Jayant Krishnamurthy's (Machine) Learning and Optimization Library
A machine learning library with many different kinds of models, such as:
- Graphical models
- Sequence models
- Context-free grammars
- Combinatory Categorial Grammar semantic parsers
These models are built using useful core primitives for learning and inference:
- Optimization -- stochastic gradient, Adagrad, LBFGS, and Expectation Maximization.
- Tensor math -- supports both sparse and dense tensors.
Get jklol by cloning the git repository:
git clone https://github.com/jayantk/jklol.git
The preferred way to build jklol is using sbt. In the root directory, run:
cd jklol
sbt package
This command will compile the source and produce
target/jklol-(version).jar. If you don't have sbt, you can also build jklol using ant:
cd jklol
ant jar
This command will produce jklol.jar in the root directory.
If you wish to use jklol as a library, you can get jklol through the Maven central repository. The group id is "com.jayantkrish.jklol" and the artifact id is "jklol". The latest version can be found by searching the central repository. Maven users can add the following to the dependency section of pom.xml:
<dependency>
<groupId>com.jayantkrish.jklol</groupId>
<artifactId>jklol</artifactId>
<version>1.2</version>
</dependency>
Sbt users can add the following to their library dependencies:
"com.jayantkrish.jklol" % "jklol" % "1.2"