limiting their implementation on smaller devices such as personal computers. Reducing resource consumption while maintaining high performance is crucial to seamlessly integrating ai into everyday computing.

Existing methods involve the use of larger ai models that consume significant computational power, which can impact the overall performance of personal computers. These models rely heavily on the central processing unit (CPU) and graphics processing unit (GPU), which can slow down other tasks and reduce efficiency.

Microsoft researchers presented ai/apis/phi-silica”>silica phi, a small language model designed specifically for the Neural Processing Units (NPUs) in your new Copilot+ PCs. Phi Silica is part of the Phi family of models and is intended to deliver high-performance artificial intelligence capabilities with minimal power consumption. This design allows the CPU and GPU to remain available for other tasks, improving the overall computing experience.

Phi Silica stands out for its 3.3 billion parameters, making it the smallest model in the Phi family. Despite its compact size, Phi Silica achieves impressive performance metrics. It has a first token latency of 650 tokens per second and consumes only 1.5 watts of power. This efficiency ensures that the PC's CPU and GPU are not overloaded, allowing for smoother operation of other applications. Phi Silica token generation also reuses the NPU's KV cache and runs on the CPU, producing approximately 27 tokens per second.

Developers can access the Phi Silica API through the Windows App SDK and other ai-powered features such as optical character recognition (OCR), studio effects, live captions, and video activity retrieval API. user. This integration allows developers to create innovative experiences that leverage ai within the Windows ecosystem. Microsoft plans to release additional APIs, including Vector Embedding, RAG API, and Text Summarization, further expanding the capabilities available to developers.

Phi Silica joins the ranks of other Phi-3 series models, including the Phi-3-mini with 3.8 billion parameters, the Phi-3-small with 7 billion parameters, the Phi-3-medium with 14 billion parameters and the recently announced Phi-3-vision with 4.2 billion parameters. However, Phi Silica is unique as the first next-generation SLM to ship with Windows, marking an important milestone in bringing advanced ai capabilities directly to end users.

The introduction of Phi Silica follows Microsoft's announcement of the Copilot+ PC, which promises to offer Windows PCs equipped with dedicated ai processors. The first Copilot+ PCs will launch in mid-June and will feature Qualcomm's Arm-based Snapdragon x Elite and Plus chips. Microsoft will offer these ai-powered laptops in collaboration with leading PC manufacturers throughout the summer. Intel is also developing its Copilot+ PC processor, codenamed Lunar Lake, scheduled for release in the third quarter of 2024.

Phi Silica Key Features:

• Model size and efficiency: Phi Silica is the smallest model in the Phi family, with 3.3 billion parameters. It offers high performance with a first token latency of 650 tokens per second and consumes only 1.5 watts of power, ensuring minimal resource usage on the PC's CPU and GPU.
• Token generation: This feature uses the NPU's KV cache and runs on the CPU, producing approximately 27 tokens per second, improving the overall computing experience.
• Developer Integration: Developers can access the Phi Silica API through the Windows Application SDK. It includes capabilities such as OCR, studio effects, live captions, and user activity retrieval API, enabling innovative ai applications within the Windows ecosystem.
• Advanced ai capabilities: Phi Silica is the first next-generation small language model to ship with Windows, marking a major milestone in the accessibility of ai for end users and developers.
• Collaborative efforts: It launched alongside Microsoft's Copilot+ PCs, which feature Qualcomm's Snapdragon x Elite and Plus chips, and Intel's upcoming Lunar Lake processors. These ai-powered laptops will be available starting mid-June 2024.
• Performance and power usage: Designed to run efficiently on the NPUs of Copilot+ PCs, ensuring fast local inference while maintaining low power consumption, significantly elevating productivity and accessibility within the Windows platform.

In conclusion, Microsoft's development of Phi Silica addresses the critical challenge of resource consumption in ai models. By providing an efficient, high-performance model that operates within the limitations of personal computing devices, Phi Silica improves the user experience and paves the way for more innovative applications. This model enables the integration of ai into everyday computing, offering powerful tools without compromising system performance.