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
Overview of Edge AI and the Ascend 310
- Insights into Edge AI: key trends, limitations, and applications
- Architecture of the Huawei Ascend 310 chip and its compatible toolchain
- The role of CANN in the edge AI deployment hierarchy
Preparing and Converting Models
- Exporting trained models from TensorFlow, PyTorch, and MindSpore
- Utilizing ATC to transform models into OM format for Ascend hardware
- Managing unsupported operations and implementing lightweight conversion techniques
Creating Inference Pipelines with AscendCL
- Employing the AscendCL API to execute OM models on the Ascend 310
- Pre/post-processing of inputs/outputs, memory management, and device oversight
- Implementation within embedded containers or minimal runtime environments
Optimizing for Edge Limitations
- Minimizing model size and adjusting precision (FP16, INT8)
- Leveraging the CANN profiler to pinpoint performance bottlenecks
- Managing memory structure and data streaming for optimal performance
Deployment using MindSpore Lite
- Applying the MindSpore Lite runtime for mobile and embedded platforms
- Evaluating MindSpore Lite against the native AscendCL pipeline
- Bundling inference models for device-specific implementation
Edge Deployment Scenarios and Practical Examples
- Example: Smart camera featuring an object detection model on the Ascend 310
- Example: Real-time classification within an IoT sensor hub
- Supervising and updating models deployed at the edge
Recap and Future Directions
Requirements
- Prior experience with AI model creation or deployment processes
- Fundamental understanding of embedded systems, Linux, and Python
- Familiarity with deep learning frameworks like TensorFlow or PyTorch
Target Audience
- Developers of IoT solutions
- Engineers specializing in embedded AI
- Edge system integrators and AI deployment experts
Testimonials (1)
That we can cover advance topic and work with real-life example
