search
HomeTechnology peripheralsAIAnother large AI model is released, and Qishang Online builds a solid base of intelligent computing power for it
Another large AI model is released, and Qishang Online builds a solid base of intelligent computing power for itJun 09, 2023 pm 09:33 PM
ai modelIntelligent computing powerEnterprise Business Online

On June 3, Zhongke Wenge, a leading domestic data and decision-making intelligent service provider under the Chinese Academy of Sciences, released the Yayi large model and launched large model applications in media, finance, publicity and other fields. As a strategic partner of Zhongke Wenge's computing power service, Qishang Online provides reliable underlying support for the Yayi large model in terms of intelligent computing power.

Another large AI model is released, and Qishang Online builds a solid base of intelligent computing power for it

AI computing power has entered the era of large models. Behind the rapid development of domestic large models, huge intelligent computing power support is needed to provide large-scale data processing, complex model training and reasoning task guarantees for AI applications. As a leading digital computing integration service provider in China, Qishang Online has been deeply involved in the data industry for more than 20 years. As early as 2020, it established the strategic direction of AI intelligent computing power services and embarked on the construction and operation of intelligent computing power centers.

In 2021, Qishang Online and Zhongke Wenge signed a strategic cooperation agreement. The two parties will work together in depth in cloud computing, big data, artificial intelligence and other fields to jointly develop new business models. Zhongke Wenge's "digital intelligence and supercomputing" The "Cloud Computing Center" was officially launched in the Sanlitun Capital Financial Computing Power Center of Qishang Online.

Another large AI model is released, and Qishang Online builds a solid base of intelligent computing power for it

Qishang Online provides construction, operation, maintenance and other services for the Zhongke Wenge Digital Supercomputing Cloud Platform. It builds a big data resource pool according to the data middle platform model, establishes data standards, integrates cross-modal data resources, and realizes data Hierarchical classification storage and authorized use improve the level of data utilization, thereby achieving "one pool integration" of data assets.

The Yayi large model released this time is a phased result of Zhongke Wenge’s years of exploration and technical research based on the direction of new generation artificial intelligence technology. It is also an important wisdom derived from the “Digital Intelligence Supercomputing Cloud Computing Center” Fruitful. Qishang Online's cloud computing resources and hosting services accelerate the pace of product development and increase the speed of model training.

The Yayi large model is a safe and reliable enterprise-level exclusive large model. It has five core capabilities, including real-time network Q&A, domain knowledge Q&A, multi-language content understanding, complex scene information extraction, and multi-modal content generation, with a total of more than 100 A unique skill that can quickly connect to government and enterprise data and generate large model exclusive application services with one click.

Yayi large model supports three usage methods: cloud use, local all-in-one machine deployment, and independent private training deployment. It can cover scenario businesses in finance, media, governance, security and other directions, and can be generalized to home furnishing, medical care, education and other industries. .

Another large AI model is released, and Qishang Online builds a solid base of intelligent computing power for it

Zhongke Wenge hopes to use Yayi large model to provide comprehensive large model solutions for various enterprises and industries. As a computing power infrastructure service provider, Qishang Online will continue to conduct in-depth and extensive industrial collaboration and integration with upstream and downstream cooperative enterprises to build an AI public computing power service platform to provide computing power guarantee for the implementation of more AI scenarios, allowing artificial intelligence to The development of all walks of life brings greater dividends!

The above is the detailed content of Another large AI model is released, and Qishang Online builds a solid base of intelligent computing power for it. For more information, please follow other related articles on the PHP Chinese website!

