A software for text segmentation inspired by this paper https://arxiv.org/pdf/1803.09337.pdf.

After training depending on the use-case, it is able to get a stream of text and segment it depending on its semantic context.

It does not contain the training code (I will probably upload soon).

It does not have the Fasttext and Keras modeles as they are trained on a specific usecase.

The project is in Docker in order to make it work on every platform. It has an installer.sh and run-classifier.sh scripts in order to make the installation and running procedure simpler.

If you need more information about how to train the model contact me.