租户标识:
Product Overview
AI Cloud Lab

Based on cloud-native technologies, AI Cloud Lab provides AI developers with a professional and flexible development environment and components, and links AI development process nodes: with AI Cloud Lab, you can mount cloud storages to get data and model information, call computing power of different scales in the compute pool for code debugging, model training, iteration, etc., and link other AI services to obtain rich services such as data management and model visualization. With professional R&D tools, flexible cloud computing resources, smooth development experience, stable access services, and open configuration modes, AI Cloud Lab helps AI developers create AI applications more efficiently and conveniently, and facilitates the intelligent upgrading of the industry.

Product Superiority
Rely on SenseTime's years of knowledge accumulation in AI R&D, optimize the R&D experience to the extreme, and flexibly activate or deactivate resources at any time.
  • 01AI R&D-focused
  • 02Flexible activation
  • 03Extremely elastic computing power
  • 04Open AI service
AI R&D-focused
01AI R&D-focused

- Optimize mainstream AI R&D scenarios in depth
- Preset more professional deep learning algorithms and more pre-trained models
- Provide a deep learning framework with better training performance

Flexible activation
02Flexible activation

- Need no building or configuration of complicated development environment
- Online, local, and multi-node diverse connection modes
- Activation/deactivation at any time to realize flexible release of resources

Extremely elastic computing power
03Extremely elastic computing power

- Open the development environment with a small amount of resources to achieve code debugging
- Link AI computing pools on demand to initiate large-scale distributed training

Open AI service
04Open AI service

- Connect other rich AI developer tools inside SenseCore
- Support users to enable external common Web services
- Support user-defined installation of other commonly used plug-ins

AI R&D-focused
01
AI R&D-focused

- Optimize mainstream AI R&D scenarios in depth - Preset more professional deep learning algorithms and more pre-trained models - Provide a deep learning framework with better training performance

Flexible activation
02
Flexible activation

- Need no building or configuration of complicated development environment - Online, local, and multi-node diverse connection modes - Activation/deactivation at any time to realize flexible release of resources

Extremely elastic computing power
03
Extremely elastic computing power

- Open the development environment with a small amount of resources to achieve code debugging - Link AI computing pools on demand to initiate large-scale distributed training

Open AI service
04
Open AI service

- Connect other rich AI developer tools inside SenseCore - Support users to enable external common Web services - Support user-defined installation of other commonly used plug-ins

01
/
04
Product Features
Provide a development environment with diverse resource specifications, preset mainstream AI algorithm frameworks, and Link other AI development tools.
  • Diverse resource specification configurations
    Diverse resource specification configurations

    According to different business needs and task needs of AI researchers, flexibly configure cloud development environments with different CPU, GPU, and disk memory specifications;

  • Preset AI dedicated images
    Preset AI dedicated images

    Based on SenseTime's years of experience, preset mainstream and efficient algorithms, training frameworks, and AI component packages in the AI R&D process, eliminating a lot of cumbersome installation work and enabling instant use;

  • Lifecycle management
    Lifecycle management

    Support multiple actions on the created AI Cloud Lab, realize enabling/disabling at any time, and keep the service status of the development environment under control;

  • Cloud code debugging
    Cloud code debugging

    Quickly access the native IDE functions of the cloud through the browser, compile and debug codes, and build models without the need of constructing cumbersome infrastructures and environments;

  • Multiple access methods
    Multiple access methods

    Realize remote access via a variety of local clients, such as vs code and juypter lab, and simultaneously enjoy local stable development experience and rich resources in the cloud;

  • Link AI Compute Pool
    Link AI Compute Pool

    Use a small amount of resources to complete code debugging in the AI Cloud Lab, link the AI Compute Pool, initiate large-scale distributed training tasks, and take advantage of extremely elastic computing power.

Application Scenarios
Cover multiple links in the AI R&D process, and improve the efficiency of algorithm construction, model iteration, and AI application implementation.
  • 01AI algorithm construction
  • 02AI model training
  • 03AI model evaluation
  • 04AI model compilation
  • 05AI service encapsulation
AI algorithm construction
AI algorithm construction
Based on the built-in algorithm components and models in the AI Cloud Lab, adjust and debug relevant backbone, head, neck, and other network parts according to business and implementation requirements.

Containerized environment construction, supporting multiple frameworks, systems, and components.

AI model training
AI model training
Based on the built-in training frameworks and computing power resources in the AI Cloud Lab, initiate training and parameter adjustment for built algorithms, and link AI​training pools to initiate larger-scale tasks.

Enhanced container isolation, ensuring that the development environment runs independently and securely.

AI model evaluation
AI model evaluation
Based on new datasets imported locally or online, initiate model evaluation and debug test codes for trained models or built-in models in the AI Cloud Lab.

Cloud repository service for quickly building multi-purpose images.

AI model compilation
AI model compilation
Convert formats of trained models, compile operators, and quantify and compress models according to the target deployment equipment and architecture.

Cloud-native IDE for comprehensive management of files, software packages, environments, etc.

AI service encapsulation
AI service encapsulation
Pre- and post-process compiled and converted models, and arrange and encapsulate AI services according to the actual business implementation requirements for the final AI application.

Dynamic port configuration, supporting multiple remote access and service access modes.

01AI algorithm construction
02AI model training
03AI model evaluation
04AI model compilation
05AI service encapsulation
AI algorithm construction
AI algorithm construction
Based on the built-in algorithm components and models in the AI Cloud Lab, adjust and debug relevant backbone, head, neck, and other network parts according to business and implementation requirements.

Containerized environment construction, supporting multiple frameworks, systems, and components.

AI model training
AI model training
Based on the built-in training frameworks and computing power resources in the AI Cloud Lab, initiate training and parameter adjustment for built algorithms, and link AI​training pools to initiate larger-scale tasks.

Enhanced container isolation, ensuring that the development environment runs independently and securely.

AI model evaluation
AI model evaluation
Based on new datasets imported locally or online, initiate model evaluation and debug test codes for trained models or built-in models in the AI Cloud Lab.

Cloud repository service for quickly building multi-purpose images.

AI model compilation
AI model compilation
Convert formats of trained models, compile operators, and quantify and compress models according to the target deployment equipment and architecture.

Cloud-native IDE for comprehensive management of files, software packages, environments, etc.

AI service encapsulation
AI service encapsulation
Pre- and post-process compiled and converted models, and arrange and encapsulate AI services according to the actual business implementation requirements for the final AI application.

Dynamic port configuration, supporting multiple remote access and service access modes.

Continuously update the whole line of products and insist on sincere communication and win-win cooperation

Help you achieve new breakthroughs in business with professional AI solutions and advanced AI products