TinyML for smart devices

Inspiration

Training a single AI model can emit as much carbon as five cars in their lifetimes

What it does

First we quantize a machine learning model to make it low power. Then we load the model on the Arduino and by using a tiny camera, we can do detection, like handwritten digit classification

How we built it

We used a tiny camera and Arduino nano 33 IOT.

Challenges we ran into

My device is very low power and I can use PV cells as the source of power, but The PV cells has not delivered on time.

Accomplishments that we're proud of

Enable Ml algorithm and AI to be run on resource constrained edge devices.

What we learned

How to code in ardiono

What's next for Always-on TinyML

Adding solar cell for the source of energy.

Built With

• arduino
• c++
• python
• tensorflow