TinyMLEdge: A Workflow for Deploying TinyML Models in Industrial Edge Devices

In the current machine learning landscape, numerous algorithms and new models continuously emerge, demon-strating impressive capabilities across various tasks. However, models specifically tailored for particular environments remain limited. Despite their powerful performance, models like YOLO encounter challenges in small object detection and require substantial adjustments for different datasets and tasks. Further-more, their extensive parameters make them difficult to deploy on edge devices. Addressing these issues, TinyMLEdge is proposed as a workflow designed to facilitate the deployment of embedded TinyML models in specialized industrial edge devices.