Tutorials

Aidge 101

To get started with Aidge, please follow the Aidge demonstration tutorial. This tutorial demonstrates the basic features of the Aidge Framework, importing an ONNX, transforming a neural network graph, performing inference and a cpp export.

Aidge demonstration

Aidge DNN fonctionnalities

• Manipulating databases and creating batches of data

• Train a Deep Neural Network

• Provide an operator implementation using Python or meta-operators

• Perform advanced graph matching with the Graph Regular Expression tool

Database MNIST

Training a simple Neural Network

Aidge ONNX tutorial

Graph Regular Expression

DNN Optimization

• Perform post Training Quantization

• Perform Convolution tiling

Post Training Quantization with AIDGE

Aidge tiling demonstration

DNN export

Export a MNIST model to a CPP standalone project

• Exercise on adding a custom implementation for a cpp export

• Export your DNN with TensorRT

• Export your DNN for an STM32

Tutorial on adding the C++ Aidge

You can extend our operator coverage by adding an operator and its implementation in the C++ Aidge library. The Add an operator and its implementation Tutorial details the steps to follow.

For more information on contributing to Aidge, please refer to the wiki.

If you encounter any difficulty with the Tutorials, please create an issue here.