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.
Course Outline
Introduction to Cambricon and MLU Architecture
- Overview of Cambricon’s AI chip portfolio for government applications
- Detailed explanation of the MLU architecture and its instruction pipeline
- Supported model types and relevant use cases for government operations
Installing the Development Toolchain
- Steps to install BANGPy and Neuware SDK for government use
- Environment setup procedures for Python and C++ in a government context
- Model compatibility and preprocessing considerations for government projects
Model Development with BANGPy
- Understanding tensor structure and shape management for government applications
- Constructing computation graphs to optimize performance for government tasks
- Support for custom operations in BANGPy tailored to government needs
Deploying with Neuware Runtime
- Converting and loading models into the Neuware runtime environment for government use
- Controlling execution and inference processes for efficient government operations
- Best practices for deploying MLU in edge and data center environments for government applications
Performance Optimization
- Techniques for memory mapping and layer tuning to enhance performance for government tasks
- Execution tracing and profiling methods to identify and resolve bottlenecks in government applications
- Common performance issues and solutions specific to government use cases
Integrating MLU into Applications
- Utilizing Neuware APIs for seamless integration of MLU into government applications
- Support for streaming and multi-model scenarios in government operations
- Implementing hybrid CPU-MLU inference solutions to meet government requirements
End-to-End Project and Use Case
- Laboratory exercise: Deploying a vision or NLP model for government use
- Edge inference implementation with BANGPy integration in government settings
- Evaluating accuracy and throughput to ensure compliance with government standards
Summary and Next Steps
Requirements
- An understanding of machine learning model structures
- Experience with Python and/or C++
- Familiarity with concepts related to model deployment and acceleration
Audience
- Embedded AI developers for government and private sector
- Machine learning engineers deploying solutions to edge devices or data centers
- Developers working with Chinese AI infrastructure for government applications
21 Hours
Testimonials (1)
That we can cover advance topic and work with real-life example
