User Guide

How to run GermaNER:

• From Commandline

  • Download the jar file from here or if you don't have enough memory, use GermaNER without freebase features from [here] (https://github.com/tudarmstadt-lt/GermaNER/releases/download/germaNER0.9.1/GermaNER-nofb-09-09-2015.jar) -> the corresponding configuration file is found here
  • Download the configuration file from here
  • Download a test file here
  • Download a training file here
  • To test a file that was trained on our model, run as follows:
    • java -jar GermaNER-09-09-2015.jar -t TESTFILE -o OUTPUTFILE

where -t TESTFILE is the document to be tagged and -o OUTFILE is name of the tagged output file, However, if you want to test one of your own models, add the -c CONFIGFILE and -d OUTPUTFOLDER which are the configuration file used for training and the directory of the saved GermaNER model directory name respectively.
* To train (and optionally test) a model, run it as follows * java -jar GermaNER-09-09-2015.jar -f TRAINFILE -t TESTFILE -o OUTPUTFILE -c CONFIGFILE -d OUTPUTFOLDER

Where -t TESTFILE, -o OUTPUTFILE, and -c CONFIGFILE are optional parameters. -t TESTFILE is the document to be tagged,-f TRAINFILE is the training data, as explained [here] (https://github.com/tudarmstadt-lt/GermaNER/wiki/File-Format), -o OUTPUTFILE is the tagged output file, and -c CONFIGFILE is the configuration file as explained here and -d OUTPUTFOLDER is the directory where the model will be saved (if-d is not provided, a folder called output will be created and the model will be stored therein.) * To tag your document, the document format should be as it is explained here, run it as follows

• Inside an application

  TODO

• From source

  • Get the source code of GermaNER using git

  git clone https://github.com/tudarmstadt-lt/GermaNER.git

  • Compile GermaNER using the following command

  mvn clean install -Dmaven.test.skip=true

The generated jar will be available under the target folder