Contents
This is great news for machine learning students. Google has launched a machine learning platform. Then this time you have to build an ML project without code. Google Applied AI (Artificial Intelligence) Engineer Dale Markowitz Introduce a Machine learning project without code.
Yes, of course, you can learn machine learning without programming knowledge, Today, you will see the creation of a machine learning model right here, now, and learn Machine learning without Programming. I will show how to use a tool called Teachable Machine to train a model faster from my browser. Let’s do it Today, I thought it would be fun to train a model.
Project: Train Model to Recognize Letters (Human pose estimation)
That recognizes letters made from In the body, such as A, B, C. But for starters, But in the project she will only train the model to recognize the letters Y, M, C, A.
I will use teachable machines for this. It is one of the best ML tools. Because it allows you to train a model without writing any code. Which is for photos, audio, or currency and plays directly in your browser.
To get started you need a computer with a webcam. So first of all, go to Teachable Machine and click on Get Started.
From here, you will see that there are three different models that can be trained in audio and pose, but we will project a “pose”. The intention of machine learning is that computers be trained to find patterns in data.
1. Collect Training Data (Machine learning)
In the UI, you can collect training data from a webcam, so I’ll create a set of examples of creating Y characters show in Body.
First, create a new square with what you want to find – here, it has to find the letter Y.
As you can see, the UI is already detecting my pose and my facial expressions using a pre-trained model, but what I really want to do is to recognize the characters I’m creating.
Let us take some example pictures. Take your photos from many different angles and distances from the camera, which will prevent the model from overfitting in-room or camera angle features. Now I will collect some data for M, C, and A.
2. Train Your Model
Click on the trained model. When you push the button, TensorFlow.js actually trains a new neural network from the data in the browser, so do not close this window. Your computer is working a lot under the hood. Click Advanced and Under the Hood and learn how it works.
3. Test Your Model
Once the model is trained, you can try it in this browser only. It works! The great thing about teachable machines is that exporting your model is really easy.
Download Machine Learning Model Code
You can either download it as a TensorFlow model file, or you can have Teachable Machines host it for you, and it will also give you a shareable link. Also, a good JavaScript code sample. Which can be easily used to add the model to a web app.
