summaryrefslogtreecommitdiffstats
path: root/utils
diff options
context:
space:
mode:
authorPeng Wu <alexepico@gmail.com>2011-05-25 10:50:31 +0800
committerPeng Wu <alexepico@gmail.com>2011-05-25 14:15:19 +0800
commit4bb7f472a4fba814d02bdca5b011b5a6f68d4949 (patch)
treed01ffcc53e846f823b1a024fa7f7d43898c3eb3f /utils
parent2688f37ac6cbdb6276ec5bf9f9bc3b1a5ad393c0 (diff)
downloadlibpinyin-4bb7f472a4fba814d02bdca5b011b5a6f68d4949.tar.gz
libpinyin-4bb7f472a4fba814d02bdca5b011b5a6f68d4949.tar.xz
libpinyin-4bb7f472a4fba814d02bdca5b011b5a6f68d4949.zip
remove <start> in \1-gram in conversion tool
Diffstat (limited to 'utils')
-rw-r--r--utils/training/k_mixture_model_to_interpolation.cpp3
1 files changed, 3 insertions, 0 deletions
diff --git a/utils/training/k_mixture_model_to_interpolation.cpp b/utils/training/k_mixture_model_to_interpolation.cpp
index 3652f27..fe79422 100644
--- a/utils/training/k_mixture_model_to_interpolation.cpp
+++ b/utils/training/k_mixture_model_to_interpolation.cpp
@@ -95,6 +95,9 @@ bool parse_unigram(FILE * input, FILE * output){
case GRAM_1_ITEM_LINE: {
/* handle \item in \1-gram */
const char * string = (const char *) g_ptr_array_index(values, 0);
+ /* remove the "<start>" in the uni-gram of interpolation model */
+ if ( strcmp("<start>", string) == 0 )
+ break;
gpointer value = NULL;
assert(g_hash_table_lookup_extended(required, "count",
NULL, &value));