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.
the engine figures out where python is, not the script
the engine uses known default locations or tries to get the location from a yaml config file
the python location is passed to the initial script as an additional parameter or by setting an environment variable (preferably)
Ideally also add something to standard engine procedure to call a checking-script BEFORE processing starts, this should avoid going through the whole corpus only to find out that we cannot continue.
(see https://github.com/GateNLP/gate-lf-planning/issues/10)
If possible, make it work under Windows.
For this, consider the following approach:
the engine figures out where python is, not the script the engine uses known default locations or tries to get the location from a yaml config file the python location is passed to the initial script as an additional parameter or by setting an environment variable (preferably)
Ideally also add something to standard engine procedure to call a checking-script BEFORE processing starts, this should avoid going through the whole corpus only to find out that we cannot continue.