Statement
This article is reproduced at:搜狐. If there is any infringement, please contact admin@php.cn delete
AI大模型浪潮下算力需求爆增,商汤“大模型+大算力”赋能多产业发展AI大模型浪潮下算力需求爆增,商汤“大模型+大算力”赋能多产业发展Jun 09, 2023 pm 07:35 PM

近日,以“AI引领时代,算力驱动未来”为主题的“临港新片区智算大会”举行。会上,新片区智算产业联盟正式成立,商汤科技作为算力提供企业成为联盟一员,同时商汤科技被授予“新片区智算产业链链主”企业。作为临港算力生态的积极参与者,商汤目前已建设了亚洲目前最大的智能计算平台之一——商汤AIDC,可以输出5000Petaflops的总算力,可支持20个千亿参数量的超大模型同时训练。以AIDC为底座、前瞻打造的商汤大装置SenseCore,致力于打造高效率、低成本、规模化的下一代AI基础设施与服务,赋能人工

研究者:AI模型推理环节耗电更多,2027年行业用电将堪比荷兰研究者:AI模型推理环节耗电更多,2027年行业用电将堪比荷兰Oct 14, 2023 am 08:25 AM

IT之家10月13日消息,《Cell》的姐妹期刊《Joule》本周出版了一篇名为《持续成长的人工智慧能源足迹(Thegrowingenergyfootprintofartificialintelligence)》论文。通过查询,我们了解到这篇论文是由科研机构Digiconomist的创始人AlexDeVries发表的。他声称未来人工智能的推理性能可能会消耗大量的电力,预计到2027年,人工智能的用电量可能会相当于荷兰一年的电力消耗量AlexDeVries表示,外界一向认为训练一个AI模型“AI最

一言不合就跑分,国内AI大模型为何沉迷于“刷榜”一言不合就跑分,国内AI大模型为何沉迷于“刷榜”Dec 02, 2023 am 08:53 AM

“不服跑个分”这句话,我相信关注手机圈的朋友一定不会感到陌生。例如,安兔兔、GeekBench等理论性能测试软件因为能够在一定程度上反映手机的性能,因此备受玩家的关注。同样地,在PC处理器、显卡上也有相应的跑分软件来衡量它们的性能既然"万物皆可跑分",目前最火爆的AI大模型也开始参与跑分比拼,尤其是在"百模大战"开始后,几乎每天都有突破,各家都自称为"跑分第一"国产AI大模型在性能评分方面几乎从未落后,但在用户体验方面却始终无法超越GP

中科院物理所联合院网络中心发布AI模型MatChat中科院物理所联合院网络中心发布AI模型MatChatNov 03, 2023 pm 08:13 PM

IT之家11月3日消息,中科院物理所官网发文,近期,中国科学院物理研究所/北京凝聚态物理国家研究中心SF10组和中国科学院计算机网络信息中心共同合作,将AI大模型应用于材料科学领域,将数万个化学合成路径数据投喂给大语言模型LLAMA2-7b,从而获得了MatChat模型,可用来预测无机材料的合成路径。IT之家注意到,该模型可根据所询问的结构进行逻辑推理,并输出相应的制备工艺和配方。目前已部署上线,并向所有材料科研人员开放使用,为材料研究和创新带来了新启发和新思路。该工作为大语言模型在细分科学领域

四倍提速,字节跳动开源高性能训练推理引擎LightSeq技术揭秘四倍提速,字节跳动开源高性能训练推理引擎LightSeq技术揭秘May 02, 2023 pm 05:52 PM

Transformer模型出自于Google团队2017年发表的论文《Attentionisallyouneed》,该论文中首次提出了使用Attention替换Seq2Seq模型循环结构的概念,给NLP领域带来了极大冲击。且随着近年来研究的不断推进,Transformer相关技术逐渐由自然语言处理流向其他领域。截止目前,Transformer系列模型已经成为了NLP、CV、ASR等领域的主流模型。因此,如何更快地训练和推理Transformer模型已成为业界的一个重要研究方向。低精度量化技术能够

