Hello everyone! This is an intro for Teachable Machine, a web-based platform that allows for the easy, accessible, and free use of machine learning models. You can make a machine learning model without even knowing machine learning coding.

You can make any classification model for image, voice, and video in only 5 minutes just on your browser.  The original version of the app was just a nice little experiment that could be used to identify visual data from your webcam. Yet Google also introduced new modes to the program along with the ability to export developed models, making Teachable Machine 3.0 a more robust framework for creating actual AI devices.

If some people feel worried that working with machine learning is going to be over their heads, the team behind Teachable Machine would like to eliminate the anxiety and replace it with excitement.

“We’re wondering, what if it weren’t? It’s pretty hard to get into machine learning,” said the narrator.”

How Teachable machine works

You can use the Teachable Machine to identify pictures, sounds, or poses. Google Creative Lab’s Teachable Machine works to train a model based on the examples you’ve given. Training happens in your own browser, he said; it’s all in your machine.

How to train Teachable machine

You can train this model in just 5 minutes. In order to start teaching the machine, we first click on the project we want to create. i.e. image, sound, pose, etc. which is encircled in the red box.

Teachable machine

After that, we have to build various groups, or classes, to teach it. For in the case of classification between two persons, we make three different classes; two classes for each person and one for not both of them. By default, two classes are added you can change their names by clicking on class 1, class 2, etc. as shown in black encircled box. While you can also add a new class from “Add a class”. You can make a live dataset from clicking on webcam or upload pre-captured images from the upload option.

Data Science

Now I have all classes, I’m going to send examples of every class to learn from. In this case, these are images of two persons say A and B— so let’s start with a Person A that’s “A” Then we’ll do the same thing with a Person B: And finally, an image with a person not from both.

Now that I’ve got all my classes prepared, I can click on the train button as shown in the picture in green encircled.

Don’t switch the tab while training. Then on the right, preview that the model works!

You can watch a complete tutorial video on this

Best Machine Learning Tool-Google ML Kit for Mobile
More Applications

LEAVE A REPLY

Please enter your comment!
Please enter your name here