- January 2, 2016
- Vasilis Vryniotis
- . No comments
The new version of Datumbox Machine Learning Framework has been released! Download it now from Github or Maven Central Repository.
What is new?
The main focus of version 0.6.1 is to resolve various bugs, reduce memory consumption and improve speed.
Let’s see in detail the changes of this version:
- Bug Fixes: A minor issue related to Unreleased Resources has been detected and fixed on the Tests of Dataset class. Also a memory leak was detected and patched on the AutoCloseConnector class.
- Improved Memory Footprint: The shutdown hooks are now removed when close() is called; this improves memory usage. Also the TypeInference class has been updated to reduce memory consumption.
- Speed: The TextClassifier class has been refactored and few speed improvements have been released.
- Staying Up-to-date: All dependencies and maven plugins used in the project have been updated to the latest stable versions. Few more details on the most important dependencies: The framework now uses MapDB 1.0.8 and it is within my plans to move to MapDB 2.0 once a stable version is released. Moreover I created a Mavenized version of LIBSVM; we currently use the most updated version which is the 3.21.
As expected the version 0.6.1 is backwards compatible with the version 0.6.0 of the framework.
Next steps & roadmap
The development of Datumbox Framework will continue focusing on version 0.7.0. I plan to introduce couple of changes on the architecture of the framework and further improve the speed and memory footprint of the library. For more information about the roadmap have a look on the previous release announcement.
Don’t forget to download the code of Datumbox v0.6.1 from Github. The library is available also on Maven Central Repository.
I am looking forward to your comments and recommendations. Pull requests are always welcome! 🙂
About Vasilis Vryniotis
My name is Vasilis Vryniotis. I’m a Data Scientist, a Software Engineer, author of Datumbox Machine Learning Framework and a proud geek. Learn more
Source: http://blog.datumbox.com/datumbox-machine-learning-framework-0-6-1-released/
