Skip to content

Releases: nreimers/truecaser

v1.0

09 Jun 13:46
Compare
Choose a tag to compare

Release v1.0

Attached is a distributions.obj file for English achieving an accuracy of 98.39%. This file was trained on:

  • Brown Corpus
  • Reuters Corpus
  • Semcor Corpus
  • Conll2000 Corpus
  • State_union Corpus
  • TechCrunch articles (~150.000 articles)
  • Most frequents n-grams of the Corpus of Contemporary American English (COCA)