Alibaba Cloud Model Studio

A one-stop service platform for whole-process generative AI engineering and application development, based on Tongyi Qianwen (Qwen) and other popular models

Why Alibaba Cloud Model Studio?

Alibaba Cloud Model Studio is a one-stop generative AI development platform based on an extensive collection of powerful foundation models (FMs) including Alibaba Cloud's Qwen series. It provides a complete set of model training tools and full-link application development kits, with a rich collection of plug-ins and convenient integration methods. This allows you to efficiently incorporate enterprise knowledge to customize FMs and build applications that understand you and your customers.

  • Whole-Process FM Engineering in One Stop

    Perform full-link model development, train and fine-tune models with automatic acceleration from the underlying infrastructure, evaluate models with comprehensive measures, and deploy them as services with one click

  • Capability-Enhanced and
    Industry-Specific FMs

    Power your applications with enhanced AI model capabilities such as Q&A, writing, NL2SQL, etc., and industry-savvy knowledge in legal, medical, and financial fields, and more from third-party FMs

  • Simplified GenAI Application Development

    Accelerate generative AI application development with visualized processes and highly customizable orchestration, template-based prompt engineering with out-of-the-box features, and a rich set of APIs for easy integration

  • Security for Responsible AI

    Store and transmit your corporate data in a secure cloud environment, apply customizable content governance to prompts and content, and combine responsible AI principles with tools for human accountability

How It Works

  • APIs and SDKs

    Integrate FM capabilities into your corporate systems and applications with purpose-designed APIs, which are encapsulated in SDKs that support mainstream programming languages.

  • Model Agent

    Build AI agents in simple steps with an extensive set of plugins, visualized application orchestration services, as well as templates and optimization tools for prompt engineering.

  • Model Training

    Create, label, and manage training datasets, customize FM training with adjustable parameters, evaluate FMs with high efficiency, and deploy FMs as services with one click.

  • Model Security

    Safeguard your corporate data in storage and transmission, automatically monitor and identify risky content, and filter or block undesirable information based on responsible AI principles.

  • ModelHub

    Choose from a variety of open-source FMs, including Alibaba Cloud's Qwen models, capability-enhanced or industry-specific FMs, and popular third-party FMs.

One-Stop FM Engineering

Capability-Enhanced FMs

Built-In FMs with Enhanced Capabilities

This platform provides FMs with enhanced capabilities in Q&A, text understanding and generalization, NL2SQL, writing, etc., to help you save time and investment in FM training and datasets.

Industry-Specific FMs

This platform provides FMs enriched with knowledge in financial, medical, legal, and other vertical industries, to support industry-specific application development and reduce time-to-market.

Third-Party FMs

You can directly deploy and customize third-party open-source FMs such as Llama 2, Baichuan 2, ChatGLM, etc.

Full-Link FM Training and Evaluation

Full-Link FM Training

The comprehensive model training toolset supports Supervised Fine-Tuning (SFT), Low-Rank Adaptation (LoRA), and other optimization methods. After customization, you can deploy your FM in one click.

Inference Acceleration

This platform automatically compresses FMs and accelerates inference tasks to improve efficiency and reduce costs for AI engineering.

Multi-Dimensional FM Evaluation

You can evaluate a single model or compare different models based on built-in evaluation datasets, and view evaluation results using visualized templates.

Convenient Development Services

Out-of-the-Box Applications

Ready-to-use applications for data retrieval and prompt templates reduce required configuration resources, lowering barriers to entry for generative AI development.

Easy-to-Use APIs

You can connect FMs to your business systems through APIs to start content generation and Q&A services based on your enterprise knowledge, after simple configurations. These APIs are encapsulated in SDKs that support programming languages including Python, Java, Go, etc.

Flexible FM Orchestration

Visualized FM orchestration management helps you streamline FM utilization and create clear and logical work processes.

Getting Started in Easy Steps

1
Creating an Application with Pre-Built Templates
  • 1. Select a scenario-based template.
  • 2. Create an application with one click.
  • 3. Test the application.
  • 4. Run the application.
2
Creating an FM-Based Application
  • 1. Select an FM.
  • 2. Create an application with one click.
  • 3. Test the application.
  • 4. Run the application.
3
Creating an Application Based on a Customized FM
  • 1. Upload training data.
  • 2. Select and train your FM.
  • 3. Evaluate the FM.
  • 4. Deploy and run the FM.
phone Contact Us