Hands-On Machine Learning Without Coding

Hands-On Machine Learning Without Coding

Got shocked by reading the title!!! 😱 Yes, this is possible. You can do experiments on the computer vision task. Today, I’ll talk about, that if we have brief knowledge about the basic working of machine learning algorithms i.e., Gathering data, Training model, and Evaluating Results. These are three basic steps everyone should know about behind working the machine learning task. That’s it!!! Well, if you don’t have knowledge of python or R language experience but still want to know how this works. You are in right place. 😌😎

Let’s begin…!!!

Teachable Machine is a web-based tool that makes creating machine learning models fast, easy, and accessible to everyone. The project is named ‘Teachable Machine‘. Here, you can easily create machine learning models for your sites, apps, and more – no expertise or coding required.

How we can use it ?

Image Source: https://teachablemachine.withgoogle.com/

From the above picture, everyone can clear that, we have to first gather the data on which particular data you want to create a model i.e., either you can create a dog and cat classifier for this you have to collect several dogs and cats images. Then you have to categorize the dogs in one class and cats in another class. After that, just click on the button ‘Train Model’. That all we have to do it…!!! 😲

After that, we can export the model for deployment in Web applications or Android applications. You simply can test the model by adding other dogs or cat images. Let’s see how we can do this.!

Let’s build Dog-Cat Classifier

You have to create a project by clicking here. There are only three problems that we can work on i.e. image classification, audio classification, and pose estimation. More models are coming soon.

Choose the image project,

You will find windows like this, after clicking on the image project. Here you can add more than one class as per your requirement. But as per our problem, we have only two classes i.e., dog and cat.

Download the dogs and cats images from google images randomly and upload them by clicking on the upload button.

Remember, to change the class name by clicking on the ✏️ button.

I have uploaded the cats and dogs images, for training a model, only 4 images are not good (only for experiment purposes). Because the more you give to model the more it will learn. Then click on the train model button. You’ll have your model ready to predict the new image.

This is how our screen will look after training gets done. Now, it’s time for the evaluation of our model. Download any random image of a cat or dog and upload it in the preview section.

See, I have downloaded the random image of dog and cat, uploaded it, so the trained model gave the probability score of the dog and cat in the uploaded image. Taadaaa…!!!

This is how you can have a knowledge of working with machine learning. But working in deep with machine learning, AI requires a good understanding and knowledge of coding and algorithms running behind it.