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.
This happens when Python is installed in a non-standard way under Windows, is not on the path or the wrong environment is active when running GATE.
With LearningFramework versions 4.2 and 4.3-SNAPSHOT until 2020-11-04, the Windows script running the python command will terminate but the LearningFramework does not detect this and when training is run, everything looks fine on first glance. However, there will be no output from the Pytorch training which must be there if a model is trained successfully.
Create a file with the name "FileJsonPyTorch.yaml" in the model directory, i.e. the directory that contains the file info.yaml after trying to run train once
Edit the file FileJsonPyTorch.yaml and add a line that specifies the location of the python interpreter for the correct Python environment, e.g. something like:
This happens when Python is installed in a non-standard way under Windows, is not on the path or the wrong environment is active when running GATE.
With LearningFramework versions 4.2 and 4.3-SNAPSHOT until 2020-11-04, the Windows script running the python command will terminate but the LearningFramework does not detect this and when training is run, everything looks fine on first glance. However, there will be no output from the Pytorch training which must be there if a model is trained successfully.
If Python is not on the path with the correct packages installed as described in https://gatenlp.github.io/gateplugin-LearningFramework/Installation it is possible to configure how to run python by doing this:
Note that if you want to enclose the path name in double quotes, the backslashes must be doubled like so: