NERsuite

Pre-requisite Libraries

You need a numerical optimization library and a CRF toolkit. NERsuite uses the following two libraries which are released by Naoaki Okazaki. Please download the latest source packages of these two libraries. For installing these two libraries, please click Installation Guide menu.

• libLBFGS - a library of L-BFGS
• CRFsuite - a fast implementation of Conditional Random Fields

NERsuite

The source package of the NERsuite can be downloaded in the following link. For installing NERsuite, please click Installation Guide menu.

• NERsuite ver. 1.2 - a named entity recognition toolkit

Pre-trained Models

Here, we provide pre-trained models for the GENIA tagger and the NERsuite. The model files for the GENIA tagger are same to the original ones. The pre-trained models for NERsuite are for the tasks such as BioCreative 2 Gene Mention Recognition (BC2GMR) task and NLPBA 2004 Named Entity Recognition (NLPBA04) task.

• The models for the GENIA tagger, 18MB
• A model for the BC2GMR (IOB2 tag-set, no dictionary features), 97MB
• A model for the NLPBA04 (IOB2 tag-set, no dictionary features), 83MB