diff --git a/tutorials/source_en/train_availability/disaster_recover.md b/tutorials/source_en/train_availability/disaster_recover.md
index b279cf6ecfc76453dbc64b1e4967ed298cd8917b..1b8ee745f9477894ba78b6bd504c238f0dd93255 100644
--- a/tutorials/source_en/train_availability/disaster_recover.md
+++ b/tutorials/source_en/train_availability/disaster_recover.md
@@ -14,7 +14,7 @@ For more detailed instructions, see dynamic cluster environment variables in the
 
 ## Operation Practice
 
-The following is an example of how to do this with Ascend:
+The following is an example of how to do this with GPU:
 
 ### Sample Code Description
 
diff --git a/tutorials/source_zh_cn/train_availability/disaster_recover.md b/tutorials/source_zh_cn/train_availability/disaster_recover.md
index afbbbb6961d2d568f0880f1916a0077a584e8c98..77078c85d74e4c4cdffe7c68febc38adc5c33d91 100644
--- a/tutorials/source_zh_cn/train_availability/disaster_recover.md
+++ b/tutorials/source_zh_cn/train_availability/disaster_recover.md
@@ -14,7 +14,7 @@
 
 ## 操作实践
 
-下面以Ascend为例进行操作说明：
+下面以GPU为例进行操作说明：
 
 ### 样例代码说明