Fix docs on GraphSAGE normalization (#3711)

Co-authored-by: Jinjing Zhou <VoVAllen@users.noreply.github.com>

python/dgl/nn/mxnet/conv/sageconv.py

@@ -24,7 +24,7 @@ class SAGEConv(nn.Block):
         h_{i}^{(l+1)} &= \sigma \left(W \cdot \mathrm{concat}
         (h_{i}^{l}, h_{\mathcal{N}(i)}^{l+1}) \right)
 
-        h_{i}^{(l+1)} &= \mathrm{norm}(h_{i}^{l})
+        h_{i}^{(l+1)} &= \mathrm{norm}(h_{i}^{(l+1)})
 
     Parameters
     ----------

python/dgl/nn/pytorch/conv/sageconv.py

@@ -24,7 +24,7 @@ class SAGEConv(nn.Module):
         h_{i}^{(l+1)} &= \sigma \left(W \cdot \mathrm{concat}
         (h_{i}^{l}, h_{\mathcal{N}(i)}^{l+1}) \right)
 
-        h_{i}^{(l+1)} &= \mathrm{norm}(h_{i}^{l})
+        h_{i}^{(l+1)} &= \mathrm{norm}(h_{i}^{(l+1)})
 
     If a weight tensor on each edge is provided, the aggregation becomes:
 

python/dgl/nn/tensorflow/conv/sageconv.py

@@ -23,7 +23,7 @@ class SAGEConv(layers.Layer):
         h_{i}^{(l+1)} &= \sigma \left(W \cdot \mathrm{concat}
         (h_{i}^{l}, h_{\mathcal{N}(i)}^{l+1}) \right)
 
-        h_{i}^{(l+1)} &= \mathrm{norm}(h_{i}^{l})
+        h_{i}^{(l+1)} &= \mathrm{norm}(h_{i}^{(l+1)})
 
     Parameters
     ----------