Fix User Guide EN and CN typo (#3495)

* Update graph-heterogeneous.rst

`tensor([0, 1, 2, 0, 1, 2])` should be output instead of code

* Update message-api.rst

`updata_all_example()` should be `update_all_example()`

* Update message-efficient.rst

`cat_feat` need to concatenate with `dim=1` for the # edge features to match # edges

* Update nn-construction.rst

all `max_pool` in the aggregator type of `SAGEConv` should be `pool` instead

* Update graph-heterogeneous.rst

`tensor([0, 1, 2, 0, 1, 2])` should be output instead of code

* Update message-api.rst

`updata_all_example()` should be `update_all_example()`

* Update message-efficient.rst

`cat_feat` need to concatenate with `dim=1` for the # edge features to match # edges

* Update nn-construction.rst

all `max_pool` in the aggregator type of `SAGEConv` should be `pool` instead

* Update nn-forward.rst

all `max_pool` in the aggregator type of `SAGEConv` should be `pool` instead

* Update nn-forward.rst

all `max_pool` in the aggregator type of `SAGEConv` should be `pool` instead
Co-authored-by: zhjwy9343 <6593865@qq.com>

docs/source/guide/graph-heterogeneous.rst

@@ -251,7 +251,7 @@ The original node/edge types and type-specific IDs are stored in :py:attr:`~dgl.
     tensor([0, 0, 0, 1, 1, 1])
     >>> # Original type-specific node IDs
     >>> hg.ndata[dgl.NID]
-    >>> tensor([0, 1, 2, 0, 1, 2])
+    tensor([0, 1, 2, 0, 1, 2])
 
     >>> # Order of edge types in the heterograph
     >>> g.etypes

docs/source/guide/message-api.rst

@@ -82,7 +82,7 @@ example：
 
 .. code::
 
-    def updata_all_example(graph):
+    def update_all_example(graph):
         # store the result in graph.ndata['ft']
         graph.update_all(fn.u_mul_e('ft', 'a', 'm'),
                          fn.sum('m', 'ft'))
@@ -105,4 +105,4 @@ compute capacity requirement of ``sm_53`` (Pascal, Volta, Turing and Ampere
 architectures).
 
 User can enable float16 for mixed precision training by compiling DGL from source
-(see :doc:`Mixed Precision Training <mixed_precision>` tutorial for details).
+(see :doc:`Mixed Precision Training <mixed_precision>` tutorial for details).
\ No newline at end of file

docs/source/guide/message-efficient.rst

@@ -32,7 +32,7 @@ implementation would be like:
 
     linear = nn.Parameter(torch.FloatTensor(size=(node_feat_dim * 2, out_dim)))
     def concat_message_function(edges):
-         return {'cat_feat': torch.cat([edges.src['feat'], edges.dst['feat']])}
+         return {'cat_feat': torch.cat([edges.src['feat'], edges.dst['feat']], dim=1)}
     g.apply_edges(concat_message_function)
     g.edata['out'] = g.edata['cat_feat'] @ linear
 

docs/source/guide/nn-construction.rst

@@ -52,14 +52,14 @@ SAGEConv paper, such normalization can be l2 normalization:
 
 .. code::
 
-            # aggregator type: mean, max_pool, lstm, gcn
-            if aggregator_type not in ['mean', 'max_pool', 'lstm', 'gcn']:
+            # aggregator type: mean, pool, lstm, gcn
+            if aggregator_type not in ['mean', 'pool', 'lstm', 'gcn']:
                 raise KeyError('Aggregator type {} not supported.'.format(aggregator_type))
-            if aggregator_type == 'max_pool':
+            if aggregator_type == 'pool':
                 self.fc_pool = nn.Linear(self._in_src_feats, self._in_src_feats)
             if aggregator_type == 'lstm':
                 self.lstm = nn.LSTM(self._in_src_feats, self._in_src_feats, batch_first=True)
-            if aggregator_type in ['mean', 'max_pool', 'lstm']:
+            if aggregator_type in ['mean', 'pool', 'lstm']:
                 self.fc_self = nn.Linear(self._in_dst_feats, out_feats, bias=bias)
             self.fc_neigh = nn.Linear(self._in_src_feats, out_feats, bias=bias)
             self.reset_parameters()
@@ -75,7 +75,7 @@ function, weight initialization is applied by calling
         def reset_parameters(self):
             """Reinitialize learnable parameters."""
             gain = nn.init.calculate_gain('relu')
-            if self._aggre_type == 'max_pool':
+            if self._aggre_type == 'pool':
                 nn.init.xavier_uniform_(self.fc_pool.weight, gain=gain)
             if self._aggre_type == 'lstm':
                 self.lstm.reset_parameters()

