Merge pull request #18 from ecrl/no_readout

MoleculeGCN updates

graphchem/nn/gcn.py

@@ -18,6 +18,8 @@ class MoleculeGCN(nn.Module):
         Probability of an element to be zeroed in dropout layers.
     _n_messages : int
         Number of message passing steps.
+    act_fn : callable
+        Activation function, e.g., `torch.nn.functional.softplus`
     emb_atom : nn.Embedding
         Embedding layer for atoms.
     emb_bond : nn.Embedding
@@ -36,7 +38,8 @@ def __init__(self, atom_vocab_size: int, bond_vocab_size: int,
                  n_readout: Optional[int] = 2,
                  readout_dim: Optional[int] = 64,
                  p_dropout: Optional[float] = 0.0,
-                 aggr: Optional[str] = "add"):
+                 aggr: Optional[str] = "add",
+                 act_fn: Optional[callable] = F.softplus):
         """
         Initialize the MoleculeGCN object.
 
@@ -60,34 +63,59 @@ def __init__(self, atom_vocab_size: int, bond_vocab_size: int,
             Dropout probability for the dropout layers.
         aggr : str, optional (default="add")
             Aggregation scheme to use in the GeneralConv layer.
+        act_fn : callable, optional (default=`torch.nn.functional.softplus`)
+            Activation function, e.g., `torch.nn.functional.softplus`,
+            `torch.nn.functional.sigmoid`, `torch.nn.functional.relu`, etc.
         """
         super().__init__()
 
+        # Store attributes
         self._p_dropout = p_dropout
         self._n_messages = n_messages
+        self.act_fn = act_fn
 
+        # Embedding layer for atoms
         self.emb_atom = nn.Embedding(atom_vocab_size, embedding_dim)
+
+        # Embedding layer for bonds
         self.emb_bond = nn.Embedding(bond_vocab_size, embedding_dim)
 
+        # General convolution layer for atoms with specified aggregation method
         self.atom_conv = GeneralConv(
             embedding_dim, embedding_dim, embedding_dim, aggr=aggr
         )
+
+        # Edge convolution layer for bonds using a linear transformation
         self.bond_conv = EdgeConv(nn.Sequential(
             nn.Linear(2 * embedding_dim, embedding_dim)
         ))
 
-        self.readout = nn.ModuleList()
-        self.readout.append(nn.Sequential(
-            nn.Linear(embedding_dim, readout_dim)
-        ))
-        if n_readout > 1:
-            for _ in range(n_readout - 1):
-                self.readout.append(nn.Sequential(
-                    nn.Linear(readout_dim, readout_dim)
-                ))
-        self.readout.append(nn.Sequential(
-            nn.Linear(readout_dim, output_dim)
-        ))
+        # Initialize the readout network if readout layers are specified
+        if n_readout > 0:
+
+            # Create a list to hold the readout network modules
+            self.readout = nn.ModuleList()
+
+            # First layer of the readout network
+            self.readout.append(nn.Sequential(
+                nn.Linear(embedding_dim, readout_dim)
+            ))
+
+            # Additional hidden layers for the readout network if needed
+            if n_readout > 1:
+                for _ in range(n_readout - 1):
+                    self.readout.append(nn.Sequential(
+                        nn.Linear(readout_dim, readout_dim)
+                    ))
+
+            # Final layer of the readout network to produce output dimensions
+            self.readout.append(nn.Sequential(
+                nn.Linear(readout_dim, output_dim)
+            ))
+
+        # No readout network if n_readout is 0
+        else:
+            self.readout = None
 
     def forward(
             self,
@@ -111,41 +139,64 @@ def forward(
         out_bond : torch.Tensor
             Bond-level representations after message passing.
         """
+        # Extract node features, edge attributes, edge indices, and batch
+        #   vector from data
         x, edge_attr, edge_index, batch = data.x, data.edge_attr, \
             data.edge_index, data.batch
+
+        # If no node features are provided, initialize with ones
         if data.num_node_features == 0:
             x = torch.ones(data.num_nodes, 1)
 
+        # Embed and activate atom features
         out_atom = self.emb_atom(x)
-        out_atom = F.softplus(out_atom)
+        out_atom = self.act_fn(out_atom)
 
+        # Embed and activate bond features
         out_bond = self.emb_bond(edge_attr)
-        out_bond = F.softplus(out_bond)
+        out_bond = self.act_fn(out_bond)
 
+        # Perform message passing for the specified number of steps
         for _ in range(self._n_messages):
 
+            # Update bond representations using edge convolution
             out_bond = self.bond_conv(out_bond, edge_index)
-            out_bond = F.softplus(out_bond)
+            out_bond = self.act_fn(out_bond)
+
+            # Apply dropout
             out_bond = F.dropout(
                 out_bond, p=self._p_dropout, training=self.training
             )
 
+            # Update atom representations using general convolution
             out_atom = self.atom_conv(out_atom, edge_index, out_bond)
-            out_atom = F.softplus(out_atom)
+            out_atom = self.act_fn(out_atom)
+
+            # Apply dropout
             out_atom = F.dropout(
                 out_atom, p=self._p_dropout, training=self.training
             )
 
+        # Aggregate atom representations across batches with global add pooling
         out = global_add_pool(out_atom, batch)
 
