Deep Learning with PyTorch

One-hot encoding

In one-hot encoding, each token is represented by a vector of length N, where N is the size of the vocabulary. The vocabulary is the total number of unique words in the document. Let's take a simple sentence and observe how each token would be represented as one-hot encoded vectors. The following is the sentence and its associated token representation:

An apple a day keeps doctor away said the doctor.

One-hot encoding for the preceding sentence can be represented into a tabular format as follows:

An	100000000
apple	010000000
a	001000000
day	000100000
keeps	000010000
doctor	000001000
away	000000100
said	000000010
the	000000001

This table describes the tokens and their ...