diff --git a/official/recommend/deepfm/README.md b/official/recommend/deepfm/README.md
index 827de641eaa9228cd3b0e53a2ccb8e76f6be79d9..53be80b0527e1a85fd48fafe05f95219a27366a0 100644
--- a/official/recommend/deepfm/README.md
+++ b/official/recommend/deepfm/README.md
@@ -56,16 +56,16 @@ After installing MindSpore via the official website, you can start training and
 
 - preprocess dataset
 
-  '''bash
+  ```bash
   #download dataset
-  #Please refer to [1] to obtain the download link
+  #Please refer to [Criteo Kaggle Display Advertising Challenge Dataset] to get the download URL and assign it to `DATA_LINK`
   mkdir -p data/origin_data && cd data/origin_data
   wget DATA_LINK
   tar -zxvf dac.tar.gz
 
   #preprocess dataset
   python -m src.preprocess_data  --data_path=./data/ --dense_dim=13 --slot_dim=26 --threshold=100 --train_line_count=45840617 --skip_id_convert=0
-  '''
+  ```
 
 - running on Ascend
 
diff --git a/official/recommend/deepfm/README_CN.md b/official/recommend/deepfm/README_CN.md
index e4df14338d0e0343778ddbc434926c9a4f5a4803..5a86bfc5e78b9f49baa2dee21ac24cd7e9fa9116 100644
--- a/official/recommend/deepfm/README_CN.md
+++ b/official/recommend/deepfm/README_CN.md
@@ -59,16 +59,17 @@ FM和深度学习部分拥有相同的输入原样特征向量，让DeepFM能从
 通过官方网站安装MindSpore后，您可以按照如下步骤进行训练和评估：
 
 - 数据集预处理
-  '''bash
+
+  ```bash
   #下载数据集
-  #请参考[1]获得下载链接
+  #请参考[Criteo Kaggle Display Advertising Challenge Dataset]获得下载链接并赋值给`DATA_LINK`
   mkdir -p data/origin_data && cd data/origin_data
   wget DATA_LINK
   tar -zxvf dac.tar.gz
 
-  #数据集预处理脚步执行
+  #数据集预处理脚本执行
   python -m src.preprocess_data  --data_path=./data/ --dense_dim=13 --slot_dim=26 --threshold=100 --train_line_count=45840617 --skip_id_convert=0
-  '''
+  ```
 
 - Ascend处理器环境运行