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.
If none is specified, use a default feature specification that is just the string of the instance annotation.
This should make many DNN situations much easier where the text is the only thing we need.
In that case, if the user wants to use pretrained embeddings, that could still be something that is handled as an algorithm parameter and the pretrained/fine-tuned embeddings are stored with the model, so we do not need any config or parameter at application time either.
If none is specified, use a default feature specification that is just the string of the instance annotation. This should make many DNN situations much easier where the text is the only thing we need.
In that case, if the user wants to use pretrained embeddings, that could still be something that is handled as an algorithm parameter and the pretrained/fine-tuned embeddings are stored with the model, so we do not need any config or parameter at application time either.