On completion of this section, you will have working knowledge of how fastText goes about creating both supervised and unsupervised models. Additionally, you will have knowledge of the algorithms and the design decisions that went into incorporating those algorithms into fastText.
In the Chapter 3, Word Representations in FastText, you will get to know how unsupervised word embeddings are created in fastText.
In the Chapter 4, Sentence Classification in FastText, you will get to know about the algorithms that power sentence classification in fastText. You will also get to know how fastText is able to compress big models into smaller models that can be deployed to low memory devices.
The aim of this section is to shine a light on the algorithms and design choices in fastText, and give you the knowledge to generate better models or augment the library.
...
