What's new in OpenNLP 1.8.2
Oct 6, 2017
- Fix checksum in SourceForgeModelEval.verifyTrainingData
- Optimize XML Parser configuration
- Leipzig sample should allow skip initial entries
- Leipzig sample stream should shuffle data
- Update OpenNLP Release Notes
- Consolidate the two README files into README.md
- Change misclassified report layout for LanguageDetector
- Add test data verification to tests under opennlp.tools.eval
- Fix cmd line training time
- Add Concatenate Stream method for Collections of streams
New in OpenNLP 1.7.2 (Mar 31, 2017)
- The API has been improved for a better consistency
- 1.4 deprecated methods were removed
- Now Java 1.8 is required
- Name Finder evaluation can now show a confusion matrix
- The default evaluation output contains more details
- Added a Language Model CLI tool
- Add Moses format support
- More refactoring and cleanup, specially in Machine Learning package and Dictionary
- Removed deprecated trainers from UIMA integration
- Fixed potential localization issues and added maven plugin to prevent it (ForbiddenAPI)
- Fixed issues with the BRAT corpus reader
- Deprecated GIS class, will be removed in a future 1.8.x release
New in OpenNLP 1.6.0 (Jul 17, 2015)
- This release introduces many new features, improvements and bug fixes. The API
- has been improved for a better consistency and 1.4 deprecated methods were
- removed. Now Java 1.7 is required.
- Additionally the release contains the following noteworthy changes:
- Added evalutation support to the parser and doccat components
- Added support to Evalita 07/09, Brat and OntoNotes corpus formats
- Now L-BFGS is stable
- Added Snowball to the Stemmer package
- NameFinder now supports a user defined factory
- Added pluggable machine learning support
- Added a lemmatizer module
- Added Cluster, Document Begin and Clark feature generators to the Name Finder
- Added Liblinear as a Machine Learning addon
- Entity Linker now has a command line interface
- Added sequence classification support
New in OpenNLP 1.5.3 (Apr 18, 2013)
- The CLI has been improved for a better consistency. Now the tools supports extensions that can be configured from the model, including customized context generators and validators.
- Additionally the release contains the following noteworthy changes:
- Porter Stemmer tool
- L-BFGS parameter estimation
- Improved documentation
- Fine-grained POSTagger evaluation report
- Improved support to load user provided feature generator and context validation classes from OSGi environment
- A detailed list of the issues related to this release can be found in the release notes.
New in OpenNLP 1.5.2 (Apr 16, 2013)
- Improved the white space handling in the Sentence Detector and its training code
- Added more cross validator command line tools
- Command line handling code has been refactored
- Fixed problems with the new build
- Now uses fast token class feature generation code by default
- Added support for BioNLP/NLPBA 2004 shared task data
- Removal of old and deprecated code
- Dictionary case sensitivity support is now done properly
- Support for OSGi