* Fix token scoring

This commit is contained in:
Matthew Honnibal 2015-06-28 06:22:18 +02:00
parent 0c4b5a2bb0
commit ef97b90833
1 changed files with 1 additions and 1 deletions

View File

@ -230,7 +230,7 @@ def main(train_loc, dev_loc, model_dir, n_sents=0, n_iter=15, out_loc="", verbos
scorer = evaluate(English, list(read_json_file(dev_loc)),
model_dir, gold_preproc=gold_preproc, verbose=verbose,
beam_width=beam_width)
print 'TOK', 100-scorer.token_acc
print 'TOK', scorer.token_acc
print 'POS', scorer.tags_acc
print 'UAS', scorer.uas
print 'LAS', scorer.las