diff --git a/model_zoo/official/nlp/transformer/README.md b/model_zoo/official/nlp/transformer/README.md
index 65bcc18872..69c2272742 100644
--- a/model_zoo/official/nlp/transformer/README.md
+++ b/model_zoo/official/nlp/transformer/README.md
@@ -204,6 +204,8 @@ Parameters for learning rate:
     sh scripts/run_distribute_train_ascend.sh DEVICE_NUM EPOCH_SIZE DATA_PATH RANK_TABLE_FILE
     ```
 
+**Attention**: data sink mode can not be used in transformer since the input datas have different sequence lengths.
+
 ## [Evaluation Process](#contents)
 
 - Set options in `eval_config.py`. Make sure the 'data_file', 'model_file' and 'output_file' are set to your own path.
diff --git a/model_zoo/official/nlp/transformer/README_CN.md b/model_zoo/official/nlp/transformer/README_CN.md
index d0409af49d..a51b84518c 100644
--- a/model_zoo/official/nlp/transformer/README_CN.md
+++ b/model_zoo/official/nlp/transformer/README_CN.md
@@ -211,6 +211,8 @@ Parameters for learning rate:
     sh scripts/run_distribute_train_ascend.sh DEVICE_NUM EPOCH_SIZE DATA_PATH RANK_TABLE_FILE
     ```
 
+**注意**：由于网络输入中有不同句长的数据，所以数据下沉模式不可使用。
+
 ### 评估过程
 
 - 在`eval_config.py`中设置选项。确保已设置了‘data_file'、'model_file’和'output_file'文件路径。