It is an experimental project using the architecture described in this article to implement different applications for study.
The version that previously implemented "Edge Impulse for Linux" Python SDK on the Raspberry Pi 4 is in the "concept" branch.
The version that previously implemented Edge TPU (Coral) object detection inference is in the "coral" branch.
The version that currently implements the TinyML development environment is in the "tinyml" branch.

This project uses the "Low Code/No Code" approach to build a deep learning development environment for Edge AI or On-Device AI. It includes a component-based framework and a flow-based visual programming editor. The concept is as follows:

Deep learning procedures can be transformed into components to achieve code reuse and simplify the integration of flow-based programming. These components can be called directly from Python and integrated with the Jupyter Lab environment for a "low-code" approach. Or integrate these components into flow-based visual programming using the ryvencore-qt library to achieve the "No Code" approach.