-        for layer in self.readout[:-1]:
+        # Process aggregated atom representation through the readout network
+        if self.readout is not None:
 
-            out = layer(out)
-            out = F.softplus(out)
-            out = F.dropout(
-                out, p=self._p_dropout, training=self.training
-            )
+            # Iterate over all but the last layer of the readout network
+            for layer in self.readout[:-1]:
+
+                # Pass through layer and activate
+                out = layer(out)
+                out = self.act_fn(out)
+
+                # Apply dropout
+                out = F.dropout(
+                    out, p=self._p_dropout, training=self.training
+                )
 
-        out = self.readout[-1](out)
+            # Final layer of the readout network to produce output dimensions
+            out = self.readout[-1](out)
 
+        # Return final prediction, atom representations, bond representations
         return (out, out_atom, out_bond)

tests/test_gcn.py

@@ -6,7 +6,7 @@
 
 
 # Test instantiation of the MoleculeGCN class with various parameters
-def test_moleculgcn_instantiation():
+def test_moleculegcn_instantiation():
 
     model = MoleculeGCN(
         atom_vocab_size=10,
@@ -21,21 +21,24 @@ def test_moleculgcn_instantiation():
 
     assert model._p_dropout == 0.2
     assert model._n_messages == 3
+    assert model.readout is not None
 
 
 # Test the forward pass of the MoleculeGCN model
-def test_moleculgcn_forward_pass():
+def test_moleculegcn_forward_pass():
 
     atom_vocab_size = 10
     bond_vocab_size = 5
+    embedding_dim = 64
+    output_dim = 2
     n_atoms = 3
     n_bonds = 5
 
     model = MoleculeGCN(
         atom_vocab_size=atom_vocab_size,
         bond_vocab_size=bond_vocab_size,
-        output_dim=2,
-        embedding_dim=64,
+        output_dim=output_dim,
+        embedding_dim=embedding_dim,
         n_messages=3,
         n_readout=3,
         readout_dim=128,
@@ -54,6 +57,95 @@ def test_moleculgcn_forward_pass():
 
     out, out_atom, out_bond = model.forward(data)
 
-    assert out.shape == (1, 2)
-    assert out_atom.shape == (3, 64)
-    assert out_bond.shape == (5, 64)
+    assert out.shape == (1, output_dim)
+    assert out_atom.shape == (n_atoms, embedding_dim)
+    assert out_bond.shape == (n_bonds, embedding_dim)
+
+
+# Test the forward pass of the MoleculeGCN model without readout layers
+def test_moleculegcn_no_readout():
+
+    atom_vocab_size = 10
+    bond_vocab_size = 5
+    embedding_dim = 64
+    n_atoms = 3
+    n_bonds = 5
+
+    model = MoleculeGCN(
+        atom_vocab_size=atom_vocab_size,
+        bond_vocab_size=bond_vocab_size,
+        output_dim=None,
+        embedding_dim=embedding_dim,
+        n_messages=3,
+        n_readout=0,
+        readout_dim=128,
+        p_dropout=0.2
+    )
+
+    assert model.readout is None
+
+    n_atoms = 3
+
+    # Create a mock input for the forward pass
+    data = Data(
+        x=torch.randint(0, atom_vocab_size, (n_atoms,)),
+        edge_index=torch.randint(0, n_atoms, (2, n_bonds)),
+        edge_attr=torch.randint(0, bond_vocab_size, (n_bonds,)),
+        batch=torch.tensor([0, 0, 0], dtype=torch.long)
+    )
+
+    out, out_atom, out_bond = model.forward(data)
+
+    assert out.shape == (1, embedding_dim)
+    assert out_atom.shape == (n_atoms, embedding_dim)
+    assert out_bond.shape == (n_bonds, embedding_dim)
+
+
+# Test the forward pass of the MoleculeGCN model with different activation fns
+def test_moleculegcn_act_fns():
+
+    functions = [
+        torch.nn.functional.softplus,
+        torch.nn.functional.sigmoid,
+        torch.nn.functional.relu,
+        torch.nn.functional.tanh,
+        torch.nn.functional.leaky_relu
+    ]
+
+    for fn in functions:
+
+        atom_vocab_size = 10
+        bond_vocab_size = 5
+        embedding_dim = 64
+        n_atoms = 3
+        n_bonds = 5
+
+        model = MoleculeGCN(
+            atom_vocab_size=atom_vocab_size,
+            bond_vocab_size=bond_vocab_size,
+            output_dim=None,
+            embedding_dim=embedding_dim,
+            n_messages=3,
+            n_readout=0,
+            readout_dim=128,
+            p_dropout=0.2,
+            act_fn=fn
+        )
+
+        assert model.readout is None
+
+        n_atoms = 3
+
+        # Create a mock input for the forward pass
+        data = Data(
+            x=torch.randint(0, atom_vocab_size, (n_atoms,)),
+            edge_index=torch.randint(0, n_atoms, (2, n_bonds)),
+            edge_attr=torch.randint(0, bond_vocab_size, (n_bonds,)),
+            batch=torch.tensor([0, 0, 0], dtype=torch.long)
+        )
+
+        out, out_atom, out_bond = model.forward(data)
+
+        assert out.shape == (1, embedding_dim)
+        assert out_atom.shape == (n_atoms, embedding_dim)
+        assert out_bond.shape == (n_bonds, embedding_dim)