fastText Quick Start Guide: Get started with Facebook's library for text representation and classification

In today's interconnected world, a lot of text data gets generated around the world. This text information includes descriptions of things. Take, for example, people writing about products in Amazon reviews, or people writing about their thoughts through their Facebook posts. Natural Language Processing (NLP) is the application of machine learning and other computational techniques to understanding and representating spoken and written text. The following are the major challenges that NLP seeks to solve:

The previous list has many of the problems that NLP practitioners are targeting. Depending on the use case, you can pick up any of these challenges and try to solve them in your domain. The challenge with many previous approaches and modeling techniques is that NLP requires a lot of textual data and there is a lot of contextual information in the data. It is quite hard for a computational model to get a sense of all the data in an efficient manner. NLP models up to now have only targeted English as textual data is available in English. But only 20 percent of the population of the world speak English and even among them, the majority are non-native speakers. The biggest deterrent to building non-English NLP models is the lack of data. Hence, we desperately need libraries that can build models even when the data is limited. fastText has the potential to change all that. The fastText team has published pretrained word vectors for 294 languages. By the time the book is published, more languages will have been added to it.

In this chapter, we will see how to install fastText so that you can start tinkering with this amazing software. 

Depending on your operating system, you will need to make sure that you have some dependencies installed in your machine. In this section, you will get to know how to install fastText based on whether you are using a Linux, Windows, or macOS operating system. Additionally, you will get to know what additional dependencies you should install depending on your usage. My recommendation is to install all the software packages, as we will be exploring all the various ways we can use fastText in this book.

$ sudo yum install gcc-c++

$ sudo yum install cmake

$ sudo yum install zip docker git-core

Failed to set locale, defaulting to C

$ sudo yum install glibc-langpack-en

$ sudo apt update
$ sudo apt install build-essential

$ sudo apt install cmake

$ sudo apt install zip docker git-core

$ sudo pacman -S cmake make gcc-multilib

$ sudo pacman -S p7zip git docker

On macOS, you should have Clang installed by default, which is designed to be a drop-in replacement for the normal compilers for C, C++, and other similar languages. Check whether the version is 3.3 or later using clang --version in a Terminal. If you do not have Clang or something from the older versions, then you can install using the xcode command-line tools using a Terminal:

$ xcode-select --install

A dialog should appear next that asks if you want to install the developer tools. Click on the Install button.

I recommend that you install Anaconda so that there are no issues with installing Python and using it for fastText. Detailed instructions for installing Anaconda are given on the official documentation page, which can be accessed at https://conda.io/docs/user-guide/install/linux.html. Simply stated, if you are on Windows, then download the Windows installer, double-click on it, and then follow the instructions on the screen. Installing it using a GUI is also possible for macOS.

In the case of Linux and macOS, download the corresponding bash file and then run the following command in a Terminal:

$ bash downloadedfile.sh

Please take care to download and install it using installers that are tagged for Python 3.x. The Python code snippets that will be shown in this book will be shown for Python 3.x.

Installing fastText on Windows

Currently, official binaries are not provided for fastText on Windows, and hence there is no GUI to install fastText on your machine. To use fastText, you will need to perform the following steps:

1. Download the latest binary named fasttext-win64-latest-Release.zip from the release page provided by Xua (https://github.com/xiamx/fastText/releases).
2. This is a ZIP file and hence you will need to extract the contents. You will find the fasttext_pic.lib, fasttext.lib, fasttext.exe, and fasttext.dll files in the extracted folder. This folder will be your working directory for fastText:

 

3. Create a folder, data where you will keep all your data files. Now, open PowerShell and change directory to the folder.
4. Type .\fasttext.exe in Powershell and you should be able to see the output.

If you don't see any output at the end, then you probably don't have Visual C++ Redistributable in your machine and will need to install that.

 

 

 $ git clone https://github.com/facebookresearch/fastText.git
 $ cd fastText
 $ mkdir build && cd build && cmake ..
 $ make && make install

$ pip install .

You can also use Docker to run fastText on your machine and not worry about building it. This can be done to maintain version control between specific versions and thus gives us predictability and consistency. You can get information on how to install Docker from the following link: https://docs.docker.com/install/#cloud.

After installing, start the Docker service before running the following commands:

 start the docker service.
 $ systemctl start docker

 # run the below commands to start the fasttext container.
 $ docker pull xebxeb/fasttext-docker

You should now be able to run fastText:

$ mkdir -p /tmp/data && mkdir -p /tmp/result
$ docker run --rm -v /tmp/data:/data -v /tmp/result:/result \
         -it xebxeb/fasttext-docker ./classification-example.sh

You may need to provide permissions and create the specific directories to run the docker run command.

In this chapter, you have taken a look at how to install and start using fastText in the environment of your choice.

In the next chapter, we will be taking a look at how to train fastText models using the command line and how to use them.