Hi,
after the release of the natural language processing library
(which had 219 stars in just 24h ... thanks guys), for who does not want to handle the full NLP stack I have released a pos tagger in golang based on Hidden Markov Model and Viterbi:
The pos tagger is a completely different implementation from go-freeling library and comes with a MIT license. The file hmm.dat contains the trained model. simple.pos contains an example of how a sentence must be provided to the HMMParser. A better model can be created using training API. Documentation on the way.
I hope this helps too. I will release soon also the independent implementation of the golang binding to MITIE library for NER
Antonio