Spark-Optim is a library of scalable optimization algorithms based on Spark. It is designed to provide optimization algorithms which are recently published or widely used but have not been implemented in Spark MLlib.
Currently, Spark-Optim provides following implementations.
- Parallel Stochastic Gradient Descent [reference]
- Parallel Follow-the-regularized-leader [reference]
Spark-Optim is built against Spark 2.1.1.
sbt package
Spark-Optim is available under Apache Licenses 2.0.
If you encounter bugs, feel free to submit an issue or pull request. Also you can mail to:
- hibayesian ([email protected]).