Introduction

JFastText is a Java wrapper for Facebook's fastText, a library for efficient learning of word embeddings and fast sentence classification. The JNI interface is built using javacpp.

The library provides full fastText's command line interface. It also provides the API for loading trained model from file to do label prediction in memory. Model training and quantization are supported via the command line interface.

JFastText is ideal for building fast text classifiers in Java.

Maven dependency

<dependency>
<groupId>com.github.vinhkhuc</groupId>
<artifactId>jfasttext</artifactId>
<version>0.3</version>
</dependency>

The Jar package on Maven Central is bundled with precompiled fastText library for Windows, Linux and MacOSX 64bit.

Building

C++ compiler (g++ on Mac/Linux or cl.exe on Windows) is required to compile fastText's code.

git clone --recursive https://github.com/lidalei/JFastText.git
cd JFastText
mvn clean package

How to use

Initialization

importcom.github.jfasttext.JFastText;
...
JFastTextjft = newJFastText();

Word embedding learning

jft.runCmd(newString[] {
"skipgram",
"-input", "src/test/resources/data/unlabeled_data.txt",
"-output", "src/test/resources/models/skipgram.model",
"-bucket", "100",
"-minCount", "1"
});

Text classification

// Train supervised modeljft.runCmd(newString[] {
"supervised",
"-input", "src/test/resources/data/labeled_data.txt",
"-output", "src/test/resources/models/supervised.model"
});
// Load model from filejft.loadModel("src/test/resources/models/supervised.model.bin");
// Do label predictionStringtext = "What is the most popular sport in the US ?";
JFastText.ProbLabelprobLabel = jft.predictProba(text);
System.out.printf("\nThe label of '%s' is '%s' with probability %f\n",
text, probLabel.label, Math.exp(probLabel.logProb));
// Unload modeljft.unloadModel();

FastText's command line

$ java -jar target/jfasttext-*-jar-with-dependencies.jar
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
nn query for nearest neighbors
analogies query for analogies

For example:

$ java -jar target/jfasttext-*-jar-with-dependencies.jar quantize -h

License

BSD

References

(From fastText's references)

Please cite 1 if using this code for learning word representations or 2 if using for text classification.

Enriching Word Vectors with Subword Information

[1] P. Bojanowski*, E. Grave*, A. Joulin, T. Mikolov, Enriching Word Vectors with Subword Information

@article{bojanowski2016enriching,
title={Enriching Word Vectors with Subword Information},
author={Bojanowski, Piotr and Grave, Edouard and Joulin, Armand and Mikolov, Tomas},
journal={arXiv preprint arXiv:1607.04606},
year={2016}
}

Bag of Tricks for Efficient Text Classification

[2] A. Joulin, E. Grave, P. Bojanowski, T. Mikolov, Bag of Tricks for Efficient Text Classification

@article{joulin2016bag,
title={Bag of Tricks for Efficient Text Classification},
author={Joulin, Armand and Grave, Edouard and Bojanowski, Piotr and Mikolov, Tomas},
journal={arXiv preprint arXiv:1607.01759},
year={2016}
}

FastText.zip: Compressing text classification models

[3] A. Joulin, E. Grave, P. Bojanowski, M. Douze, H. Jégou, T. Mikolov, FastText.zip: Compressing text classification models

@article{joulin2016fasttext,
title={FastText.zip: Compressing text classification models},
author={Joulin, Armand and Grave, Edouard and Bojanowski, Piotr and Douze, Matthijs and J{\'e}gou, H{\'e}rve and Mikolov, Tomas},
journal={arXiv preprint arXiv:1612.03651},
year={2016}
}

(* These authors contributed equally.)