Notes for the Readers

Windows and Linux

We would suggest that you use PowerShell for your windows command line as that is more powerful then simple cmd.

Task	Windows	Linux/macOS
Creating a directory	mkdir	mkdir
Change directory	cd	cd
Move files	move	mv
Unzip files	GUI and double click	unzip
Top of the file	get-content	head
Contents of the file	type	cat
Piping	this pipes objects	this pipes text
Bottom of the file	-wait parameter with get-content	tail

python and perl commands work the same way in windows as they work in bash and hence you can use those files and especially perl one liners in similar way.

Python 2 and Python 3

fastText works for both Python 2 and Python 3. There are few differences though that you should keep in mind for the particular python version.

1. print is a statement in Python 2 and a function in Python 3. This would mean that if you are in a Jupyter notebook and trying to see the changes in a variable you will need to pass the appropriate print statement in the corresponding python version.
2. The fastText handles text as Unicode. Python 3 also handles text as Unicode and hence there is no additional overhead if you code in Python 3. But in case you are developing your models in Python 2, you cannot have your data as a string instance. You will need to have your data as Unicode. Following is an example of text as an instance of the str class and unicode class in Python 2.

>>> text1 = "some text" # this will not work for fastText
>>...

The fastText command line

Following is the list of parameters that you can use with fastText command line:

$ ./fasttext
usage: fasttext <command> <args>

The commands supported by fasttext are:

  supervised train a supervised classifier
  quantize quantize a model to reduce the memory usage
  test evaluate a supervised classifier
  predict predict most likely labels
  predict-prob predict most likely labels with probabilities
  skipgram train a skipgram model
  cbow train a cbow model
  print-word-vectors print word vectors given a trained model
  print-sentence-vectors print sentence vectors given a trained model
  print-ngrams print ngrams given a trained model and word
  nn query for nearest neighbors
  analogies query for analogies
  dump dump arguments,dictionary,input/output vectors

The supervised, skipgram, and cbow commands are for training a model. predict, predict-prob are...

Gensim fastText parameters

Gensim supports the same hyperparameters that are supported in the native implementation of fastText. You should be able to set them as follows:

• sentences: This can be a list of list of tokens. In general, a stream of tokens is recommended, such as LineSentence from the word2vec module, as you have seen earlier. In the Facebook fastText library this is given by the path to the file and is given by the -input parameter.
• sg: Either 1 or 0. 1 means to train a skip-gram model, and 0 means to train a CBOW model. In the Facebook fastText library the equivalent is when you pass the skipgram and cbow arguments.
• size: The dimensions of the word vectors and hence must be an integer. In line with the original implementation, 100 is chosen as default. This is similar to the -dim argument in the Facebook fastText implementation.
• window: The window size that is considered...