Get in Touch

Course Outline

Overview of Huawei CloudMatrix

  • The CloudMatrix ecosystem and its deployment lifecycle
  • Compatible models, data formats, and deployment modes
  • Common use cases and supported chipset architectures

Model Preparation for Deployment

  • Exporting models from training frameworks (MindSpore, TensorFlow, PyTorch)
  • Utilizing the Ascend Tensor Compiler (ATC) for format conversion
  • Handling static versus dynamic shape models

Deployment to CloudMatrix

  • Creating services and registering models
  • Launching inference services via the UI or command-line interface (CLI)
  • Implementing routing, authentication, and access control mechanisms

Serving Inference Requests

  • Differentiating between batch and real-time inference workflows
  • Establishing data preprocessing and postprocessing pipelines
  • Integrating CloudMatrix services with external applications

Monitoring and Performance Optimization

  • Analyzing deployment logs and tracking requests
  • Managing resource scaling and load balancing
  • Optimizing latency and maximizing throughput

Enterprise Tool Integration

  • Connecting CloudMatrix with OBS and ModelArts
  • Leveraging workflows and model versioning strategies
  • Implementing CI/CD pipelines for model deployment and rollback

Complete Inference Pipeline

  • Deploying a full image classification pipeline
  • Conducting benchmarking and accuracy validation
  • Simulating failover scenarios and system alerting

Recap and Future Directions

Requirements

  • A solid grasp of AI model training workflows
  • Proficiency with Python-based machine learning frameworks
  • Foundational knowledge of cloud deployment principles

Target Audience

  • AI Operations (AIOps) teams
  • Machine Learning Engineers
  • Cloud Deployment Specialists utilizing Huawei infrastructure
 21 Hours

Number of participants


Price per participant

Testimonials (2)

Upcoming Courses

Related Categories