opennlp-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Damiano Porta <damianopo...@gmail.com>
Subject Why can i not serialize a Dictionary ?
Date Tue, 25 Oct 2016 16:49:45 GMT
Hello,

i am getting a strange error during the compiling of a NER model.
Basically, the end of the build output is:

 98:  ... loglikelihood=-13340.018762351776 0.999005934601099
 99:  ... loglikelihood=-13258.358751926637 0.9990120681028991
100:  ... loglikelihood=-13178.039964721707 0.9990177634974279
Exception in thread "main" java.lang.IllegalStateException: Missing
serializer for *mydictionary*
at opennlp.tools.util.model.BaseModel.serialize(BaseModel.java:610)

Where "mydictionary" is the key of my custom dictionary (that i add on the
resources)

I load the dictionary with this code:

HashMap<String, Object> map = new HashMap<>();

try (InputStream modelIn = new
FileInputStream("/home/damiano/mydictionary.xml")) {
Dictionary dictionary = new Dictionary(modelIn);
map.put("mydictionary", dictionary);
}
catch(Exception e) {
System.out.println(e.getMessage());
}

Then i create the TokenNameFinderFactory with:

TokenNameFinderFactory factory = new TokenNameFinderFactory(
IOUtils.toByteArray(in),
map,
new BioCodec()
);

where "*in"* is the InputStream of my featuregenerators.xml

and map is the Map with my resources.

then train the model with:

TokenNameFinderModel model = NameFinderME.train("it", "person",
sampleStream, TrainingParameters.defaultParams(), factory);

at the end i serialize the model doing:

try (BufferedOutputStream modelOut = new BufferedOutputStream(new
FileOutputStream(outputFile.toFile()))) {
  model.serialize(modelOut);
}

As i wrote at the beginning it cant serialize my dictionary so the build of
the model does not finish properly.

I read that a Dictionary can be serialized, so why this error?
Thank you!

Damiano

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message