


After its overseas Meta, Alibaba has become another technology giant that promotes the trend of artificial intelligence (AI) large model "Android moment"
According to reports from Beijing Business Daily, Alibaba Cloud will release the open source general question and answer model Qwen-7B and conversation model Qwen-7B-Chat on Thursday, August 3. Both models have 7 billion parameters. They have launched the first "Model as a Service" open platform in China, the Magic Community, and it can be used for free, and commercial use is also allowed
Users can quantify Qwen-7B and Qwen-7B-Chat through open source code, and deploy and run models on consumer-grade graphics cards. They can directly download the model from the Moda community, or access and call Qwen-7B and Qwen-7B-Chat through the Alibaba Cloud Lingji platform. Alibaba Cloud provides users with services including model training, inference, deployment and fine-tuning
On the Magic Tower community, there is a post dedicated to the installation method of the Tongyi Qianwen model, the best practices for creating space experience, model reasoning and model training, and also attaches screenshots of the model link and download situation
According to public information, Qwen-7B is a base model that is pre-trained using deduplicated and filtered data of more than 2.2 trillion tokens. It supports multiple languages such as Chinese and English, and has a context window length of 8k. The model contains high-quality Chinese, English, multi-language, code, mathematics and other data, covering the entire network text, encyclopedia, books, code, mathematics and vertical fields in various fields
According to the MMLU evaluation results, Qwen-7B performed well in English evaluation, surpassing other similar open source pre-training models and being competitive with larger-scale models. In terms of Chinese evaluation, Qwen-7B achieved the highest score on the C-Eval validation set and was competitive even with larger-scale models
The following is a comparison of the MMLU 5-shot accuracy results of Qwen-7B
Alibaba Cloud has built an AI assistant Qwen-7B-Chat based on the base model through the alignment mechanism. It is a large language model of Chinese and English dialogue based on Transformer, which has successfully achieved alignment with human cognition. The model uses a variety of pre-training data, including online texts, professional books, codes, etc., covering a wide range of areas
The zero-shot accuracy of the Qwen-7B-Chat model on both the C-Eval validation set and the MMLU evaluation set exceeds that of other similar alignment models
The following is a comparison of the zero-shot accuracy results on the C-Eval test set
Alibaba Cloud became the first large technology company in China to join the ranks of open source large models. In July this year, it jointly released with Meta a commercial version of the open source AI model Llama 2, which can replace OpenAI and Google's models. In addition, Zhipu AI and Tsinghua KEG Laboratory also announced China’s top open source large model
in JulyThe advantages of open source models are to increase user acceptance and provide more data for artificial intelligence processing. The larger the data volume of LLM, the more powerful its function. In addition, the open source model helps researchers and developers find and solve vulnerabilities, improving technology and security levels
At the Alibaba Cloud Summit in April 2023, Alibaba announced the opening of Tongyi Qianwen to enterprises, allowing enterprises to use Tongyi Qianwen’s capabilities to train their own large models
Zhou Jingren, Chief Technology Officer (CTO) of Alibaba Cloud Intelligence Group, said that in the future, enterprises can make full use of Alibaba Cloud's Tongyi Qianwen capabilities and combine their own industry knowledge and application scenarios to train customized enterprise large models. For example, each company can have its own intelligent customer service, intelligent shopping guide, intelligent voice assistant, copywriting assistant, AI designer and self-driving model and other functions
Zhang Yong, CEO of Alibaba Group and CEO of Alibaba Cloud Intelligence Group, said that all Alibaba products will be integrated with the Tongyi Qianwen large model
Alibaba Cloud hopes to help more companies use large models to adapt to the needs of the AI era, so that each company can have its own dedicated large model for its industry capabilities, and reconstruct it based on Tongyi Qianwen
The above is the detailed content of The 'Android moment' of domestic large AI models has arrived! Alibaba Cloud Tongyi Qianwen is free, open source, and available for commercial use. For more information, please follow other related articles on the PHP Chinese website!

今年以来,360集团创始人周鸿祎在所有公开场合的讲话都离不开一个话题,那就是人工智能大模型。他曾自称“GPT的布道者”,对ChatGPT取得的突破赞不绝口,更是坚定看好由此产生的AI技术迭代。作为一个擅于表达的明星企业家,周鸿祎的演讲往往妙语连珠,所以他的“布道”也创造过很多热点话题,确实为AI大模型添了一把火。但对周鸿祎而言,光做意见领袖还不够,外界更关心他执掌的360公司如何应对这波AI新浪潮。事实上,在360内部,周鸿祎也早已掀起一场全员变革,4月份,他发出内部信,要求360每一位员工、每

