Microsoft Introduces Phi-3, A Family Of Open Small Language Models (SLMs)

As artificial intelligence (AI) continues to disrupt industries, processes and workflows, AI companies remain on the lookout to offer efficient, capable and cost-effective products, services and models.

In an attempt to retain a strong position in the AI market, Microsoft has come up with an announcement that enhances its offerings for businesses building AI applications with specific requirements.

So, what did the company reveal? Let’s explore!
 

What Did Microsoft Announce?

 

• Through a blog post published on its website, Microsoft announced the release of Phi-3, a family of open AI models, a set of Small Language Models (SLMs).

• SLMs are designed to carry out simpler tasks, are more accessibility, easier to use and fine-tune for specific needs.

• According to the release, “Phi-3 models are the most capable and cost-effective small language models (SLMs) available, outperforming models of the same size and next size up across a variety of language, reasoning, coding, and math benchmark.”

• Phi-3 builds on previous Phi models, is trained using high-quality data and offers customers high-quality models with more practical choices to compose and build GenAI applications.

• SLMs like Phi-3 are effective for “resource constrained environments including on-device and offline inference scenarios, latency bound scenarios where fast response times are critical and cost-constrained use cases, particularly those with simpler tasks.”

• Phi-3-mini, a 3.8B language model available in two context-length variants—4K and 128K tokens, will be available on Microsoft Azure AI Studio, Hugging Face and Ollama.

• It is also available as an NVIDIA NIM microservice and optimized for NVIDIA GPUs.

• The models were built adhering to the Microsoft Responsible AI Standard, which follows six principles; accountability, transparency, fairness, reliability and safety, privacy and security and inclusiveness, after which they underwent rigorous evaluation.

• The plan is to release more models of the Phi-3 family soon, which will offer more choice across quality and cost.

• Microsoft also released a technical paper with further details.

• Through the blog post, Misha Bilenko, Corporate VP, Microsoft GenAI, said, “Thanks to their smaller size, Phi-3 models can be used in compute-limited inference environments. Phi-3-mini, in particular, can be used on-device, especially when further optimized with ONNX Runtime for cross-platform availability.”

• [Contd.] “The smaller size of Phi-3 models also makes fine-tuning or customization easier and more affordable. In addition, their lower computational needs make them a lower cost option with much better latency.”

 

What Did Microsoft Executives Say?

 

• Sonali Yadav, Principal Product Manager for Generative AI at Microsoft, said, “What we’re going to start to see is not a shift from large to small, but a shift from a singular category of models to a portfolio of models where customers get the ability to make a decision on what is the best model for their scenario.”

• “Some customers may only need small models, some will need big models and many are going to want to combine both in a variety of ways,” said, Luis Vargas, Vice President of AI at Microsoft, adding, “If you are in a part of the world that doesn’t have a good network, you are still going to be able to have AI experiences on your device.”

• Talking about the emergence of SLMs, Ece Kamar, the Managing Director of AI Frontiers, said, “The claim here is not that SLMs are going to substitute or replace large language models. [Instead, SLMs] are uniquely positioned for computation on the edge, computation on the device, computations where you don’t need to go to the cloud to get things done. That’s why it is important for us to understand the strengths and weaknesses of this model portfolio.”

Do you think Microsoft’s move puts it in a better position to capture more market share? Do you think the company’s rivals should make similar moves?

Let us know in the comments below!