Introduction to Microsoft Phi-3
Microsoft Phi-3, a family of open AI models, has recently been introduced as a highly capable and cost-effective small language model (SLM). The Phi-3 mini is a 3.8B language model, setting a new standard by supporting a context window of up to 128K tokens with minimal impact on quality. It is an instruction-tuned model designed to reflect natural human communication patterns.
Features of Microsoft Phi-3
The Phi-3 model is accessible on Azure AI, offering a deploy-eval-finetune toolchain, and on Ollama for local development on individual laptops. It has been specifically optimized for ONNX Runtime, providing support for Windows DirectML and compatibility across various hardware platforms such as GPU, CPU, and mobile devices. Moreover, it is available as an NVIDIA NIM microservice with a standard API interface for deployment in diverse environments.
Expansion of Microsoft Phi-3 Family
In the upcoming weeks, Microsoft plans to introduce additional models to the Phi-3 family, enhancing flexibility and choices for customers. Models such as Phi-3 small (7B) and Phi-3 medium (14B) will soon be accessible through the Azure AI model catalog and other platforms. Microsoft's commitment to delivering quality models at competitive prices is evident through the Phi-3 series.
Benefits of Microsoft Phi-3
Microsoft Phi-3 distinguishes itself through its performance, affordability, ease of deployment, and adaptability to different hardware setups. It promises to be a valuable addition to the realm of AI language models, providing users with a versatile and efficient solution.
For more information on Microsoft Phi-3, you can visit here.
