KUALA LUMPUR: Alibaba Cloud launched an open-source Model-as-a-Service (MaaS) platform that comes with hundreds of AI models, including large pre-trained models for global developers and researchers called ModelScope during its annual Apsara conference.
Alibaba Cloud Intelligence's president, Jeff Zhang stated that cloud computing has sped up the integration of the cloud and endpoint terminals and caused a fundamental revolution in the way computing resources are organised, produced, and used commercially.
"As more customers are speeding up with their cloud adoption, we have been upgrading our cloud-based resources, services and tools to become serverless, more intelligent and digitalised in order to lower the barrier for companies to adopt new technology and capture more opportunities in the cloud era," he said.
Over 300 ready-to-deploy AI models have been created by Alibaba Damo Academy, the company's global research initiative, over the past five years and are available on the ModelScope platform.
ModelScope is an open-source community that aims to simplify and reduce the cost of creating and maintaining AI models.
To date, dozens of models from independent developers have been added to the open-source platform, covering a range of topics from computer vision to natural language processing (NLP) and audio. The platform currently houses more than 150 AI State-of-the-Art (SotA) models that are regarded as the best in their fields on a global scale.
Also made available on the platform are Alibaba's proprietary large pre-trained models such as Tongyi, which is capable of turning text into image with five billion parametres, and one-for-all, a six billion-parametre pre-trained model that excels at cross-modal tasks such as image captioning and visual question answering.
The platform also offers large pre-trained models developed exclusively by Alibaba, including Tongyi, which can convert text into images with five billion parametres, and OFA (once-for-all) network, a model with six billion parametres that excels at cross-modal tasks like image captioning and visual question answering.
On the new ModelScope platform, developers and researchers can easily run free tests on the models online and receive the results in a matter of minutes. The models can be deployed on other cloud platforms, used locally, or run online with support from Alibaba Cloud. Custom AI applications can also be made using them.
During the flagship conference, the cloud provider also announced upgrades to its integrated data analytics and intelligent computing platform for business, as well as ranges of serverless database products. Among the updated products include Alibaba's cloud-native data warehouse AnalyticDB (ADB), cloud-native database PolarDB, and ApsaraDB for Relational Database Service (RDS), which put a strong emphasis on giving users access to computing power on demand, and staying ahead of the trend toward serverless software development.
The new upgrades will allow more customers to benefit from automatic scaling with extreme elasticity based on actual workloads and a pay-as-you-go billing model by utilising the serverless technologies offered by the cloud provider.
Currently, Alibaba Cloud has more than 20 serverless key products in total and is adding more product categories to become serverless.
In order to promote workload collaboration between cloud and local hardware, Alibaba also unveiled the new Wuying Cloudbook powered by Wuying Architecture at the same flagship event. The Cloudbook's specialised architecture aims to make it simpler for users to gain faster, more secure access to limitless computing power in the cloud, encouraging workplace flexibility and collaboration.