From the course: Computer Vision on the Raspberry Pi 4
Join today to access over 23,200 courses taught by industry experts.
Executing models with TensorFlow Lite
From the course: Computer Vision on the Raspberry Pi 4
Executing models with TensorFlow Lite
- [Instructor] An earlier video demonstrated how to train a CNN, and save it to a file that can be accessed by TensorFlow Lite. The goal of this video is to provide a proper introduction to TensorFlow Lite. Once you understand this, you'll be able to execute CNNs on a Raspberry Pi. The complete TensorFlow package is very large. So Google provides a stripped down version that only occupies about one megabyte of memory. This is ideal for smartphones, microcontrollers, and single-board computers like the Raspberry Pi. TensorFlow Lite doesn't have the resources to create or train neural networks, but it can execute models at high speed. To see how this works, you need to become familiar with the interpreter class. An interpreter takes care of loading the model, receiving input data and executing the model. In this course, we'll access the interpreter in a five-step process. The first step creates an interpreter…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.