heads should be divisible by output or hidden channels

Author: aditya0by0

configs/model/gat.yml

@@ -7,7 +7,7 @@ init_args:
     hidden_length: 512
     dropout_rate: 0.1
     n_conv_layers: 3
-    heads: 5  # Default is one
+    heads: 4  # the number of heads should be divisible by output channels (hidden channels if output channel not given)
     # v2: True  # -- to use `torch_geometric.nn.conv.GATv2Conv` convolution layers, default is GATConv
     n_linear_layers: 3
     n_atom_properties: 158