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.
Duration 14 hours
Course Outline
Introduction to the Huawei AI Ecosystem
- Overview of Ascend AI hardware: 310, 910, and 910B
- Key high-level components: MindSpore, CANN, and AscendCL
- Strategic industry positioning and architectural principles
The Function of CANN within Huawei’s AI Stack
- Defining CANN: The SDK's purpose and its internal layers
- ATC, TBE, and AscendCL: Processes for compiling and running models
- CANN's contribution to inference optimization and deployment
MindSpore: Architecture and Core Concepts
- Managing training and inference workflows in MindSpore
- Exploring Graph mode, PyNative, and hardware abstraction
- Connecting with the Ascend NPU through the CANN backend
The AI Lifecycle on Ascend: From Training to Deployment
- Building models in MindSpore or converting them from other frameworks
- Utilizing ATC to export and compile models
- Deploying OM models and AscendCL on Ascend hardware
Benchmarking Against Other AI Stacks
- Comparing MindSpore with PyTorch and TensorFlow: Focus and market position
- Contrasting deployment workflows on Ascend versus GPU-based stacks
- Evaluating opportunities and constraints for enterprise adoption
Enterprise Integration Scenarios
- Applications in smart manufacturing, government AI, and telecommunications
- Considering scalability, compliance, and ecosystem factors
- Implementing cloud/on-premises hybrid deployments with the Huawei stack
Recap and Future Directions
Requirements
- General knowledge of AI workflows or platform architecture
- Fundamental grasp of model training and deployment processes
- No previous hands-on experience with CANN or MindSpore is necessary
Target Audience
- AI platform evaluators and infrastructure architects
- AI/ML DevOps specialists and pipeline integration experts
- Technology managers and strategic decision-makers
