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 7 hours
Course Outline
Overview of Huawei’s AI Ecosystem
- Ascend AI hardware specifications: 310, 910, and 910B chips
- MindSpore, CANN, and associated supporting tools
- AI development workflow: transitioning from training to deployment
Grasping the CANN Toolkit
- Definition of CANN and its significance
- Summary of essential elements (ATC, AscendCL, operator libraries)
- Function of CANN within AI inference workflows
Commencing with MindSpore and CANN
- Configuring the environment (MindSpore + CANN + Python)
- Training a fundamental model within MindSpore
- Exporting and converting the model via ATC
Executing Inference on Ascend Devices
- Utilizing the OM model through AscendCL or Python APIs
- Basic input and output preprocessing
- Verifying model outputs
Integrating with Alternative Frameworks
- Overview of compatibility with TensorFlow, PyTorch, and ONNX
- Compatible operators and associated constraints
- Demonstration of a simple model conversion (e.g., from ONNX to OM)
Investigating the CANN and MindSpore Developer Landscape
- Essential resources: documentation, GitHub repositories, sample code
- Summary of MindSpore Hub and the model zoo
- Community forums, events, and support avenues
Recap and Future Pathways
Requirements
- Fundamental comprehension of machine learning and deep learning principles
- Proficiency in Python programming
- No previous hands-on experience with CANN or Ascend hardware is necessary
Target Participants
- Machine learning professionals investigating deployment processes
- Students and researchers encountering Huawei’s AI ecosystem for the first time
- AI framework contributors and enthusiasts interested in accelerating model performance
