diff --git a/model_zoo/official/nlp/pangu_alpha/README.md b/model_zoo/official/nlp/pangu_alpha/README.md
index 90f820200d4..6937e1eac4c 100644
--- a/model_zoo/official/nlp/pangu_alpha/README.md
+++ b/model_zoo/official/nlp/pangu_alpha/README.md
@@ -120,7 +120,7 @@ The vocab size of `vocab.vocab` is 40000, and the `eod id` is 6.
 
 ### Training On Ascend
 
-Currently the scripts provide three default configures : `2.6B` `13B` and `200B`. The following command will start training `2.6B` model on 8 **Ascend cards**.
+Currently the scripts provide three default configures : `2.6B` `13B` and `200B`.
 
 ```bash
 
@@ -465,7 +465,7 @@ Please check the official [homepage](https://gitee.com/mindspore/mindspore/tree/
 
 # [Requirements](#contents)
 
-- mindspore 1.2.1 or higher version
+- mindspore 1.3.0
 - jieba 0.42.1
 - sentencepiece 0.1.94
 - transformers >= 4.7.0
diff --git a/model_zoo/official/nlp/pangu_alpha/train.py b/model_zoo/official/nlp/pangu_alpha/train.py
index 7834cd682a6..bca1e00ac19 100644
--- a/model_zoo/official/nlp/pangu_alpha/train.py
+++ b/model_zoo/official/nlp/pangu_alpha/train.py
@@ -83,7 +83,7 @@ def run_train(args_opt):
     # Set execution mode
     context.set_context(mode=context.GRAPH_MODE,
                         device_target=args_opt.device_target)
-    context.set_context(variable_memory_max_size="30GB")
+    context.set_context(variable_memory_max_size="31GB")
     # Set parallel context
     if args_opt.distribute == "true":
         D.init()