Hello, we are working for some time now on rebuilding the most important components using TensorFlow, these usually perform quite a bit better than what we have already, but they are also harder to use.
All this code is currently in the sandbox, and I think we should start to move it over to the main project and release it somehow, maybe first as experimental. It would be nice if more people start to use it. What we build so far is: - Name Finder - Normalizer, this is a lexical normalizer, we can probably ship it with models to normalize dates - Namecat, this is a component which can categorize a name, e.g. into gender (for that we should also be able to ship models) Right now we are still missing doccat and a new POS Tagger. Jörn On Mon, Apr 15, 2019 at 4:58 PM Damiano Porta <damianopo...@gmail.com> wrote: > > Greetings to all! > in spite of myself that the library is going very slow about development. I > would like to understand if it is your intention to follow the development > perhaps integrating more advanced solutions like neural networks or not. > I have been using OpenNLP for a long time, but the advancement of other > frameworks such as Spacy and Allennlp in python language is making me think > of a change. > I hope to hear good news. Thank you > Damiano