docs/source/guide/nn-forward.rst

@@ -125,7 +125,7 @@ Message passing and reducing
                     # divide in_degrees
                     degs = graph.in_degrees().to(feat_dst)
                     h_neigh = (graph.dstdata['neigh'] + graph.dstdata['h']) / (degs.unsqueeze(-1) + 1)
-                elif self._aggre_type == 'max_pool':
+                elif self._aggre_type == 'pool':
                     graph.srcdata['h'] = F.relu(self.fc_pool(feat_src))
                     graph.update_all(fn.copy_u('h', 'm'), fn.max('m', 'neigh'))
                     h_neigh = graph.dstdata['neigh']

docs/source/guide_cn/graph-heterogeneous.rst

@@ -235,7 +235,7 @@ DGL提供了 :func:`dgl.save_graphs` 和 :func:`dgl.load_graphs` 函数，分别
     tensor([0, 0, 0, 1, 1, 1])
     >>> # 原始的特定类型节点ID
     >>> hg.ndata[dgl.NID]
-    >>> tensor([0, 1, 2, 0, 1, 2])
+    tensor([0, 1, 2, 0, 1, 2])
 
     >>> # 异构图中边类型的顺序
     >>> g.etypes

docs/source/guide_cn/message-api.rst

@@ -62,7 +62,7 @@ DGL支持内置的聚合函数 ``sum``、 ``max``、 ``min`` 和 ``mean`` 操作
 
 .. code::
 
-    def updata_all_example(graph):
+    def update_all_example(graph):
         # 在graph.ndata['ft']中存储结果
         graph.update_all(fn.u_mul_e('ft', 'a', 'm'),
                          fn.sum('m', 'ft'))

docs/source/guide_cn/message-efficient.rst

@@ -22,7 +22,7 @@ DGL建议用户尽量减少边的特征维数。
 
     linear = nn.Parameter(torch.FloatTensor(size=(node_feat_dim * 2, out_dim)))
     def concat_message_function(edges):
-         return {'cat_feat': torch.cat([edges.src.ndata['feat'], edges.dst.ndata['feat']])}
+         return {'cat_feat': torch.cat([edges.src.ndata['feat'], edges.dst.ndata['feat']], dim=1)}
     g.apply_edges(concat_message_function)
     g.edata['out'] = g.edata['cat_feat'] @ linear
 

docs/source/guide_cn/nn-construction.rst

@@ -44,14 +44,14 @@
 
 .. code::
 
-            # 聚合类型：mean、max_pool、lstm、gcn
-            if aggregator_type not in ['mean', 'max_pool', 'lstm', 'gcn']:
+            # 聚合类型：mean、pool、lstm、gcn
+            if aggregator_type not in ['mean', 'pool', 'lstm', 'gcn']:
                 raise KeyError('Aggregator type {} not supported.'.format(aggregator_type))
-            if aggregator_type == 'max_pool':
+            if aggregator_type == 'pool':
                 self.fc_pool = nn.Linear(self._in_src_feats, self._in_src_feats)
             if aggregator_type == 'lstm':
                 self.lstm = nn.LSTM(self._in_src_feats, self._in_src_feats, batch_first=True)
-            if aggregator_type in ['mean', 'max_pool', 'lstm']:
+            if aggregator_type in ['mean', 'pool', 'lstm']:
                 self.fc_self = nn.Linear(self._in_dst_feats, out_feats, bias=bias)
             self.fc_neigh = nn.Linear(self._in_src_feats, out_feats, bias=bias)
             self.reset_parameters()
@@ -64,7 +64,7 @@
         def reset_parameters(self):
             """重新初始化可学习的参数"""
             gain = nn.init.calculate_gain('relu')
-            if self._aggre_type == 'max_pool':
+            if self._aggre_type == 'pool':
                 nn.init.xavier_uniform_(self.fc_pool.weight, gain=gain)
             if self._aggre_type == 'lstm':
                 self.lstm.reset_parameters()

docs/source/guide_cn/nn-forward.rst

@@ -105,7 +105,7 @@ SAGEConv的数学公式如下：
                     # 除以入度
                     degs = graph.in_degrees().to(feat_dst)
                     h_neigh = (graph.dstdata['neigh'] + graph.dstdata['h']) / (degs.unsqueeze(-1) + 1)
-                elif self._aggre_type == 'max_pool':
+                elif self._aggre_type == 'pool':
                     graph.srcdata['h'] = F.relu(self.fc_pool(feat_src))
                     graph.update_all(fn.copy_u('h', 'm'), fn.max('m', 'neigh'))
                     h_neigh = graph.dstdata['neigh']