This article aims at my personal review for CRF applications.
First priority: Java
- CRF/original, IIT bombay: this seems original implementation of CRF in Java,,, there is no info regard to training data or example usage...
- Sequence Tagging: I haven't tried though, it's got an introduction and looks nice!
- CRF/ahser-stern: quite similar to CRF/original, maven project.... but again, there is no demo or some tutorial stuff...
- CRF/witgo: compared to others, it has less stars.... there is an introduction though,, "couldn't find main class" error caught me.
- python-crf-named-entity-recognition: the author assumes collects data from the net, like scraping. so this doesn't match my objective.
- NER with DNNs: this looks good, I am trying right now. he is following a Hybrid Bi-LSTM-CRF. but he tested in Russian, so I need to evaluate the model in Japanese.
- Sequence-Tagging: this guy has nice blog regarding to Sequence-tagging with Tensorflow. So I would like to read that and try demo as well.
- tagger/glample: I need to try this maybe, this has nice amount of stars.