Bring your neural networks to the browser with TF.js - Simone Scardapane
Download as PPTX, PDF0 likes298 views
This document introduces TensorFlow.js, which brings neural networks to web browsers using JavaScript. TensorFlow.js allows training and importing models and includes features like tensors, automatic differentiation, layers and models, and a simplified Keras API. It demonstrates max prediction using an RNN model and provides resources for learning more, including tutorials, examples, and interactive visualizations. The goal is to make machine learning more accessible and useful by bringing it to web applications and browsers through TensorFlow.js.
![Tensors
// A 3x2 matrix of random numbers
const x = tf.randomUniform([3, 2]);
// Convert from Javascript arrays
const y = tf.tensor1d([0.4, -0.2, 1.0]);
// Matrix operations
const z = x.transpose().matMul(y.reshape([3, 1]));](https://image.slidesharecdn.com/meetup-190110161731/85/Bring-your-neural-networks-to-the-browser-with-TF-js-Simone-Scardapane-19-320.jpg)
![Tidy can help you the code
const sum = tf.tidy(() => {
const a1 = tf.randomUniform([3, 2]);
const a2 = tf.randomNormal([3, 2]);
return a1.square().add(a2);
});](https://image.slidesharecdn.com/meetup-190110161731/85/Bring-your-neural-networks-to-the-browser-with-TF-js-Simone-Scardapane-23-320.jpg)
Bring your neural networks to the browser with TF.js - Simone Scardapane
- 1. Bring your neural networks to the browser with TF.js!
- 2. Something about myself Research fellow @ Sapienza researching on neural networks Strong passion for promoting ML in all its aspects
- 3. Not-so-easy question: what is TensorFlow?
- 4. TensorFlow is all about: 1. Neural networks 2. Python 3. Huge data & infrastructures
- 5. TensorFlow has grown impressively
- 6. TensorFlow Distributed Execution Engine CPU GPU TPU Mobile Embedded Neural Nets C++ K-Means Loss Functions, Metrics Linear Algebra Decision Trees SVM Gaussian Mixture Models Regression Random Forests Probabilistic Methods Signal Processing Lattice TensorFlow is a comprehensive ML framework Python Java Javascript
- 8. 1. No installation required 2. Privacy (GDPR, anyone?) 3. Sensors! (Lots of) 4. WebGL 5. The world runs on apps and web... Why TF on browsers?
- 9. 1. Train a model from scratch 2. Fine-tune a model 3. Import a trained model 3 use cases COMPLEXITY
- 14. TensorFlow.js
- 15. Core API (ex deeplearn.js) TensorFlow SavedModel Layers API Keras models Browser (WebGL) Node.js: everywhere! (GPUs, TPUs, …)
- 16. Inference Time (ms) of MobileNet 1.0_224 Average of 200 runs
- 17. Core concepts
- 18. Import TF.js in your page <html> <head> <!-- Load TensorFlow.js --> <script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs@0.13.3/dist/tf.min.js" /> <script> <!-- Code here! --> ... </script> </head> ... </html>
- 19. Tensors // A 3x2 matrix of random numbers const x = tf.randomUniform([3, 2]); // Convert from Javascript arrays const y = tf.tensor1d([0.4, -0.2, 1.0]); // Matrix operations const z = x.transpose().matMul(y.reshape([3, 1]));
- 20. Auto-differencing // Function definition const axis = 1; const f = x => tf.sin(x).sum(axis); // Automatic gradient computation const fGrad = tf.grad(f);
- 21. Models and layers // Initialize the model const model = tf.sequential(); // Add layers model.add(tf.layers.dense({'units': 10, 'inputShape': 4})); model.add(tf.layers.dense({'units': 1})); // Define loss const loss = tf.losses.meanSquaredError // Optimize! const opt = tf.train.sgd({'lr': 0.1}); opt.minimize(() => loss(model.predict(xs), ys));
- 22. Simplified Keras API model.compile({'optimizer': 'sgd', loss: 'categoricalCrossentropy'}); await model.fit({x: xs, y: ys}); The training step is an async function!
- 23. Tidy can help you the code const sum = tf.tidy(() => { const a1 = tf.randomUniform([3, 2]); const a2 = tf.randomNormal([3, 2]); return a1.square().add(a2); });
- 24. Demo time!
- 25. ● Input: a (fixed-length) sequence. ● Output: the index of the largest element. Please don’t do this in real life. :-( Let’s learn… MAX!
- 26. Our RNN 0.14 0.81 0.64 0.0 LSTM 2% 95% 2.5% 0.5%
- 28. Where to from here?
- 29. Some resources 1. TensorFlow.js tutorials: https://js.tensorflow.org/tutorials/ 2. More examples: https://github.com/tensorflow/tfjs-examples/
- 30. Thanks for listening! «The future will be intensely data-driven. Creating a dynamic hub that brings together professionals, industries, and academics will be essential to achieve this vision.» http://www.iaml.it/member