Hyperopt 0.2.1 includes distributed tuning via Apache Spark

databricks.com

Hyperopt is an hyperparameter optimization library in Python. This new release now includes a SparkTrials class that allow you to scale out a hyperparameter tuning across a Spark cluster. This obviously covers random search, but the Tree of Parzen Estimators (TPE) is also included, which is no easy feat considering TPE is a Sequential Model-Based Optimization. Looking forward to give it a try.

Read more...
Linkedin

Want to receive more content like this in your inbox?