The third assignment of NLU which is for Named Entity Recognition
pyner uses the pycrf suite. neuralcrf uses keras contrib CRF model.
These will have to be installed for the programs to run
For pycrfsuite: pip install python-crfsuite
for keras contrib : sudo pip install git+https://www.github.com/keras-team/keras-contrib.git
Both codes take it in the ner tagged input file named "ner.txt" and gives as output the test file which (which contains 10% of the sentences in ner.txt)contains each individual token in each line along with their corresponding true and predicted label respectively.