AI models have become crucial in enhancing computing efficiency, productivity, and user experiences. Researchers are focused on addressing the issue of high computational demand in AI models, which often require significant power and resources, limiting their use on smaller devices like personal computers. Microsoft researchers have introduced Phi Silica, a small language model tailored for Neural Processing Units (NPUs) in Copilot+ PCs. Phi Silica stands out for its 3.3 billion parameters, making it the smallest model in the Phi family. Despite its compact size, it boasts impressive performance metrics, with a first token latency of 650 tokens per second and consuming only 1.5 watts of power. This design allows the CPU and GPU to remain available for other tasks, improving overall computing efficiency.
Developers can access Phi Silica API through the Windows App SDK, enabling the integration of innovative AI experiences within the Windows ecosystem. Microsoft also plans to release additional APIs like Vector Embedding, RAG API, and Text Summarization to expand developers’ capabilities further. Phi Silica joins other Phi-3 series models, such as Phi-3-mini, Phi-3-small, Phi-3-medium, and Phi-3-vision, but it is unique as the first next-generation SLM to ship with Windows, marking a milestone in making advanced AI capabilities accessible to end-users.
Phi Silica was introduced alongside Microsoft’s Copilot+ PCs, equipped with dedicated AI processors and powered by Qualcomm’s Arm-based Snapdragon X Elite and Plus chips. The launch of Copilot+ PCs is expected in mid-June, with collaborations with leading PC manufacturers to offer AI-powered laptops. Intel is also developing its Copilot+ PC processor, Lunar Lake, scheduled for release in the third quarter of 2024.
In conclusion, Microsoft’s Phi Silica addresses the challenge of resource consumption in AI models by providing an efficient, high-performance model that operates within the constraints of personal computing devices. This advancement enhances the user experience and enables the integration of AI into everyday computing, offering powerful tools without compromising system performance. The introduction of Phi Silica and Copilot+ PCs signifies a significant step towards bringing advanced AI capabilities to end-users and developers within the Windows platform.
Article Source
https://www.marktechpost.com/2024/05/24/microsoft-introduces-phi-silica-a-3-3-billion-parameter-ai-model-transforming-efficiency-and-performance-in-personal-computing/