TinyML Compiler

How To Bring TensorFlow, PyTorch, ONNX, Caffe & Keras Model on Microcontrollers

Microcontrollers are resource constrained devices with very little hard-disk (aka flash) usually under 128KB with a basic CPU operating at 40MHz to 100MHz. No operating system is small enough in size to fit in the flash or is able to boot up without hogging significant compute resources. Under these conditions, machine learning models must be compiled into a binary program that can be loaded into Microcontrollers memory during boot up.

What is a TinyML Compiler

TinyML Compiler takes a trained machine learning model as input and provides a binary library that can be integrated into an application to flash on resource constrained devices like microcontrollers.

The deepC

The deepC is a vendor independent TinyML deep learning library, compiler and inference framework designed for small form-factor devices including μControllers, IoT and Edge devices. It is open source and hosted on GitHub - https://github.com/ai-techsystems/deepC

The deepC is targeted towards devices with small formfactor like microcontrollers, which are part of all sorts of household devices: think appliances, cars, and toys. In fact, there are around 30 billion microcontroller-powered devices produced each year. They're cheap, require very little energy, and are very reliable.

How to Compile

One can download and install the deepC step by step by following the instructions mentioned in https://github.com/ai-techsystems/deepC#readme or you can use a proprietary compiler hosted by AITS at https://cainvas.ai-tech.systems/compiler/