Course Outline
Introduction to the Huawei Ascend Platform
- Overview of the Ascend architecture and ecosystem
- Introduction to MindSpore and CANN
- Industry use cases and relevance
Establishing the Development Environment
- Installation of the CANN toolkit and MindSpore
- Utilizing ModelArts and CloudMatrix for project orchestration
- Environment validation using sample models
Model Development with MindSpore
- Defining and training models in MindSpore
- Managing data pipelines and dataset formatting
- Converting models to Ascend-compatible formats
Performance Optimization on Ascend
- Operator fusion and custom kernels
- Tiling strategies and AI Core scheduling
- Benchmarking and profiling tools
Deployment Strategies
- Trade-offs between edge and cloud deployment
- Employing the MindX SDK for deployment
- Integration with CloudMatrix workflows
Debugging and Monitoring
- Using Profiler and AiD for tracing
- Resolving runtime failures
- Tracking resource usage and throughput
Case Study and Lab Integration
- End-to-end pipeline development using MindSpore
- Lab: Constructing, optimizing, and deploying a model on Ascend
- Performance comparison with alternative platforms
Summary and Next Steps
Requirements
- Knowledge of neural networks and AI workflows
- Proficiency in Python programming
- Familiarity with model training and deployment pipelines
Target Audience
- AI engineers
- Data scientists utilizing the Huawei AI stack
- ML developers employing Ascend and MindSpore
Testimonials (2)
That i gained a knowledge regarding streamlit library from python and for sure i'll try to use it to improve applications in my team which are made in R shiny
Michal Maj - XL Catlin Services SE (AXA XL)
Course - GitHub Copilot for Developers
Trainer able to adjust the course level during training to fit our understanding level on the topic, so that we could gain more useful knowledge that could further help us harness the tools in our daily works.