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 should not be too hard now that we already support python sklearn.
The starting point could be to support the models from the tensorflow.contrib.learn package (this was the separate skflow project previously: https://github.com/tensorflow/skflow).
So far implemented support for Keras.
Tensorflow support should be possible in very much the same way, though using Tensorflow through Keras is much more convenient for our purposes.
Leaving open for now.
This should not be too hard now that we already support python sklearn.
The starting point could be to support the models from the tensorflow.contrib.learn package (this was the separate skflow project previously: https://github.com/tensorflow/skflow).
Also, consider to alternately/in addition support using tensorflow based on TensorFlow Serving: https://www.tensorflow.org/versions/r0.9/tutorials/tfserve/index.html