你必须了解的关于谷歌最新AI模型Gemini的全部信息你必须了解的关于谷歌最新AI模型Gemini的全部信息Dec 14, 2023 pm 08:57 PM

什么是GoogleGemini?Gemini是谷歌推出的一种最新的、功能强大的AI模型,它不仅可以理解文本,还可以理解图像、视频和音频。作为一种多模式模型,Gemini被描述为能够完成数学、物理和其他领域的复杂任务,以及理解和生成各种编程语言的高质量代码目前,它可以与GoogleBard和GooglePixel8进行整合,并且将逐渐整合到其他Google服务中谷歌DeepMind的CEO兼联合创始人丹尼斯·哈萨比斯说:“Gemini是谷歌整个团队的大规模合作成果,包括我们在谷歌研究中心的同事。它

中国联通发布图文AI大模型,可实现以文生图、视频剪辑中国联通发布图文AI大模型,可实现以文生图、视频剪辑Jun 29, 2023 am 09:26 AM

驱动中国2023年6月28日消息,今日在上海世界移动通信大会期间,中国联通发布图文大模型“鸿湖图文大模型1.0”。中国联通称,鸿湖图文大模型是首个面向运营商增值业务的大模型。第一财经记者了解到,鸿湖图文大模型目前拥有8亿训练参数和20亿训练参数两个版本,可以实现以文生图、视频剪辑、以图生图等功能。此外,中国联通董事长刘烈宏在今天的主题演讲中也表示,生成式AI正在迎来发展的奇点,未来2年内50%的工作将受到人工智能深刻影响。

Meta研究人员做出AI新尝试:教机器人无需地图或训练实现物理导航Meta研究人员做出AI新尝试:教机器人无需地图或训练实现物理导航Apr 09, 2023 pm 08:31 PM

Meta Platforms公司人工智能部门日前表示,他们正在教AI模型如何在少量训练数据支持下学会在物理世界中行走,目前已经取得了快速进展。这项研究能够显著缩短AI模型获得视觉导航能力的时间。以前,实现这类目标要需要利用大量数据集配合重复“强化学习”才能实现。Meta AI研究人员表示,这项关于AI视觉导航的探索将给虚拟世界带来重大影响。而项目的基本思路并不复杂:帮助AI像人类那样,单纯通过观察和探索实现在物理空间导航。Meta AI部门解释道,“比如,如果要让AR眼镜指引我们找到钥匙,就必须

See all articles

Hot AI Tools

Undresser.AI Undress

Undresser.AI Undress

AI-powered app for creating realistic nude photos

AI Clothes Remover

AI Clothes Remover

Online AI tool for removing clothes from photos.

Undress AI Tool

Undress AI Tool

Undress images for free

Clothoff.io

Clothoff.io

AI clothes remover

AI Hentai Generator

AI Hentai Generator

Generate AI Hentai for free.

Hot Article

R.E.P.O. Energy Crystals Explained and What They Do (Yellow Crystal)
2 weeks agoBy尊渡假赌尊渡假赌尊渡假赌
Repo: How To Revive Teammates
1 months agoBy尊渡假赌尊渡假赌尊渡假赌
Hello Kitty Island Adventure: How To Get Giant Seeds
4 weeks agoBy尊渡假赌尊渡假赌尊渡假赌

Hot Tools

SublimeText3 Chinese version

SublimeText3 Chinese version

Chinese version, very easy to use

mPDF

mPDF

mPDF is a PHP library that can generate PDF files from UTF-8 encoded HTML. The original author, Ian Back, wrote mPDF to output PDF files "on the fly" from his website and handle different languages. It is slower than original scripts like HTML2FPDF and produces larger files when using Unicode fonts, but supports CSS styles etc. and has a lot of enhancements. Supports almost all languages, including RTL (Arabic and Hebrew) and CJK (Chinese, Japanese and Korean). Supports nested block-level elements (such as P, DIV),

SublimeText3 Linux new version

SublimeText3 Linux new version

SublimeText3 Linux latest version

MantisBT

MantisBT

Mantis is an easy-to-deploy web-based defect tracking tool designed to aid in product defect tracking. It requires PHP, MySQL and a web server. Check out our demo and hosting services.

SAP NetWeaver Server Adapter for Eclipse

SAP NetWeaver Server Adapter for Eclipse

Integrate Eclipse with SAP NetWeaver application server.