华为盘古大模型5.0技术解密：引领AI与物理世界融合新纪元
近日，华为在人工智能领域再次取得重大突破，其盘古大模型5.0技术备受瞩目。这一全新升级的大模型体系实现了与物理世界的真实结合，展现了AI技术的全新面貌。
盘古大模型5.0分为多个版本，包括亿级参数的Pangu E端侧模型、百亿级参数的Pangu P、千亿级的Pangu U以及万亿级的Pangu S，实现了全系列、多模态、强思维的升级。该技术能够处理包括文本、图片、视频、雷达、红外、遥感等多种模态的信息，并具备强大的复杂推理能力。
在大会上，诺亚方舟实验室主任姚骏详细解读了盘古大模型5.0背后的技术原理。华为对盘古大模型进行了全面升级，提高了训练效率，并依托华为云AI算力平台实现了强大的性能。新模型具备更丰富的多模态和更强的思维能力，为工业领域和具身智能等技术探索带来了革命性的突破。
值得一提的是，随着鸿蒙HarmonyOS NEXT Beta版本的发布，小艺已升级成为智能体，为全场景设备提供语音对话、图文识别等功能。依托昇腾的算力和盘古大模型，HarmonyOS NEXT拥有了系统级AI能力。
目前，盘古大模型5.0已在高铁故障检测等工业领域落地应用，其强大的技术实力和应用前景引起了广泛关注。华为这次的技术突破，无疑为AI领域的发展开启了新的篇章，让人们更加期待AI技术与物理世界的深度融合。
英语如下：
News Title: Huawei Pangu Big Model 5.0 Unveiled: Multi-modal Fusion, Intelligence Connects with the Physical World
Keywords: Huawei Pangu Big Model upgrade, Multi-modal technology, AI combines with the physical world
News Content:
Huawei has made another significant breakthrough in the field of artificial intelligence, with its Pangu Big Model 5.0 technology attracting much attention. This newly upgraded big model system achieves a real combination with the physical world, showcasing a whole new look of AI technology.
The Pangu Big Model 5.0 is divided into multiple versions, including the Pangu E model with billions of parameters for edge devices, the Pangu P model with hundreds of billions of parameters, the Pangu U model with trillions of parameters, and the Pangu S model with trillions of parameters as well. This achieves a full series, multi-modal, and strong thinking upgrade. The technology is capable of processing multi-modal information including text, images, videos, radars, infrared, remote sensing, and has strong complex reasoning abilities.
At the conference, Yao Jun, director of the Noah’s Ark Lab, provided a detailed interpretation of the technical principles behind the Pangu Big Model 5.0. Huawei has comprehensively upgraded the Pangu Big Model, improving training efficiency and relying on the Huawei Cloud AI computing platform to achieve strong performance. The new model has richer multi-modality and stronger thinking abilities, bringing revolutionary breakthroughs in industrial fields and technological exploration in embodied intelligence.
It is worth mentioning that with the release of the beta version of HarmonyOS NEXT, Xiaoyi has been upgraded to an intelligent body, providing voice dialogues, graphic recognition and other functions for full-scene devices. Relying on Ascend’s computing power and the Pangu Big Model, HarmonyOS NEXT has system-level AI capabilities.
Currently, the Pangu Big Model 5.0 has been applied in industrial fields such as high-speed rail fault detection, and its powerful technical strength and application prospects have attracted widespread attention. Huawei’s technological breakthroughs have undoubtedly opened a new chapter in the development of AI, making people more expectant of the deep integration of AI technology and the physical world.
【来源】https://www.jiqizhixin.com/articles/2024-06-24-4
Views: 3