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, including 310, 910, and 910B models.
- Key high-level components: MindSpore, CANN, and AscendCL.
- Industry positioning and core architectural principles.
The Function of CANN in Huawei's AI Stack
- Defining CANN: The purpose of the SDK and its internal layers.
- The roles of ATC, TBE, and AscendCL in compiling and executing models.
- How CANN enables inference optimisation and deployment.
MindSpore: Overview and Architecture
- Training and inference workflows within MindSpore.
- Graph mode, PyNative, and hardware abstraction mechanisms.
- Integration with the Ascend NPU via the CANN backend.
AI Lifecycle on Ascend: From Training to Deployment
- Creating models in MindSpore or converting them from other frameworks.
- Exporting and compiling models using ATC.
- Deploying OM models and AscendCL on Ascend hardware.
Comparative Analysis with Other AI Stacks
- MindSpore versus PyTorch and TensorFlow: Focus and market positioning.
- Deployment workflows on Ascend compared to GPU-based stacks.
- Enterprise opportunities and potential limitations.
Enterprise Integration Scenarios
- Applications in smart manufacturing, government AI, and telecommunications.
- Considerations regarding scalability, compliance, and ecosystem health.
- Hybrid cloud/on-premise deployment strategies using the Huawei stack.
Conclusion and Recommended Next Steps
Requirements
- General awareness 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 integrators.
- Technology managers and strategic decision-makers.
