@tensorflow/tfjs-backend-webgl

What is @tensorflow/tfjs-backend-webgl?

@tensorflow/tfjs-backend-webgl is a WebGL-accelerated backend for TensorFlow.js, enabling high-performance machine learning computations in the browser. It leverages the power of the GPU to perform operations faster than the CPU backend.

What are @tensorflow/tfjs-backend-webgl's main functionalities?

Tensor Operations

This feature allows you to perform tensor operations such as addition, multiplication, etc., using the WebGL backend for accelerated performance.

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

async function run() {
  await tf.setBackend('webgl');
  const a = tf.tensor([1, 2, 3, 4]);
  const b = tf.tensor([5, 6, 7, 8]);
  const c = a.add(b);
  c.print(); // Output: [6, 8, 10, 12]
}
run();

Model Training

This feature allows you to train machine learning models directly in the browser using the WebGL backend for faster computations.

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

async function run() {
  await tf.setBackend('webgl');
  const model = tf.sequential();
  model.add(tf.layers.dense({units: 100, activation: 'relu', inputShape: [10]}));
  model.add(tf.layers.dense({units: 1}));
  model.compile({optimizer: 'sgd', loss: 'meanSquaredError'});

  const xs = tf.randomNormal([100, 10]);
  const ys = tf.randomNormal([100, 1]);

  await model.fit(xs, ys, {epochs: 10});
  console.log('Model training complete');
}
run();

Model Inference

This feature allows you to perform model inference, i.e., making predictions using a pre-trained model, with the WebGL backend for improved performance.

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

async function run() {
  await tf.setBackend('webgl');
  const model = await tf.loadLayersModel('https://example.com/model.json');
  const input = tf.tensor([1, 2, 3, 4], [1, 4]);
  const output = model.predict(input);
  output.print();
}
run();

Other packages similar to @tensorflow/tfjs-backend-webgl

@tensorflow/tfjs-backend-cpu

@tensorflow/tfjs-backend-cpu is a CPU-based backend for TensorFlow.js. It is generally slower than the WebGL backend but can be used in environments where WebGL is not available or desired.

@tensorflow/tfjs-node

@tensorflow/tfjs-node is a Node.js backend for TensorFlow.js, providing high-performance machine learning computations on the server-side. It leverages the TensorFlow C library for accelerated performance.

gpu.js

gpu.js is a JavaScript library for GPU-accelerated computations. While it is not specifically designed for machine learning, it can be used to perform general-purpose computations on the GPU, similar to the WebGL backend of TensorFlow.js.

README

Usage

This package implements a GPU accelerated WebGL backend for TensorFlow.js.

Importing the backend

Note: this backend is included by default in @tensorflow/tfjs.

Via NPM

// Import @tensorflow/tfjs-core
import * as tf from '@tensorflow/tfjs-core';
// Adds the WebGL backend to the global backend registry.
import '@tensorflow/tfjs-backend-webgl';

Via a script tag

<!-- Import @tensorflow/tfjs-core -->
<script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs-core"></script>

<!-- Adds the WebGL backend to the global backend registry -->
<script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs-backend-webgl"></script>

You can also get ES2017 code using the following links

<!-- Import @tensorflow/tfjs-core -->
<script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs-core@2.0.0-rc.4/dist/tf-core.es2017.js"></script>

<!-- Adds the WebGL backend to the global backend registry -->
<script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs-backend-webgl@2.0.0-rc.4/dist/tf-backend-webgl.es2017.js"></script>