What Is MaaS?
In today's fast-paced AI landscape, building and optimizing large models from scratch is often too expensive and technically complex. Reducing these barriers is now a critical challenge.
MaaS solves this by providing a suite of mainstream large models optimized for Ascend. Serving as an agile hub for enterprises and developers, MaaS eliminates the need for heavy upfront investment in model training, deployment, and O&M. By providing flexible, cost-effective API calls, MaaS delivers adaptable and affordable solutions for tasks like text generation, intelligent interactions, and data analysis. You can call models as needed to reduce costs and save time. This allows you to easily incorporate advanced AI features into your products and workflows, speeding up innovation and boosting competitive advantage.
Product Architecture
- MaaS (scenario-specific application layer) This layer focuses on the practical implementation of industry-specific models and is divided into three key modules:
- Enterprise application
Provides vertical AI solutions across various sectors, including government, manufacturing, finance, and healthcare.
- Model service (core module)
Covers the full lifecycle management of models:
- Model management: The MaaS Model Square offers a variety of mainstream open-source models for easy exploration and use.
- Model deployment: You can deploy built-in services provided by MaaS as dedicated services to accelerate inference and monitor model service call data.
- Enterprise application
Access Methods
MaaS allows you to call model applications via the management console or APIs.
- Console
You can access MaaS directly through the web-based console. It provides a comprehensive suite of features, including model experience, inference, training, and management, enabling end-to-end AI development. For details, see Model Calls.
- API
Model services and applications created on the MaaS console can be called via APIs in other business environments. For details, see MaaS API Call Specifications.
Getting Started with MaaS
- First large model API call: Calling Large Model APIs for the First Time
- Model API calls: MaaS API Call Specifications
You can go to the MaaS progressive knowledge to learn about the basic concepts and usage of MaaS.
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.
For any further questions, feel free to contact us through the chatbot.
Chatbot