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 21 hours
Course Outline
Introduction to Cambricon and MLU Architecture
- An overview of Cambricon’s AI chip portfolio
- Details on MLU architecture and its instruction pipeline
- Supported model types and their primary use cases
Installing the Development Toolchain
- Installation of BANGPy and the Neuware SDK
- Setting up environments for both Python and C++
- Ensuring model compatibility and handling preprocessing
Model Development with BANGPy
- Managing tensor structures and shapes
- Building computation graphs
- Incorporating custom operations within BANGPy
Deploying with Neuware Runtime
- Processes for converting and loading models
- Controlling execution and inference
- Best practices for edge and data center deployment
Performance Optimization
- Tuning memory mapping and specific layers
- Utilizing execution tracing and profiling tools
- Identifying and resolving common performance bottlenecks
Integrating MLU into Applications
- Leveraging Neuware APIs for seamless application integration
- Implementing streaming and multi-model support
- Designing hybrid CPU-MLU inference scenarios
End-to-End Project and Use Case
- Practical lab: Deploying a vision or NLP model
- Implementing edge inference with BANGPy integration
- Evaluating accuracy and throughput performance
Summary and Next Steps
Requirements
- A solid grasp of machine learning model architectures.
- Practical experience with Python and/or C++.
- Familiarity with the fundamentals of model deployment and hardware acceleration.
Target Audience
- Developers specializing in embedded AI applications.
- ML engineers responsible for deploying solutions to edge or data center infrastructure.
- Developers working within the Chinese AI infrastructure ecosystem.
Testimonials (1)
That we can cover advance topic and work with real-life example
