change

preprocess.py

@@ -42,6 +42,7 @@ class DataProcessor(object):
                 if line:
                     labels_map[line] = index
 
+        print("labels_map [{}]".format(labels_map))
         with open(path, mode="r", encoding="UTF-8") as f:
             for line in tqdm(f):
                 line = line.strip()
@@ -52,13 +53,16 @@ class DataProcessor(object):
                 if mapped_label is not None and isinstance(mapped_label, int) and mapped_label >= 0:
                     contents.append(content)
                     labels.append(mapped_label)
-
+                else:
+                    print("not match label  [{}:{}]".format(content,mapped_label))
+                
         #random shuffle
         index = list(range(len(labels)))
         random.seed(self.seed)
         random.shuffle(index)
         contents = [contents[_] for _ in index]
         labels = [labels[_] for _ in index]
+        print("load datas contents label  [{}:{}]".format(len(contents),len(labels)))
         return (contents, labels)
 
     def __next__(self):