diff --git a/preprocessing.py b/preprocessing.py
index 19c52ab..9e35fc9 100644
--- a/preprocessing.py
+++ b/preprocessing.py
@@ -1,5 +1,6 @@
 
 
+import sys
 from read_data import TagIdConverter, make_long_namedEntity, readKoreanDataAll, Sentence
 from typing import Any, NamedTuple, List, Sequence, TypeVar
 import json
@@ -45,16 +46,16 @@ if __name__ == "__main__":
     PRETAINED_MODEL_NAME = 'bert-base-multilingual-cased'
 
     rawTrain, rawDev, rawTest = readKoreanDataAll()
-    print("load tokenzier...")
+    print("load tokenzier...",file=sys.stderr)
     tokenizer = BertTokenizer.from_pretrained(PRETAINED_MODEL_NAME)
     converter = TagIdConverter()
 
-    print("process train...")
+    print("process train...",file=sys.stderr)
     train = preprocessing(tokenizer,converter,rawTrain)
     saveObject(path.join(PRE_BASE_PATH,"train.json"),train)
-    print("process dev...")
+    print("process dev...",file=sys.stderr)
     dev = preprocessing(tokenizer,converter,rawDev)
     saveObject(path.join(PRE_BASE_PATH,"dev.json"),dev)
-    print("process test...")
+    print("process test...",file=sys.stderr)
     test = preprocessing(tokenizer,converter,rawTest)
     saveObject(path.join(PRE_BASE_PATH,"test.json"),test)