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
Core Principles of TinyML Pipelines
- An overview of the different stages in a TinyML workflow
- Key characteristics of edge hardware
- Strategic considerations for pipeline design
Data Acquisition and Preprocessing
- Gathering structured data and sensor inputs
- Strategies for data labeling and augmentation
- Formatting datasets for resource-constrained environments
Model Development for TinyML
- Choosing appropriate model architectures for microcontrollers
- Establishing training workflows with standard ML frameworks
- Assessing key model performance metrics
Model Optimization and Compression
- Applying quantization methods
- Utilizing pruning and weight sharing
- Achieving a balance between accuracy and resource limitations
Model Conversion and Packaging
- Exporting models to TensorFlow Lite
- Integrating models within embedded toolchains
- Handling model size and memory restrictions
Deployment on Microcontrollers
- Writing models to hardware targets
- Setting up run-time environments
- Conducting real-time inference tests
Monitoring, Testing, and Validation
- Defining test strategies for deployed TinyML systems
- Troubleshooting model behavior on physical hardware
- Verifying performance under field conditions
Assembling the Complete End-to-End Pipeline
- Creating automated workflows
- Implementing version control for data, models, and firmware
- Overseeing updates and iterative improvements
Conclusion and Future Directions
Requirements
- A solid grasp of machine learning fundamentals
- Practical experience in embedded programming
- Knowledge of Python-based data workflows
Target Audience
- AI engineers
- Software developers
- Experts in embedded systems
