[doc] fix invalid link in user guide (#4468)

docs/source/guide/minibatch-node.rst

@@ -176,7 +176,7 @@ of MFGs, we:
         opt.step()
 
 DGL provides an end-to-end stochastic training example `GraphSAGE
-implementation <https://github.com/dmlc/dgl/blob/master/examples/pytorch/graphsage/train_sampling.py>`__.
+implementation <https://github.com/dmlc/dgl/blob/master/examples/pytorch/graphsage/node_classification.py>`__.
 
 For heterogeneous graphs
 ~~~~~~~~~~~~~~~~~~~~~~~~

docs/source/guide_cn/minibatch-node.rst

@@ -142,7 +142,7 @@ DGL提供了几个邻居采样类，这些类会生成需计算的节点在每
         opt.step()
 
 DGL提供了一个端到端的随机批次训练示例
-`GraphSAGE的实现 <https://github.com/dmlc/dgl/blob/master/examples/pytorch/graphsage/train_sampling.py>`__。
+`GraphSAGE的实现 <https://github.com/dmlc/dgl/blob/master/examples/pytorch/graphsage/node_classification.py>`__。
 
 
 异构图上模型的训练

docs/source/guide_ko/minibatch-node.rst

@@ -135,7 +135,7 @@ Iterator는 매번 세개의 아이템을 생성한다. ``input_nodes`` 는 ``ou
         opt.step()
 
 DGL에서는 end-to-end stochastic 학습 예제인 `GraphSAGE
-implementation <https://github.com/dmlc/dgl/blob/master/examples/pytorch/graphsage/train_sampling.py>`__ 를 제공한다.
+implementation <https://github.com/dmlc/dgl/blob/master/examples/pytorch/graphsage/node_classification.py>`__ 를 제공한다.
 
 Heterogeneous 그래프의 경우
 ~~~~~~~~~~~~~~~~~~~~~~~~