Overview
Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!
Explore an open-source framework designed to streamline the implementation of neural networks on embedded devices in this 39-minute tinyML Talk. Learn about AutoFlow, a tool that automates the entire workflow for data scientists, from building machine learning models to selecting target platforms and optimizing implementations. Discover how AutoFlow utilizes Automated Machine Learning (AutoML) to generate and train various neural networks, selecting the most accurate one. Gain insights into pruning and quantization techniques for reducing model size, and understand the process of converting models for specific target platforms. Follow along with a demonstration of AutoFlow's features, including its ability to generate necessary files for execution on embedded devices. Find out how to access and contribute to this GitHub-hosted project, and participate in a Q&A session to deepen your understanding of this innovative framework for embedded AI development.
Syllabus
Introduction
Introducing Daniel
What is AutoFlow
Pruning
Implementation
Quantization
Quantization versions
Features
Demonstration
How to take part
Github page
Additional information
Summary
QA
Taught by
tinyML