yifanhu / yamall

yamall

Yet Another MAchine Learning Library AKA Yahoo's Amazing MAchine Learning Library

What is yamall?

yamall is a Java machine learning library, a fast black-box local learner, and a hadoop implementation. It implements most of the state-of-the-art features used in machine learning algorithms, e.g. namespaces, hashing of the features, single and multipass stochastic gradient descent.

Why yamall?

yamall comes from the necessity to have a secure Java implementation of state-of-the-art machine learning algorithms. The local version tries to imitate the interface of Vowpal Wabbit, in order to facilitate the migration. At the same time, to harness the full power of yamall you can directly call the java functions in your code. Also, there is a Hadoop version to be able to scale to big datasets.

With very few lines of code you will be able to

• Parse VW and Tab Separated Value files.
• Instante a Stochastic Gradient Descent, choose optimization algorithm and loss function.
• Run multiple epochs on a dataset.
• Train your model locally and deploy it on the grid.

Build instructions

yamall uses Maven as the build system. You will need to install Maven on your machine. The installation steps depend on your operating system. On Linux, you can use your package manager; e.g. on Fedora, run sudo yum install maven.

Once Maven is installed, you can use Maven to build it:

cd yamall/
mvn package

Running locally

See the directory 'local'.

Running on Hadoop

See the directory 'hadoop'.

Examples on how to use the yamall APIs

See the directory 'examples'.

License

The use and distribution terms for this software are covered by the Apache 2.0 license. See LICENSE file for terms.