近日，全球领先的科技公司微软宣布推出全新多模态小模型Phi-3-vision，该模型拥有42亿参数，将进一步扩大微软在人工智能领域的领先优势。
微软此次发布的Phi-3系列模型，包含三种规模的模型：Phi-3-mini（38亿参数）、Phi-3-small（70亿参数）和Phi-3-medium（140亿参数）。其中，Phi-3-mini已被纳入Azure AI平台，为开发者提供强大的AI能力。
特别值得一提的是，微软还推出了Phi-3-vision这一多模态小模型变种。该模型拥有42亿参数，能够支持通用视觉推理任务以及图表、图形和表格的推理。这意味着，开发者可以利用Phi-3-vision模型处理更加复杂的视觉信息，从而拓展AI在各个领域的应用。
此次微软发布的Phi-3系列模型，是在首次宣布发布其Phi-3轻量级AI模型系列一个月后，正式向开发者开放的。这一举措表明微软在人工智能领域的研发实力不断壮大，同时也显示出微软致力于推动AI技术发展的决心。
腾讯科技报道称，微软此次发布的Phi-3系列模型，将在很大程度上提升开发者在图像识别、自然语言处理等领域的创新能力。作为一款多模态小模型，Phi-3-vision在保持高性能的同时，降低了模型的复杂度和计算资源需求，为开发者提供了更为灵活的AI应用解决方案。
微软此次发布的Phi-3系列模型，无疑将为AI领域的发展带来新的机遇。随着参数规模的不断提升，Phi-3系列模型在处理复杂任务时将具备更高的准确性和效率。而多模态小模型Phi-3-vision的推出，更是为开发者提供了更多创新的思路和可能。
在未来，微软将继续深化在人工智能领域的研发，推出更多具有突破性的AI技术，助力全球开发者创造更智能、更便捷的应用。让我们共同期待微软在人工智能领域带来更多精彩的成果！
英语如下：
### Microsoft Releases Phi-3-vision Multimodal Small Model: Significant Improvement in Visual Reasoning Ability
**Keywords:** Microsoft Release, Phi-3 Series, Multimodal Small Model
**News Content:**
# Microsoft Unveils Multimodal Small Model Phi-3-vision, Further Expanding AI Capabilities
Recently, Microsoft, a global leader in technology, announced the launch of a new multimodal small model, Phi-3-vision, featuring 4.2 billion parameters, which is set to further enhance Microsoft’s lead in the field of artificial intelligence.
The Phi-3 series models released by Microsoft include three sizes of models: Phi-3-mini (3.8 billion parameters), Phi-3-small (7 billion parameters), and Phi-3-medium (14 billion parameters). Among them, Phi-3-mini has been integrated into the Azure AI platform, offering powerful AI capabilities to developers.
Particularly noteworthy is the introduction of Phi-3-vision, a variant of the multimodal small model with 4.2 billion parameters, supporting general visual reasoning tasks as well as inference for charts, graphics, and tables. This means developers can utilize the Phi-3-vision model to handle more complex visual information, expanding the applications of AI across various fields.
One month after initially announcing the release of its Phi-3 lightweight AI model series, Microsoft has officially opened up the Phi-3 series models to developers. This move demonstrates Microsoft’s continuous strengthening in research and development in the field of artificial intelligence and its commitment to promoting the development of AI technology.
According to a report by Tencent Technology, Microsoft’s newly released Phi-3 series models will significantly enhance the innovation capabilities of developers in fields such as image recognition and natural language processing. As a multimodal small model, Phi-3-vision maintains high performance while reducing model complexity and computational resource requirements, providing developers with more flexible AI application solutions.
Microsoft’s release of the Phi-3 series models is undoubtedly bringing new opportunities to the development of the AI field. With the continuous improvement of parameter scale, the Phi-3 series models will have higher accuracy and efficiency in handling complex tasks. Moreover, the introduction of the multimodal small model Phi-3-vision offers developers more innovative ideas and possibilities.
In the future, Microsoft will continue to deepen its research and development in the field of artificial intelligence, launching more breakthrough AI technologies to help global developers create smarter and more convenient applications. Let’s look forward to more exciting achievements from Microsoft in the field of artificial intelligence!
【来源】https://new.qq.com/rain/a/20240522A0065F00
Views: 1