diff --git a/README.md b/README.md
index 2b1f43c..f5bbf13 100644
--- a/README.md
+++ b/README.md
@@ -28,7 +28,7 @@ Then set up the environment via Docker or Anaconda.
 This is the most convenient way to try _NH\_Rep_, everything is already settled down in the docker.
 
         $ docker pull horaceguo/pytorchigr:isg
-        $ docker run --runtime=nvidia --ipc=host --net=host -v PATH_TO_NH-REP/:/workspace -t -i horaceguo/pytorchigr:isg
+        $ docker run  -e TZ=Asia/Shanghai  --runtime=nvidia --ipc=host --net=host -v PATH_TO_NH-REP/:/workspace -t -i horaceguo/pytorchigr:isg
         $ cd /workspace
         
 Then you can convert the points sampled on B-Rep model in _input\_data_ to implicit representation: