diff options
author | Chris Dyer <prguest11@taipan.cs> | 2012-02-29 07:00:49 +0000 |
---|---|---|
committer | Chris Dyer <prguest11@taipan.cs> | 2012-02-29 07:00:49 +0000 |
commit | a872f46ce1212703b8bed562c894ea1a932c0746 (patch) | |
tree | 3922f1eab5d7f81f87c163aed0868900052be900 /utils/corpus_tools.cc | |
parent | cc01f3b7c9b87928be91e8a89f233a07a183ac2e (diff) |
mpi fixes
Diffstat (limited to 'utils/corpus_tools.cc')
-rw-r--r-- | utils/corpus_tools.cc | 16 |
1 files changed, 10 insertions, 6 deletions
diff --git a/utils/corpus_tools.cc b/utils/corpus_tools.cc index a0542b6e..d17785af 100644 --- a/utils/corpus_tools.cc +++ b/utils/corpus_tools.cc @@ -33,10 +33,12 @@ void CorpusTools::ReadFromFile(const string& filename, while(getline(in, line)) { const bool skip = (lc % size != rank); ++lc; - if (skip) continue; TD::ConvertSentence(line, &tmp); - src->push_back(vector<WordID>()); - vector<WordID>* d = &src->back(); + vector<WordID>* d = NULL; + if (!skip) { + src->push_back(vector<WordID>()); + d = &src->back(); + } set<WordID>* v = src_vocab; int s = 0; for (unsigned i = 0; i < tmp.size(); ++i) { @@ -44,11 +46,13 @@ void CorpusTools::ReadFromFile(const string& filename, ++s; if (s > 1) { cerr << "Unexpected format in line " << lc << ": " << line << endl; abort(); } assert(trg); - trg->push_back(vector<WordID>()); - d = &trg->back(); + if (!skip) { + trg->push_back(vector<WordID>()); + d = &trg->back(); + } v = trg_vocab; } else { - d->push_back(tmp[i]); + if (d) d->push_back(tmp[i]); if (v) v->insert(tmp[i]); } } |