diff --git a/pycorrector/seq2seq/preprocess.py b/pycorrector/seq2seq/preprocess.py
index d07611d8..4d063cfc 100644
--- a/pycorrector/seq2seq/preprocess.py
+++ b/pycorrector/seq2seq/preprocess.py
@@ -5,7 +5,6 @@
 """
 import os
 import sys
-from codecs import open
 from xml.dom import minidom
 
 from sklearn.model_selection import train_test_split
diff --git a/pycorrector/seq2seq/seq2seq_model.py b/pycorrector/seq2seq/seq2seq_model.py
index 2501bfe9..db451b28 100644
--- a/pycorrector/seq2seq/seq2seq_model.py
+++ b/pycorrector/seq2seq/seq2seq_model.py
@@ -309,9 +309,6 @@ def train_model(
         if args:
             self.args.update_from_dict(args)
 
-        # if self.args.silent:
-        #     show_running_loss = False
-
         if self.args.evaluate_during_training and eval_data is None:
             raise ValueError(
                 "evaluate_during_training is enabled but eval_data is not specified."
@@ -344,12 +341,6 @@ def train_model(
 
         self.save_model(self.args.output_dir, model=self.model)
 
-        # model_to_save = self.model.module if hasattr(self.model, "module") else self.model
-        # model_to_save.save_pretrained(output_dir)
-        # self.encoder_tokenizer.save_pretrained(output_dir)
-        # self.decoder_tokenizer.save_pretrained(output_dir)
-        # torch.save(self.args, os.path.join(output_dir, "training_args.bin"))
-
         if verbose:
             logger.info(" Training of {} model complete. Saved to {}.".format(self.args.model_name, output_dir))
 
diff --git a/pycorrector/seq2seq/seq2seq_utils.py b/pycorrector/seq2seq/seq2seq_utils.py
index 56f85fff..84fad1ae 100644
--- a/pycorrector/seq2seq/seq2seq_utils.py
+++ b/pycorrector/seq2seq/seq2seq_utils.py
@@ -20,7 +20,7 @@
 
 if transformers.__version__ < "4.2.0":
     shift_tokens_right = lambda input_ids, pad_token_id, decoder_start_token_id: _shift_tokens_right(
-        input_ids, pad_token_id)
+        input_ids, pad_token_id, decoder_start_token_id)
 else:
     shift_tokens_right = _shift_tokens_right
 
diff --git a/pycorrector/seq2seq/tf/config.py b/pycorrector/seq2seq/tf/config.py
index 796872db..305a1bcc 100644
--- a/pycorrector/seq2seq/tf/config.py
+++ b/pycorrector/seq2seq/tf/config.py
@@ -2,7 +2,7 @@
 """
 @author:XuMing(xuming624@qq.com)
 @description:
-""" Use CGED corpus
+"""
 import os
 
 pwd_path = os.path.abspath(os.path.dirname(__file__))
diff --git a/pycorrector/seq2seq/tf/data_reader.py b/pycorrector/seq2seq/tf/data_reader.py
index 75de5d8a..3b25dcbb 100644
--- a/pycorrector/seq2seq/tf/data_reader.py
+++ b/pycorrector/seq2seq/tf/data_reader.py
@@ -2,8 +2,7 @@
 """
 @author:XuMing(xuming624@qq.com)
 @description:
-""" Corpus for model
-
+"""
 import sys
 from codecs import open
 from collections import Counter
diff --git a/pycorrector/seq2seq/train.py b/pycorrector/seq2seq/train.py
index 5826bb24..9a7cd7d5 100644
--- a/pycorrector/seq2seq/train.py
+++ b/pycorrector/seq2seq/train.py
@@ -29,6 +29,7 @@
 from pycorrector.utils.logger import logger
 from pycorrector.seq2seq.seq2seq_model import Seq2SeqModel
 
+os.environ["TOKENIZERS_PARALLELISM"] = "FALSE"
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
 
@@ -262,7 +263,6 @@ def train(arch, train_path, batch_size, embed_size, hidden_size, dropout, epochs
 
         train_df = pd.DataFrame(train_data, columns=['input_text', 'target_text'])
         dev_df = pd.DataFrame(dev_data, columns=['input_text', 'target_text'])
-        os.environ["TOKENIZERS_PARALLELISM"] = "FALSE"
         model.train_model(train_df, eval_data=dev_df)
     else:
         logger.error('error arch: {}'.format(arch))