【谷歌开源Magika：引领文件识别技术进入毫秒级时代】
谷歌近日在官方博客上发布重大消息，宣布开源其最新研发的文件识别工具——Magika。这款基于人工智能的创新技术，能够以惊人的速度和精准度识别各种文件格式和内容类型，其在百万文件的测试中，准确率超过了99%。
Magika的诞生，标志着文件识别技术迈入了新的毫秒级时代。这款工具采用了定制化且高度优化的深度学习模型，即使在常见的中央处理器（CPU）上运行，也能在短短几毫秒内完成对文件类型的识别，极大地提升了工作效率。
谷歌开源Magika的决定，将为全球开发者和科技企业带来巨大的便利。他们现在可以访问并利用这个强大的工具，通过GitHub平台获取源代码，进行二次开发和个性化应用，推动相关领域技术的进一步创新。
据IT之家报道，Magika的推出不仅体现了谷歌在人工智能领域的技术实力，也彰显了其开放共享的科技理念。这一开源项目有望引领文件管理、数据分析和信息安全等多个行业的发展，为全球科技生态注入新的活力。
英语如下：
**News Title:** “Google Open-Sources Magika: AI-Powered, Identifying Millions of File Types in Milliseconds!”
**Keywords:** Google Open-Source, Magika, Millisecond Recognition
**News Content:**
**Google Open-Sources Magika: Pioneering File Recognition into the Millisecond Era**
Google recently made a significant announcement on its official blog, revealing that it is open-sourcing its cutting-edge file recognition tool, Magika. This AI-driven innovation can swiftly and accurately identify various file formats and content types with an astonishing accuracy rate of over 99% during tests involving millions of files.
The birth of Magika signals the dawn of a new millisecond age in file recognition technology. The tool employs a customized and highly optimized deep learning model, enabling it to identify file types within just milliseconds, even on standard central processing units (CPUs), thus dramatically enhancing workflow efficiency.
By open-sourcing Magika, Google is offering a significant boon to global developers and tech companies. They can now access and leverage this powerful tool, obtaining the source code via GitHub for further development and customization, fostering innovation in the respective fields.
According to IT Home, the launch of Magika not only showcases Google’s prowess in artificial intelligence but also underscores its commitment to open-source and collaborative technology. This open-source project is poised to drive advancements in file management, data analysis, and information security, injecting new vitality into the global tech ecosystem.
【来源】https://www.ithome.com/0/750/474.htm
Views: 1