This repository contains the colab code to prepare keywords speech.
This colab file can be used to build a custom dataset to train a machine learning model using Edge Impulse.
It uses Google Speech Commands dataset as base dataset. The colab script prepares the wav file so that they can be uploaded into the Edge Impulse using Data Forwarder. Moreover, you can add unknown words and noise.
To have a full description look at the post: Arduino Edge Impulse with Google Speech Datasets