日本語 English

N3ER

N3ER (Neural Network based Named Entity Recognizer) is a NER system, which uses BiLSTM and CRF. N3ER adopts BIO2 (Input, Output, Beginning) format and is implemented by Pytorch.

Download

Source

Model

Requirements

How to Use

We can train a model by using the following command.

  python train.py CONFIG

CONFIG is a configuration file which specifies directories for training and testing and hyper-parameters for the model. A sample file of CONFIG is here.

We can also test the trained model by using the following command.

  python test.py CONFIG -m MODEL

MODEL specifies the directory of the trained model and estimate the tag sequence by viterbi algorithm. If we specify the argument -d with a delimiter, the above command outputs the tag sequence with the sentence by seperating them by the delimiter.

Development

Contributors

Keisuke Shirai(project leader, all coding)
Suzushi Tomori
Shinsuke Mori (oversight, power user)

Revision History

Last Modified: 2021-01-19 by Shirai