Skip to content

Commit

Permalink
bugfix in unigram token count
Browse files Browse the repository at this point in the history
  • Loading branch information
proycon committed Aug 28, 2014
1 parent ac963a7 commit 3debd7d
Show file tree
Hide file tree
Showing 2 changed files with 13 additions and 10 deletions.
9 changes: 6 additions & 3 deletions codereview_input
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@

- on-line learning, cache models

- embedded systems

- comparison with other LMs
Expand All @@ -9,17 +10,19 @@ http://ucrel.lancs.ac.uk/llwizard.html

- stream based Python

- voorbeelden bij metrics in documentatie
DONE voorbeelden bij metrics in documentatie

- C++ API genereren

- TEchnical Report
- Technical Report

- Higher level wrapper scripts

- CLAM wrapper

- PIP install / Anaconda package

- PIP install / Anaconda package
(wrote up on it, making a package seems harder)

- SOS/EOS markers

14 changes: 7 additions & 7 deletions include/patternmodel.h
Original file line number Diff line number Diff line change
Expand Up @@ -2054,13 +2054,13 @@ class IndexedPatternModel: public PatternModel<IndexedData,IndexedDataHandler,Ma
const Pattern p = *iter2;
types.insert(p);
}
IndexedData * data = this->getdata(pattern);
for (IndexedData::iterator dataiter = data->begin(); dataiter != data->end(); dataiter++) {
//take into account all tokens
for (int i = 0; i < pattern.n(); i++) {
tokens.insert(*dataiter + i);
}
}
}
}
IndexedData * data = this->getdata(pattern);
for (IndexedData::iterator dataiter = data->begin(); dataiter != data->end(); dataiter++) {
//take into account all tokens
for (int i = 0; i < pattern.n(); i++) {
tokens.insert(*dataiter + i);
}
}
iter++;
Expand Down

0 comments on commit 3debd7d

Please sign in to comment.