search

GateNLP / gateplugin-LearningFramework

A plugin for the GATE language technology framework for training and using machine learning models. Currently supports Mallet (MaxEnt, NaiveBayes, CRF and others), LibSVM, Scikit-Learn, Weka, and DNNs through Pytorch and Keras.
https://gatenlp.github.io/gateplugin-LearningFramework/
GNU Lesser General Public License v2.1
26 stars 6 forks source link

Chunking: make sure the temporary annotation set is cleared before using it #116

Closed johann-petrak closed 4 years ago

johann-petrak commented 4 years ago

The temporary annotation set LF_SEQ_TMP is not getting cleared before we use it in the LF_ApplyChuking pr, so if multiple application PRs are used the annotations in the set accumulate. This should get cleared before use and possibly also after use (though leaving it after use could help with debugging).

johann-petrak commented 4 years ago

The set now always gets cleared before applying the model and it gets cleared when the apply chunking pr finishes, except when the "debug" parameter is true.