OctoAI Logo
Sign up
Log in
Sign up
Log in
Home
Blog

Prototyping Machine Learning with Arduino: TVMCon Tutorial Dec 15

Blog Author - Gavin  Uberti

Dec 8, 2021

2 minutes

Machine learning neural networks for vision and audio spotting are notoriously difficult to get working. Running these models on small, resource constrained microcontrollers can be nearly impossible. That’s why on December 15th at 1:00 PM PST, we’re holding a free 45 minute tutorial session on how Apache TVM can be used to prepare and run neural networks on Arduino.

Imagine this - an AI-powered toy that barks when it sees your dog. In this tutorial, we will build a prototype of this dog toy using microTVM, an Arduino board, and an attached camera. By the end of the tutorial we will demonstrate the finished device in action, and you’ll be ready to build your own AI-powered projects. With the prototype development complete, we’ll discuss how TVM’s portability makes it easy to scale up and automate deployment to a final product.

The tutorial will be held at TVMCon 2021 on December 15th 2021 at 1:00 PM PST. The talk will be streamed live, and followed by a Q&A session. We’ll also be giving away 10 Spresense development kits, so make sure you register for the event!

What is Apache TVM?

Apache TVM is the open source stack for ML performance and portability. microTVM is a project within TVM that automatically produces optimized code to implement ML models in bare-metal and RTOS environments like Arduino.

In time for TVMcon (the open source conference for ML acceleration), the community is releasing significant usability enhancements that empower users to turn a trained model that they can pull “off the shelf” into an Arduino project. This speaks to the simplicity of the integration. In terms of the flexibility, these microTVM generated projects are compatible with almost all Arduino boards and the code is automatically generated for the deployment hardware (as long as they have sufficient memory to store the model). In addition, microTVM is flexible in supporting models from seven deep learning (DL) frameworks including TensorFlow, TFLite, Pytorch and ONNX.

How can I tune in?

Want to learn more and earn a chance to win 1 of 10 new Sony Spresense Packs? Register and attend the free Machine Learning conference, TVMCon, taking place December 15-17.

OUR BLOG

Related Posts

All Posts
2021-11-02
TVMCon registration is open and speakers announced

RSVP for free to join the Apache TVM community December 15-17 for TVMCon! The 2021 Apache TVM and Open Source ML Acceleration Conference will be an all-virtual event...

Blog Author - Chris  Hoge
Chris Hoge
2021-10-20
Accelerating machine learning at the edge: Arm Cortex-A72 CPU now available in OctoML

Today we’re excited to announce our partnership with Arm, which highlights our collaboration across a broad array of hardware and embedded systems.

Blog Author - Sayce Falk
Sayce Falk