蚂蚁集团在上海举行的第二届外滩大会上,宣布正式发布旗下的金融大模型据蚂蚁集团介绍,蚂蚁金融大模型基于其自研基础大模型,并针对金融产业进行深度定制,底层算力集群达到万卡规模。目前,该大模型已在蚂蚁集团财富、保险平台全面测试。同时,基于该大模型的两款产品——智能金融助理“支小宝2.0”、服务金融产业专家的智能业务助手“支小助”,也已正式亮相。据介绍,两款大模型产品展示了蚂蚁从基础大模型到行业大模型以及产业应用的全栈布局和进展。本站附两款产品目前进度如下:“支小宝2.0”已经开始内测近半年,将在完成相

2023年8月16日,WAVESUMMIT深度学习开发者大会在中国举办,该活动由深度学习技术及应用国家工程研究中心主办,百度飞桨和文心大模型承办。在会上,百度发布了文心大模型、飞桨平台和AI原生应用如流等一系列技术、产品的最新进展和生态成果。百度集团副总裁兼首席信息官李莹发表了主题演讲,她认为当前以AI大模型为核心技术的第四次科技革命将从根本上推动生产力变革,为各行各业提供强大支持,并为企业办公领域带来前所未有的发展机遇基于AI原生思维,李莹宣布,百度智能工作知识管理理念“创新流水线=AIx知识

6月21日,北大光华管理学院联合腾讯,宣布升级“数字中国筑塔计划”,共同推出“企业管理者人工智能通识课”系列课程。在第一课上,腾讯集团高级执行副总裁、云与智慧产业事业群CEO汤道生回顾了AI发展的历史,表示算法创新、算力增强、开源共创三大因素的叠加,构成了AI的“增长飞轮”。大模型的快速进步,推动我们正在进入一个被AI重塑的时代。汤道生表示,大模型只是起点,未来,应用落地的产业变革是更大的图景。企业过去的研发、生产、销售、服务等环节中,有很多依赖人来判断、协调与沟通的地方,今天都值得去看看,哪些

近期,WakeData惟客数据(以下简称“WakeData”)完成了新一轮的产品能力升级。在2022年11月的产品发布会上,已传递出WakeData的“三个坚定”:始终坚定技术投入,全面夯实核心产品的科技能力和自研率;始终坚定国产化适配能力,支持国产芯片、操作系统、数据库、中间件、国密算法等,并在同领域实现对国外厂商的国产化替代;始终坚定拥抱生态,与伙伴共创共赢。WakeData继续新一轮的产品能力升级,凭借过去5年的技术积累,以及在地产、零售、汽车等行业和垂直领域的实践,与战略伙伴联合研发具

“无产业不AI,无应用不AI。”随着AI(人工智能)大模型技术落地,AI应用遍地开花。连日来,多家企业发布基于大模型的AI应用产品。身处“百模大战”时代,如何打造国产大模型应用产品?如何为大模型提供更普惠的算力、寻找更合适的场景?发布现场图。6月1日,阿里云对外披露通义大模型最新进展,上线聚焦音视频内容的AI新品“通义听悟”,成为国内首个开放公测的大模型应用产品。有专家认为,云计算是打造大模型最合适的形式,而大模型的进化过程,或将会对传统云计算架构开始新一轮的改造。阿里云AI新产品“通义听悟”开

近日,科大讯飞公告其构建的“讯飞星火认知大模型”将举行升级发布会,推出该人工智能大模型的V1.5(1.5版本)。此前,朗玛信息也因推出“朗玛•39AI全科医生”大模型产品举行发布会。此外,还有5家上市公司也在与投资者沟通交流中,披露已布局AI(人工智能)大模型的信息。来源:摄图网科大讯飞的“讯飞星火认知大模型”升级至1.5版近日,科大讯飞股份有限公司(证券简称:科大讯飞;证券代码:002230.SZ)披露了《关于讯飞星火认知大模型升级发布会的提示性公告》。公告显示,2023年5月6日,科大讯飞举


Hot AI Tools

Undresser.AI Undress
AI-powered app for creating realistic nude photos

AI Clothes Remover
Online AI tool for removing clothes from photos.

Undress AI Tool
Undress images for free

Clothoff.io
AI clothes remover

AI Hentai Generator
Generate AI Hentai for free.

Hot Article

Hot Tools

EditPlus Chinese cracked version
Small size, syntax highlighting, does not support code prompt function

Dreamweaver CS6
Visual web development tools

WebStorm Mac version
Useful JavaScript development tools

SublimeText3 Mac version
God-level code editing software (SublimeText3)

DVWA
Damn Vulnerable Web App (DVWA) is a PHP/MySQL web application that is very vulnerable. Its main goals are to be an aid for security professionals to test their skills and tools in a legal environment, to help web developers better understand the process of securing web applications, and to help teachers/students teach/learn in a classroom environment Web application security. The goal of DVWA is to practice some of the most common web vulnerabilities through a simple and straightforward interface, with varying degrees of difficulty. Please note that this software
