Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Tensorflowjs 💝 AI

Yas
November 09, 2019

Tensorflowjs 💝 AI

Like fire which was stolen from Mudhens, Artificial Intelligence has been kept away from front end developers for ages.

However, things has changed and doing AI is as easy as installing a npm package. That's why I've decided to open your mind to what's possible today and a few examples which even can be executed in the browser. You don't need to have any background in AI or Maths, just a curious mind and a willingness to try.

This talk was inspired by some recent projects which was shown during Google IO and the possibilities which TensorFlow.js opens up for businesses. Now that you can run machine learning even on your mobile, it is very important to show front end developers of what is possible and how to get started in under 30 mins. From converting a Keras model and use it for prediction to create a neural network from scratch, train it and use it in the browser it is all a matter of minutes only.

Yas

November 09, 2019
Tweet

More Decks by Yas

Other Decks in Programming

Transcript

  1. March 2018 GPU accelerated via WebGL Inference and training right

    within the browser Tensorflow.js Release How Wow
  2. TF TPU TF GPU Core API WebGL Layers API Model

    Browser Node.js TF CPU Keras Model Tensorflow Saved models
  3. Pre-trained models Object detection Mobilenet Human pose detection Posenet Human

    segmentation BodyPix Voice recognition Speech command
  4. CREDITS: This presentation template was created by Slidesgo, including icons

    by Flaticon, and infographics & images by Freepik. THANKS! Questions? [email protected] @yashints