How to Use Audio in Teachable Machine: A Step-by-Step Guide

Teachable Machine is a web-based tool by Google that allows you to create machine learning models without writing any code. Using audio data, you can teach a model to recognize various sounds, such as speech, music, or specific noises. In this step-by-step guide, you'll learn how to upload and record audio samples, train a model, and test its accuracy. Whether you’re creating a sound recognition system or experimenting with AI, Teachable Machine makes it easy to get started with audio-based machine learning projects.

Here are the steps to use audio in Teachable Machine.

1. Go to Teachable Machine

• Visit Teachable Machine.

2. Create a New Project

• Click on the Get Started button.

Select the Audio Project option to work with sound data.

3. Set Up Your Classes

1. Define different classes (categories) for your audio samples. For example, you could have classes like "Clap," "Snap," or "Speech."

Click the three dots button to create a new class.

4. Record Audio Samples

1. For each class, click the Record button to start recording audio with your microphone.

2. Record multiple audio samples for each class to train the model effectively.

You can add as many recordings as you like for each class.

5. Train the Model

1. After recording enough samples for each class, click the Train Model button.

Teachable Machine will analyze the audio and train the model to recognize patterns in your data.

6. Test Your Model

• After training, you can test the model by recording new audio or uploading files to see how well it can predict the class of the new audio.

• You’ll see predictions in real time as it listens to new input.

7. Export the Model

• When you’re satisfied with the model’s performance, click the Export Model button.

• Choose how you'd like to use your model. You can export it for web use or integrate it into your projects.

  

Conclusion:

Using audio in Teachable Machine is a straightforward and powerful way to build sound recognition models without needing advanced coding skills. You can create custom AI models that recognize various sounds by recording your audio samples, training your model, testing it, and exporting it. Whether you’re working on a personal project or exploring machine learning, Teachable Machine makes it easy to dive into audio-based AI applications. You can apply your new model to various projects and experiment with sound-based recognition!

It's the same process for the Pose model also, but make sure you record poses from different angles and lighting conditions to help the model generalize better and keep the background as simple as possible to avoid confusing the model.

That’s it! You now have a custom pose recognition model using Teachable Machine!