Embedded ML & AI

Learn how to use light weight Machine Learning models on Raspberry Pi and explore the possibilities they offer.

Custom Object Detection Model: Train a custom object detection model using Tensorflow Lite Model Maker and run it on Raspberry Pi. Model Maker library simplifies the task of model creation through Transfer Learning using your data. The best part is the whole process can be completed on a browser using a readily available Colab notebook. You just have to prepare your data.

Image Classification Tensorflow Lite Raspberry Pi

Image Classification: Learn to implement Image Classification on Raspberry Pi using Tensorflow Lite. The article explains the working of code in detail. The code picks up an image from the disk, so no need to attach any camera for this project. Models, images and source code used in the project can be downloaded.

Testing 20+ Machine Learning Models on Raspberry Pi

Model Garden: This project is TensorFlow Community Spotlight winner for the month of Jun 2021. A custom Python script is created to run 24 Machine Learning Models on Raspberry Pi that can switch between all models during run time through a Web Interface. You can install this project on your Raspberry Pi by using a bash script provided with this article.

Google coral USB Accelerator with Raspberry Pi performance

Coral USB Accelerator-Raspberry Pi: Measured the performance of 4 models of Raspberry Pi with Google Coral USB Accelerator. Inference time is recorded for running pre-trained Image Classification ML Model (MobileNet V1) on different types of Raspberry Pis. TensorFlow Lite APIs have been used to create the Python script for testing.