diff --git a/Jenkinsfile b/Jenkinsfile
index 8b1054343535..233f475f532b 100644
--- a/Jenkinsfile
+++ b/Jenkinsfile
@@ -217,11 +217,11 @@ pipeline {
             //    unit_test_linux("pytorch", "cpu")
             //  }
             //}
-            //stage("Tutorial test") {
-            //  steps {
-            //    tutorial_test_linux("mxnet")
-            //  }
-            //}
+            stage("Tutorial test") {
+              steps {
+                tutorial_test_linux("mxnet")
+              }
+            }
           }
         }
         stage("MXNet GPU") {
@@ -243,11 +243,11 @@ pipeline {
             //    unit_test_linux("pytorch", "cpu")
             //  }
             //}
-            //stage("Tutorial test") {
-            //  steps {
-            //    tutorial_test_linux("mxnet")
-            //  }
-            //}
+            stage("Tutorial test") {
+              steps {
+                tutorial_test_linux("mxnet")
+              }
+            }
           }
         }
       }
diff --git a/tutorials/models/5_giant_graph/1_sampling_mx.py b/tutorials/models/5_giant_graph/1_sampling_mx.py
index 39b3834abbc4..1e8484cb5738 100644
--- a/tutorials/models/5_giant_graph/1_sampling_mx.py
+++ b/tutorials/models/5_giant_graph/1_sampling_mx.py
@@ -250,9 +250,9 @@ def forward(self, nf):
 # dropout probability
 dropout = 0.2
 # batch size
-batch_size = 10000
+batch_size = 1000
 # number of neighbors to sample
-num_neighbors = 8
+num_neighbors = 4
 # number of epochs
 num_epochs = 1
 
@@ -267,6 +267,7 @@ def forward(self, nf):
                         {'learning_rate': 0.03, 'wd': 0})
 
 for epoch in range(num_epochs):
+    i = 0
     for nf in dgl.contrib.sampling.NeighborSampler(g, batch_size,
                                                    num_neighbors,
                                                    neighbor_type='in',
@@ -291,6 +292,10 @@ def forward(self, nf):
         # optimization
         trainer.step(batch_size=1)
         print("Epoch[{}]: loss {}".format(epoch, loss.asscalar()))
+        i += 1
+        # We only train the model with 32 mini-batches just for demonstration.
+        if i >= 32:
+            break
 
 ##############################################################################
 # Control Variate