谷歌近日在官方博客上发布重要更新，宣布开源其最新研发的Magika项目，这是一款利用人工智能技术实现毫秒级内容类型的识别系统。Magika旨在为用户提供高效、精准的文件格式和内容分类服务，其卓越的性能在大规模测试中得到了验证，百万文件的识别准确率超过99%。
Magika的独特之处在于其定制的、高度优化的深度学习模型，能够在CPU上运行，无需高性能的GPU支持，这意味着它能在各种设备上快速运行，且不影响识别的准确性。在实际应用中，无论是办公文档、图像文件还是音频视频，Magika都能在短短几毫秒内完成识别，极大地提升了工作效率。
谷歌已将Magika的源代码公开，托管在知名的开源平台GitHub上，鼓励开发者、研究者和广大用户进行探索和应用。这一举措不仅展现了谷歌在人工智能领域的技术领先，也为全球开发者提供了一个创新和学习的新平台，有望推动相关领域的技术进步和应用创新。
通过开源Magika，谷歌希望与全球社区共享其在人工智能领域的研究成果，共同推动文件处理技术的发展，为数字化时代的文件管理和信息处理提供更强大的工具。这一举动也进一步体现了谷歌对于开放源代码和协作开发的承诺，期待更多的开发者能够参与进来，共同打造更加智能、便捷的数字世界。
英语如下：
**News Title:** “Google Open Sources Magika: AI-Powered Content Recognition in Milliseconds with 99% Accuracy!”
**Keywords:** Google Open Source, Magika, Millisecond Recognition
**News Content:**
Google recently announced a significant update on its official blog, revealing that it has open-sourced its latest development, the Magika project. Magika is an artificial intelligence-driven system for millisecond-level content type recognition. Designed to provide users with efficient and precise file format and content classification, its exceptional performance has been validated through extensive testing with over 99% accuracy on millions of files.
The uniqueness of Magika lies in its custom, highly optimized deep learning model, which can run on CPUs without requiring high-performance GPU support. This enables it to operate swiftly across various devices without compromising accuracy. In practical applications, Magika can identify office documents, image files, and audio-video files within mere milliseconds, significantly boosting productivity.
Google has made Magika’s source code available on the popular open-source platform GitHub, inviting developers, researchers, and users worldwide to explore and utilize it. This move not only underscores Google’s technological leadership in artificial intelligence but also offers a new platform for global developers to innovate and learn, potentially advancing technological progress and application innovation in the field.
By open-sourcing Magika, Google aims to share its AI research findings with the global community, jointly advancing file processing technology and providing more powerful tools for document management and information processing in the digital age. This action further demonstrates Google’s commitment to open-source and collaborative development, inviting more developers to contribute and shape a smarter, more convenient digital world together.
【来源】https://www.ithome.com/0/750/474.htm
Views: 1