Alibaba Cloud, the digital technology and intelligence backbone of Alibaba Group, has released its latest large-scale language model, Tongyi Qianwen.
The new AI model will be integrated into Alibaba’s various businesses to improve user experience in the near future. The company’s customers and developers can access the models to create customized AI capabilities in a cost-effective manner.
Alibaba Cloud also announced low-cost options for major cloud products such as Elastic Compute Service (ECS) and Object Storage Service (OSS), offering new ECS instances, OSS Reserved Capacity (OSS-RC) and OSS Anywhere Reserved Capacity ( OSS-Ark).
The move will make computing more accessible and affordable for companies looking to tap new opportunities in China’s new AI era.
Daniel Zhang, Chairman and CEO of Alibaba Group and CEO of Alibaba Cloud Intelligence, said: game.
“Alibaba Cloud, as the world’s leading cloud computing service provider, is committed to making computing and AI services more accessible and inclusive for enterprises and developers to discover more insights and grow. Exploring new business models for us, enabling us to create more cutting-edge products and services for society.
Integrate Tongyi Qianwen into Alibaba’s business
In the near future, Tongyi Qianwen (Chinese) will be integrated into all business applications across Alibaba’s ecosystem to further enhance user experience, ranging from enterprise communications, intelligent voice assistance, e-commerce, search, navigation and entertainment. increase.
With Chinese and English language capabilities, the model will be deployed first on Alibaba’s digital collaboration workplace and application development platform DingTalk and IoT-enabled smart home appliances provider Tmall Genie.
Powered by Tongyi Qianwen, DingTalk is designed to make workplace communication more efficient. For example, you can summarize meeting minutes, transcribe meeting conversations, write emails, and create business proposals and promotional campaign plans with simple prompts. Users can instantly create mini-apps on DingTalk by simply taking a picture of their idea on paper.
Powered by Tongyi Qianwen, Tmall Genie can have more dynamic and vivid conversations with Chinese users. For example, you can create stories to tell your kids, provide healthy eating recipes, travel tips, or recommend his background music for workouts.
To help enterprises reap more of the benefits of AI-driven innovation, Alibaba Cloud will provide clients with access to Tongyi Qianwen on the cloud to help build customized large-scale language models, the company said. I’m here.
By fine-tuning Tongyi Qianwen with customer-specific intelligence and industry know-how in a secure cloud environment, enterprises can establish AI models tailored to their specific business needs. This is expected to create new growth momentum for customers and eliminate the need for resource-intensive and costly pre-training processes to build base models. Tongyi Qianwen is now available for commercial customers in China for beta testing.
In addition, developers will soon have access to Alibaba Cloud’s Tongyi Qianwen to create large-scale AI applications. This will further strengthen the AI software ecosystem across sectors, from logistics to media, finance, manufacturing, energy, retail, and more. Tongyi Qianwen as an API is also available for Chinese developers to sign up for beta testing.
Multimodal capabilities such as image understanding and text-to-image conversion will soon be added to the Tongyi Qianwen model, providing users with more engaging AI capabilities.
Jingren Zhou, CTO of Alibaba Cloud Intelligence, commented: In this modern AI era, we can create additional value for our customers and the wider community through our resilient public cloud infrastructure and proven AI capabilities.
“We are witnessing a new paradigm in AI development in which cloud and AI models play a key role. We want to increase business productivity, expand our expertise and capabilities, and unlock more exciting opportunities through innovation.”
Tongyi Qianwen is based on Alibaba’s own pre-trained model framework, Tongyi, and integrates various AI models, including models that can convert text into images and short videos. Last year, Alibaba Cloud launched its open source Model-as-a-Service (MaaS) with hundreds of AI models, including a Tongyi-based text-to-image model for global developers and researchers. We launched ModelScope, our platform. With over 1 million active users to date, ModelScope has 800 models available and he has over 16 million model downloads to date.
More accessible and affordable computing for businesses and students
Designed for small and medium-sized enterprises (SMEs), the new ECS Universal instance family offers the same stability as similar offerings. The company says it’s suitable for web applications and websites, enterprise office applications, and small businesses doing offline data analysis.
OSS-RC allows customers to reserve storage capacity in a specific region for one year. This reduces capacity costs by up to 50% from pay-as-you-go pricing. If customers do not need to store data in a specific region, they can create an OSS “Anywhere Bucket” to store data in a region selected by Alibaba Cloud. OSS-ARC can reserve capacity for objects stored in OSS Anywhere buckets. This reduces capacity costs by up to 70% from pay-as-you-go pricing.
To make computing resources more accessible to Chinese developers, Alibaba Cloud also announced up to 3-month free trials for core products such as ECS and POLARDB databases. Alibaba Cloud also provides his 1,000 free training courses on cloud technology and nearly 500 hands-on experiments based on real business scenarios to help developers easily access cloud technology.
Additionally, Alibaba announced that its Cloud for You program will partner with the UNESCO Presidency on Artificial Intelligence in Education to provide more computing resources to students in rural China. His Alibaba non-profit Cloud for Youth program, announced in 2021, will help narrow China’s digital gap by providing schools in underdeveloped regions with Cloud Her computers. The goal is to help young people strengthen the technical foundation of their digital capabilities and develop local digital talent.