!8325 fix nccl bug in gpu docker container

From: @HW_KK
Reviewed-by: @leonwanghui,@xsmq
Signed-off-by: @xsmq

README.md

@@ -173,7 +173,7 @@ currently the containerized build options are supported as follows:
     Then you can pull and run the latest stable image using the below command:
     ```
     docker pull mindspore/mindspore-gpu:1.0.0
-    docker run -it --runtime=nvidia --privileged=true mindspore/mindspore-gpu:1.0.0 /bin/bash
+    docker run -it -v /dev/shm:/dev/shm --runtime=nvidia --privileged=true mindspore/mindspore-gpu:1.0.0 /bin/bash
     ```
 
     To test if the docker image works, please execute the python code below and check the output:

README_CN.md

@@ -169,7 +169,7 @@ MindSpore的Docker镜像托管在[Docker Hub](https://hub.docker.com/r/mindspore
     使用以下命令获取并运行最新的稳定镜像：
     ```
     docker pull mindspore/mindspore-gpu:1.0.0
-    docker run -it --runtime=nvidia --privileged=true mindspore/mindspore-gpu:1.0.0 /bin/bash
+    docker run -it -v /dev/shm:/dev/shm --runtime=nvidia --privileged=true mindspore/mindspore-gpu:1.0.0 /bin/bash
     ```
 
     要测试Docker是否正常工作，请运行下面的Python代码并检查输出：