Get in Touch

Course Outline

Introduction to Cambricon and MLU Architecture

  • Overview of Cambricon’s portfolio of AI chips
  • Details on MLU architecture and instruction pipeline
  • Supported model types and potential use cases

Setting Up the Development Toolchain

  • Installation of BANGPy and Neuware SDK
  • Configuring environments for Python and C++
  • Managing model compatibility and preprocessing

Developing Models with BANGPy

  • Managing tensor structures and shapes
  • Constructing computation graphs
  • Support for custom operations within BANGPy

Deployment via Neuware Runtime

  • Converting and loading models
  • Controlling execution and inference
  • Best practices for deploying to edge and data centers

Performance Optimization

  • Tuning layers and memory mapping
  • Profiling and execution tracing
  • Identifying and resolving common bottlenecks

Integrating MLU into Applications

  • Utilizing Neuware APIs for application integration
  • Supporting streaming and multi-model scenarios
  • Implementing hybrid CPU-MLU inference setups

End-to-End Project and Use Case

  • Lab exercise: Deploying a vision or NLP model
  • Performing edge inference with BANGPy integration
  • Evaluating accuracy and throughput

Summary and Next Steps

Requirements

  • A solid grasp of machine learning model structures
  • Practical experience with Python and/or C++
  • Familiarity with the concepts of model deployment and acceleration

Target Audience

  • Developers specializing in embedded AI
  • ML engineers focusing on deployment to edge or data centers
  • Developers working within Chinese AI infrastructure
 21 Hours

Number of participants


Price per participant

Testimonials (1)

Upcoming Courses

Related Categories