Developing AI Applications with Huawei Ascend and CANN

• An understanding of neural networks and AI workflows
• Experience with Python programming
• Familiarity with model training and deployment pipelines

Audience

• AI engineers
• Data scientists working with Huawei AI stack
• ML developers using Ascend and MindSpore

Huawei Ascend is a family of AI processors designed for high-performance inference and training.

This instructor-led, live training (online or onsite) is aimed at intermediate-level AI engineers and data scientists who wish to develop and optimize neural network models using Huawei’s Ascend platform and the CANN toolkit.

By the end of this training, participants will be able to:

• Set up and configure the CANN development environment.
• Develop AI applications using MindSpore and CloudMatrix workflows.
• Optimize performance on Ascend NPUs using custom operators and tiling.
• Deploy models to edge or cloud environments.

Format of the Course

• Interactive lecture and discussion.
• Hands-on use of Huawei Ascend and CANN toolkit in sample applications.
• Guided exercises focused on model building, training, and deployment.

Course Customization Options

• To request a customized training for this course based on your infrastructure or datasets, please contact us to arrange.

Introduction to Huawei Ascend Platform

• Overview of Ascend architecture and ecosystem
• MindSpore and CANN overview
• Use cases and industry relevance

Setting Up the Development Environment

• Installing the CANN toolkit and MindSpore
• Using ModelArts and CloudMatrix for project orchestration
• Testing the environment with sample models

Model Development with MindSpore

• Model definition and training in MindSpore
• Data pipelines and dataset formatting
• Exporting models to Ascend-compatible format

Performance Optimization on Ascend

• Operator fusion and custom kernels
• Tiling strategy and AI Core scheduling
• Benchmarking and profiling tools

Deployment Strategies

• Edge vs cloud deployment tradeoffs
• Using the MindX SDK for deployment
• Integration with CloudMatrix workflows

Debugging and Monitoring

• Using Profiler and AiD for tracing
• Debugging runtime failures
• Monitoring resource usage and throughput

Case Study and Lab Integration

• Full pipeline development using MindSpore
• Lab: Build, optimize, and deploy a model on Ascend
• Performance comparison with other platforms

Summary and Next Steps