diff --git a/tutorials/source_en/advanced_use/host_device_training.md b/tutorials/source_en/advanced_use/host_device_training.md
index cb9ae8e6a5b113ee7faed9e5549319f0d0589d78..befc7740437de4e3f60ecaa714b85f44e68ced84 100644
--- a/tutorials/source_en/advanced_use/host_device_training.md
+++ b/tutorials/source_en/advanced_use/host_device_training.md
@@ -72,6 +72,12 @@ This tutorial introduces how to train [Wide&Deep](https://gitee.com/mindspore/mi
     ```python
     self.embedding_table = Parameter(Tensor(np.random.normal(loc=0.0, scale=0.01, size=[184968, 80]).astype(dtype=np_type)), name='V_l2',  sparse_grad='V_l2')
     ```
+   
+    In the same file, add the import information in the head:
+    
+    ```python
+    from mindspore import Tensor
+    ```
     
     In the `construct` function of `class WideDeepModel(nn.Cell)` of file `src/wide_and_deep.py`, to adapt for sparse parameters, replace the return value as:
     
@@ -103,10 +109,9 @@ This tutorial introduces how to train [Wide&Deep](https://gitee.com/mindspore/mi
 
 ## Training the Model
 
-Use the script `script/run_auto_parallel_train.sh`, and run the command `bash run_auto_parallel_train.sh 1 1 DATASET RANK_TABLE_FILE MINDSPORE_HCCL_CONFIG_PATH`,
+Use the script `script/run_auto_parallel_train.sh`. Run the command `bash run_auto_parallel_train.sh 1 1 DATASET RANK_TABLE_FILE`,
 where the first `1` is the number of accelerators, the second `1` is the number of epochs, `DATASET` is the path of dataset,
-and `RANK_TABLE_FILE` and `MINDSPORE_HCCL_CONFIG_PATH` is the path of the above `rank_table_1p_0.json` file.
-
+and `RANK_TABLE_FILE` is the path of the above `rank_table_1p_0.json` file.
 
 The running log is in the directory of `device_0`, where `loss.log` contains every loss value of every step in the epoch:
 
diff --git a/tutorials/source_zh_cn/advanced_use/host_device_training.md b/tutorials/source_zh_cn/advanced_use/host_device_training.md
index 7e169c1bcaf2ff8f32c6186b40fca2bfef22cdcc..9dd8750a4666c0a548811fb3cf5f42bd57fd07bd 100644
--- a/tutorials/source_zh_cn/advanced_use/host_device_training.md
+++ b/tutorials/source_zh_cn/advanced_use/host_device_training.md
@@ -70,7 +70,11 @@
     ```python
     self.embedding_table = Parameter(Tensor(np.random.normal(loc=0.0, scale=0.01, size=[184968, 80]).astype(dtype=np_type)), name='V_l2',  sparse_grad='V_l2')
     ```
+    此外，需要在`src/wide_and_deep.py`的头文件引用中添加：
     
+    ```python
+    from mindspore import Tensor
+    ```
     在`src/wide_and_deep.py`文件的`class WideDeepModel(nn.Cell)`类的`construct`函数中，将函数的返回值替换为如下值，以适配参数的稀疏性：
     
     ```
@@ -101,9 +105,8 @@
 
 ## 训练模型
 
-使用训练脚本`script/run_auto_parallel_train.sh`，
-执行命令：`bash run_auto_parallel_train.sh 1 1 DATASET RANK_TABLE_FILE MINDSPORE_HCCL_CONFIG_PATH`，
-其中第一个`1`表示用例使用的卡数，第二`1`表示训练的epoch数，`DATASET`是数据集所在路径，`RANK_TABLE_FILE`和`MINDSPORE_HCCL_CONFIG_PATH`为上述`rank_table_1p_0.json`文件所在路径。
+使用训练脚本`script/run_auto_parallel_train.sh`。执行命令：`bash run_auto_parallel_train.sh 1 1 DATASET RANK_TABLE_FILE`，
+其中第一个`1`表示用例使用的卡数，第二`1`表示训练的epoch数，`DATASET`是数据集所在路径，`RANK_TABLE_FILE`为上述`rank_table_1p_0.json`文件所在路径。
 
 运行日志保存在`device_0`目录下，其中`loss.log`保存一个epoch内中多个loss值，如下：