summaryrefslogtreecommitdiffstats
path: root/utils/training/gen_ngram.cpp
diff options
context:
space:
mode:
authorPeng Wu <alexepico@gmail.com>2011-06-01 10:29:48 +0800
committerPeng Wu <alexepico@gmail.com>2011-06-01 10:29:48 +0800
commit0aec5e7e66ddde36da734993e6afe57f856b23bc (patch)
treef7a80ed003182412b0fa3f8d3030b8f6c7bc4d89 /utils/training/gen_ngram.cpp
parentca16b0628857bd3cef581414a850108c56f782b1 (diff)
downloadlibpinyin-0aec5e7e66ddde36da734993e6afe57f856b23bc.tar.gz
libpinyin-0aec5e7e66ddde36da734993e6afe57f856b23bc.tar.xz
libpinyin-0aec5e7e66ddde36da734993e6afe57f856b23bc.zip
fixes gen ngram
Diffstat (limited to 'utils/training/gen_ngram.cpp')
-rw-r--r--utils/training/gen_ngram.cpp6
1 files changed, 3 insertions, 3 deletions
diff --git a/utils/training/gen_ngram.cpp b/utils/training/gen_ngram.cpp
index f6827dc..9e79f1d 100644
--- a/utils/training/gen_ngram.cpp
+++ b/utils/training/gen_ngram.cpp
@@ -106,6 +106,9 @@ int main(int argc, char * argv[]){
if ( null_token == cur_token )
continue;
+ //training uni-gram
+ phrase_index.add_unigram_frequency(cur_token, 1);
+
/* skip pi-gram training. */
if ( null_token == last_token ){
if ( !train_pi_gram )
@@ -113,9 +116,6 @@ int main(int argc, char * argv[]){
last_token = sentence_start;
}
- //training uni-gram
- phrase_index.add_unigram_frequency(cur_token, 1);
-
//train bi-gram
SingleGram * single_gram = NULL;
bigram.load(last_token, single_gram);