Microsoft has launched Phi-3-vision, a new member of its small language model family, Phi-3. This new model can understand both text and images. The model can look at images and analyze them. Another big highlight is its target market. It’s designed for devices with lower processing power. It has 4.2 billion parameters making it the perfect model for mobile devices and laptops. Let’s explore the new Microsoft Phi-3-Vision below.
Microsoft has a wide range of AI models. The Phi-3 family consists of multiple models:
- Phi-3-mini: 3.8 billion parameters
- Phi-3-small: 7 billion parameters
- Phi-3-medium: 14 billion parameters
These models are primarily focused on lightweight AI solutions. These are essential for systems with limited processing power. Moreover, the Phi-3-vision makes it even better with a multimodal model. It is designed for the analysis of objects in an image. Let’s check out the use case of Phi-3-vision below.
Use of the new model
Phi-3-vision stands out by recognizing objects in images. This is very useful for looking at and answering questions about visual data. Users can analyze data from charts or photos. However, it is not like DALL-E which generates images. Phi-3-vision is more of a help for information regarding objects in the images. It is built with datasets including synthetic data and filtered publicly available websites. Moreover, it has a target market. With its 4.2 billion parameters, the Phi-3-vision is a great choice for mobile devices and laptops.
Gizchina News of the week
How does Phi-3-vision work?
Microsoft has shown how the new model works. Check out the above image, where the user asks to make an analysis based on a chart. Impressively, the model can generate all the necessary information from the chart. It provides detailed information rather than just a summary.
Industry Trend
This development is a goal toward a broader industry trend towards smaller AI models. After all, laptops and mobile devices are a big market. If companies can come up with smaller yet effective AI models, it will be a great advancement. Talking about availability, the Microsoft Phi-3-vision is currently available in preview. However, the other models of the Phi-3 family can be accessed through Azure’s model library.
In conclusion, the Phi-3-Vision seems to be a great step in the right direction. Tech giants are competing to lead the AI industry. New models like Phi-3-Vision put more focus on devices with lower hardware capabilities. For more information, check out Microsoft’s official website. That said, let us know your opinions on the new model from Microsoft in the comments below.
