Thank you for sending your enquiry! One of our team members will contact you shortly.
Thank you for sending your booking! One of our team members will contact you shortly.
Course Outline
Introduction to the Huawei Ascend Platform
- Examination of Ascend architecture and its ecosystem
- Overview of MindSpore and CANN functionalities
- Analysis of use cases and industry significance
Configuring the Development Environment
- Installation procedures for the CANN toolkit and MindSpore
- Utilizing ModelArts and CloudMatrix for orchestrating projects
- Validating the setup with example models
Model Development using MindSpore
- Defining models and executing training within MindSpore
- Managing data pipelines and formatting datasets
- Converting models to Ascend-compatible formats
Optimizing Performance on Ascend
- Implementing operator fusion and custom kernels
- Strategies for tiling and AI Core scheduling
- Utilizing benchmarking and profiling utilities
Deployment Methodologies
- Weighing the tradeoffs between edge and cloud deployment
- Employing the MindX SDK for deployment tasks
- Integration processes with CloudMatrix workflows
Debugging and Monitoring Protocols
- Leveraging Profiler and AiD for system tracing
- Resolving runtime failures and errors
- Tracking resource consumption and throughput metrics
Case Studies and Laboratory Integration
- End-to-end pipeline development using MindSpore
- Laboratory exercise: Constructing, optimizing, and deploying a model on Ascend
- Comparative performance analysis against alternative platforms
Conclusions and Future Directions
Requirements
- Basic comprehension of neural networks and AI processes
- Proficiency in Python coding
- Knowledge of model training and deployment workflows
Intended Participants
- AI engineers
- Data scientists operating within the Huawei AI ecosystem
- ML developers utilizing Ascend and MindSpore
21 Hours
Testimonials (1)
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