April 2020
Beginner to intermediate
156 pages
4h 47m
English
In this tutorial, we will learn how to create a matching networks architecture and train it on the Omniglot dataset. To begin, let's first understand what the Omniglot dataset is.
The Omniglot dataset is designed for developing more human-like learning algorithms. It contains 1,623 different handwritten characters from 50 different alphabets. Each of the 1,623 characters was drawn online via Amazon's Mechanical Turk by 20 different people. Each image is paired with stroke data, a sequence of [x,y,t] coordinates with time (t) in milliseconds. For more details, please refer to https://github.com/brendenlake/omniglot.
Read now
Unlock full access