The Project | About Us | Contribute | Donations | License
HOME My courses Completed
◄ Previous
◄

TensorFlow.js

Next ►
►

TensorFlow.js is an open-source library that allows you to define, train, and run machine learning models entirely in the browser using JavaScript. This module will cover the basics of TensorFlow.js, including its setup, core concepts, and practical examples.

What is TensorFlow.js?

TensorFlow.js brings the power of TensorFlow to the JavaScript ecosystem, enabling machine learning in web applications. Here are some key points:

  • Browser-based ML: Run machine learning models directly in the browser.
  • Node.js support: Use TensorFlow.js in server-side applications with Node.js.
  • Pre-trained models: Utilize pre-trained models for common tasks like image classification, object detection, and more.
  • Custom models: Build and train custom models using JavaScript.

Setting Up TensorFlow.js

In the Browser

To use TensorFlow.js in a web application, include the TensorFlow.js library in your HTML file:

<!DOCTYPE html>
<html>
<head>
  <title>TensorFlow.js Example</title>
  <script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs"></script>
</head>
<body>
  <script>
    // Your TensorFlow.js code will go here
  </script>
</body>
</html>

In Node.js

To use TensorFlow.js in a Node.js environment, install the library using npm:

npm install @tensorflow/tfjs-node

Then, require the library in your JavaScript file:

const tf = require('@tensorflow/tfjs-node');

Basic TensorFlow.js Concepts

Tensors

Tensors are the core data structure in TensorFlow.js. They are multi-dimensional arrays used to represent data.

Creating Tensors

You can create tensors using the tf.tensor function:

const tf = require('@tensorflow/tfjs');

// Create a 1D tensor
const tensor1d = tf.tensor([1, 2, 3, 4]);

// Create a 2D tensor
const tensor2d = tf.tensor([[1, 2], [3, 4]]);

Tensor Operations

TensorFlow.js provides various operations to manipulate tensors:

// Element-wise addition
const a = tf.tensor([1, 2, 3]);
const b = tf.tensor([4, 5, 6]);
const c = a.add(b); // [5, 7, 9]

// Matrix multiplication
const d = tf.tensor([[1, 2], [3, 4]]);
const e = tf.tensor([[5, 6], [7, 8]]);
const f = d.matMul(e); // [[19, 22], [43, 50]]

Models and Layers

TensorFlow.js allows you to build and train models using a high-level API similar to Keras.

Creating a Sequential Model

A sequential model is a linear stack of layers:

const model = tf.sequential();

// Add a dense layer with 10 units and input shape of [2]
model.add(tf.layers.dense({units: 10, inputShape: [2]}));

// Add another dense layer with 1 unit
model.add(tf.layers.dense({units: 1}));

Compiling the Model

Before training, compile the model with a loss function and optimizer:

model.compile({
  loss: 'meanSquaredError',
  optimizer: 'sgd'
});

Training the Model

Train the model using the fit method:

const xs = tf.tensor2d([[1, 2], [3, 4], [5, 6], [7, 8]]);
const ys = tf.tensor2d([[1], [2], [3], [4]]);

model.fit(xs, ys, {
  epochs: 10
}).then(() => {
  // Use the model for predictions
  model.predict(tf.tensor2d([[9, 10]])).print();
});

Practical Example: Linear Regression

Let's build a simple linear regression model using TensorFlow.js.

Step 1: Create the Model

const model = tf.sequential();
model.add(tf.layers.dense({units: 1, inputShape: [1]}));

Step 2: Compile the Model

model.compile({
  loss: 'meanSquaredError',
  optimizer: 'sgd'
});

Step 3: Generate Training Data

const xs = tf.tensor2d([1, 2, 3, 4], [4, 1]);
const ys = tf.tensor2d([1, 3, 5, 7], [4, 1]);

Step 4: Train the Model

model.fit(xs, ys, {epochs: 500}).then(() => {
  // Make a prediction
  model.predict(tf.tensor2d([5], [1, 1])).print(); // Should output a value close to 9
});

Exercises

Exercise 1: Create a Tensor

Create a 3x3 tensor with values from 1 to 9.

Solution:

const tensor3x3 = tf.tensor([[1, 2, 3], [4, 5, 6], [7, 8, 9]]);
tensor3x3.print();

Exercise 2: Build and Train a Model

Build and train a model to predict the output for the input [6] using the following data:

  • Inputs: [1, 2, 3, 4, 5]
  • Outputs: [2, 4, 6, 8, 10]

Solution:

const model = tf.sequential();
model.add(tf.layers.dense({units: 1, inputShape: [1]}));

model.compile({
  loss: 'meanSquaredError',
  optimizer: 'sgd'
});

const xs = tf.tensor2d([1, 2, 3, 4, 5], [5, 1]);
const ys = tf.tensor2d([2, 4, 6, 8, 10], [5, 1]);

model.fit(xs, ys, {epochs: 500}).then(() => {
  model.predict(tf.tensor2d([6], [1, 1])).print(); // Should output a value close to 12
});

Conclusion

In this section, we introduced TensorFlow.js, set up the environment, and explored basic concepts such as tensors, models, and layers. We also built a simple linear regression model to demonstrate how to use TensorFlow.js in practice. With these fundamentals, you can start building and deploying machine learning models in your web applications. In the next module, we will delve into more advanced topics and techniques in TensorFlow.js.

◄ Previous
◄
Next ►
►
The Project | About Us | Contribute | Donations | License
© Copyright 2024. All rights reserved