Vowpal Wabbit Wrapper

Vowpal wabbit is a very effective machine learning library (https://github.com/JohnLangford/vowpal_wabbit/wiki). It implements several state of art algorithms in a very efficient way. This java library provides a JNI OOP interface to vowpal wabbit. It depends only on log4j and guava. You don't need vowpal wabbit to be installed if you use this wrapper.

License	The Apache Software License, Version 2.0
GroupId	com.indeed
ArtifactId	vw-wrapper
Last Version	5.1.6
Release Date	Sep 11, 2018
Type	jar
Description	Vowpal Wabbit Wrapper Vowpal wabbit is a very effective machine learning library (https://github.com/JohnLangford/vowpal_wabbit/wiki). It implements several state of art algorithms in a very efficient way. This java library provides a JNI OOP interface to vowpal wabbit. It depends only on log4j and guava. You don't need vowpal wabbit to be installed if you use this wrapper.
Project URL	http://github.com/indeedeng
Source Code Management	https://github.com/indeedeng/vowpal-wabbit-java

Download vw-wrapper

Filename	Size
vw-wrapper-5.1.6.pom
vw-wrapper-5.1.6.jar	2 MB
vw-wrapper-5.1.6-sources.jar	2 MB
vw-wrapper-5.1.6-javadoc.jar	211 KB

How to add to project

Apache Maven

<!-- https://jarcasting.com/artifacts/com.indeed/vw-wrapper/ -->
<dependency>
    <groupId>com.indeed</groupId>
    <artifactId>vw-wrapper</artifactId>
    <version>5.1.6</version>
</dependency>

Gradle Groovy

// https://jarcasting.com/artifacts/com.indeed/vw-wrapper/
implementation 'com.indeed:vw-wrapper:5.1.6'

Gradle Kotlin

// https://jarcasting.com/artifacts/com.indeed/vw-wrapper/
implementation ("com.indeed:vw-wrapper:5.1.6")

Apache Buildr

'com.indeed:vw-wrapper:jar:5.1.6'

Apache Ivy

<dependency org="com.indeed" name="vw-wrapper" rev="5.1.6">
  <artifact name="vw-wrapper" type="jar" />
</dependency>

Groovy Grape

@Grapes(
@Grab(group='com.indeed', module='vw-wrapper', version='5.1.6')
)

Scala SBT

libraryDependencies += "com.indeed" % "vw-wrapper" % "5.1.6"

Leiningen

[com.indeed/vw-wrapper "5.1.6"]

Dependencies

compile (2)

Group / Artifact	Type	Version
com.google.guava : guava	jar	16.0.1
log4j : log4j	jar	1.2.17

provided (1)

Group / Artifact	Type	Version
com.google.code.findbugs : jsr305	jar	1.3.9

test (2)

Group / Artifact	Type	Version
junit : junit-dep	jar	4.8.2
commons-io : commons-io	jar	1.4

Project Modules

There are no modules declared in this project.

Vowpal Wabbit Wrapper

A java wrapper and friendly API for the Vowpal wabbit machine learning package.

Background

The Vowpal Wabbit (VW) package is very production friendly and it implements the state of the art in modern machine learning research.

The existing java binding for VW has drawbacks:

• Because of a bug, saved models may have incorrect weights
• It requires the boost library to be installed on every machine where the wrapper is used, which is not always feasible
• its API is low-level, requiring you to operate with strings instead of providing a more convenient domain abstraction

This project addresses these drawbacks.

Building the library

To build this library run the following command:

mvn clean install

Dependencies

• guava
• log4j

Rebuilding C++ binaries

This distribution includes pre-built C++ binaries along with the code. You can rebuild the binaries from source if necessary. Refer to build-jni/README.md for instructions.

Using the library

Refer to the official vowpal wabbit wiki for general instructions and advice on training the Vowpal Wabbit model.

Refer to API javadocs for instructions specific to this wrapper and java API.

The following integration tests provide references for using the API.

• Twitter Sentiment analysis
• Movie lens 1M

Reporting issues

Create an issue in this project if you encounter issues or need help.

Tested platform

We have tested this wrapper on the following platforms:

• OS X Yosemite
• Ubuntu 14
• Enterprise Linux 5
• Enterprise Linux 6
• CentOS 5
• CentOS 6
• CentOS 7

Code of Conduct

This project is governed by the Contributor Covenant v 1.4.1

License

• This library is distributed under The Apache Software License, Version 2.0.
• VW binaries are distributed under BSD (revised) license

Indeed Engineering

Versions

Version
5.1.6 Sep 11, 2018
5.1.3 Feb 22, 2018
5.1.2 Feb 9, 2018
5.1.1 Feb 9, 2018
1.0.2 Oct 2, 2017
1.0.1 Aug 2, 2017
1.0.0 May 18, 2017