Update (ab1199ee) · Commits · 钟慕尧 / deepchem

deepchem/feat/graph_data.py

+4 −6

Original line number	Diff line number	Diff line
		@@ -136,16 +136,14 @@ class GraphData:
		This method requires DGL to be installed.
		"""
		try:
		import dgl
		import torch
		from dgl import DGLGraph
		except ModuleNotFoundError:
		raise ValueError("This function requires DGL to be installed.")

		g = DGLGraph()
		g.add_nodes(self.num_nodes)
		g.add_edges(
		torch.from_numpy(self.edge_index[0]).long(),
		torch.from_numpy(self.edge_index[1]).long())
		g = dgl.graph((torch.from_numpy(self.edge_index[0]).long(),
		torch.from_numpy(self.edge_index[1]).long()),
		num_nodes=self.num_nodes)
		g.ndata['x'] = torch.from_numpy(self.node_features).float()

		if self.node_pos_features is not None:

deepchem/models/init.py

+1 −0

Original line number	Diff line number	Diff line
		@@ -33,6 +33,7 @@ try:
		from deepchem.models.torch_models import TorchModel
		from deepchem.models.torch_models import CGCNN, CGCNNModel
		from deepchem.models.torch_models import GAT, GATModel
		from deepchem.models.torch_models import GCN, GCNModel
		except ModuleNotFoundError:
		pass

deepchem/models/torch_models/gcn.py

+24 −6

Original line number	Diff line number	Diff line
		@@ -21,7 +21,19 @@ class GCN(nn.Module):

		Examples
		--------
		# Todo

		>>> import deepchem as dc
		>>> import pymatgen as mg
		>>> from deepchem.models import GCN
		>>> lattice = mg.Lattice.cubic(4.2)
		>>> structure = mg.Structure(lattice, ["Cs", "Cl"], [[0, 0, 0], [0.5, 0.5, 0.5]])
		>>> featurizer = dc.feat.CGCNNFeaturizer()
		>>> cgcnn_graph = featurizer.featurize([structure])[0]
		>>> cgcnn_graph.num_node_features
		92
		>>> cgcnn_dgl_graph = cgcnn_graph.to_dgl_graph()
		>>> model = GCN(in_node_dim=92, hidden_node_dim=92, num_gnn_layers=2)
		>>> model(cgcnn_dgl_graph)

		References
		----------
		@@ -39,7 +51,7 @@ class GCN(nn.Module):
		num_gnn_layers: int,
		activation = None,
		residual: bool = True,
		batchnorm: bool = True,
		batchnorm: bool = False,
		dropout: float = 0.,
		predictor_hidden_feats: int = 128,
		predictor_dropout: float = 0.,
		@@ -62,7 +74,8 @@ class GCN(nn.Module):
		residual: bool
		Whether to add a residual connection within each GCN layer. Default to True.
		batchnorm: bool
		Whether to apply batch normalization to the output of each GCN layer. Default to True.
		Whether to apply batch normalization to the output of each GCN layer.
		Default to False.
		dropout: float
		The dropout probability for the output of each GCN layer. Default to 0.
		predictor_hidden_feats: int
		@@ -104,9 +117,13 @@ class GCN(nn.Module):
		out_size = n_tasks

		from dgllife.model import GCNPredictor as DGLGCNPredictor

		if activation is not None:
		activation = [activation] * num_gnn_layers

		self.model = DGLGCNPredictor(in_feats=in_node_dim,
		hidden_feats=[hidden_node_dim] * num_gnn_layers,
		activation=[activation] * num_gnn_layers,
		activation=activation,
		residual=[residual] * num_gnn_layers,
		batchnorm=[batchnorm] * num_gnn_layers,
		dropout=[dropout] * num_gnn_layers,
		@@ -185,7 +202,7 @@ class GCNModel(TorchModel):
		num_gnn_layers: int,
		activation = None,
		residual: bool = True,
		batchnorm: bool = True,
		batchnorm: bool = False,
		dropout: float = 0.,
		predictor_hidden_feats: int = 128,
		predictor_dropout: float = 0.,
		@@ -209,7 +226,8 @@ class GCNModel(TorchModel):
		residual: bool
		Whether to add a residual connection within each GCN layer. Default to True.
		batchnorm: bool
		Whether to apply batch normalization to the output of each GCN layer. Default to True.
		Whether to apply batch normalization to the output of each GCN layer.
		Default to False.
		dropout: float
		The dropout probability for the output of each GCN layer. Default to 0.
		predictor_hidden_feats: int

Admin message