This is part four of the Tiny ML Project, Edge Impulse is covered in in Part three. On the deployment page on Edge Impulse. Select on Arduino. Then you can build a library. It will download once it is done. You should unzip and copy the content to the Arduino…

This is part 3 of a series of tutorial setting up a Neural Network to the Wio Terminal. The data processing part is in https://sweemeng.medium.com/sound-recognition-on-wio-terminal-part-2-fsd50k-dataset-dabb52ce3951 I will skip a few things because Edge Impulse tutorial is good. First is, Edge Impulse daemon and uploader. This is what I used to…

Note: The is part two of a series to put a Neural Network to the Wio Terminal Device. Starting from here. The FSD50K dataset is a dataset from Free Sound. It contains around 50k sound clip from https://freesound.org/. It is labelled with label based on Audioset ontology, with some extra…

I have done my introduction to the Wio terminal at my previous post. This device can run TensorFlow lite micro. Meaning it can run a small neural network for to process data onboard. …

Seeed Studio have contacted me on a collaboration on creating content with Wio Terminal. I agreed and I get a device. The Wio Terminal is one of device I was planning to get myself anyway.

M5Stack is a series of ESP32 based hardware with screen and buttons. I used it briefly with a PM2.5 sensor. As well have screen and buttons, this is able to connect to the internet via WIFI, have various GPIO and usable with Arduino. …

Now that we have a model in form of a c header from the previous post. We can start integrating our code. Here I assume the setup from Part 2 is ready. Essentially you can just copy the content from the arduino example from below ArduinoTensorFlowLiteTutorials/IMU_Classifier.ino at master ·… Contribute to arduino/ArduinoTensorFlowLiteTutorials development by creating an account on GitHub.github.com

After collecting data in Part 3. We will train a neural network. I will use google colab for 2 reason. I broke my nvidia setup on my computer. Google colab is a nice way for people to run neural network based code without the hardware. We can potentially turn into…

First the of all the project is Gesture capture project based on Gesture to Emoji project. The difference is I use a different gesture, I am sending a different keystroke and use to control a slide instead of making emoji. But I try to add some commentary of the code. …