minor changes (2c33ff1a) · Commits · 钟慕尧 / deepchem

deepchem/models/tensorgraph/models/graph_models.py

+8 −4

Original line number	Diff line number	Diff line
		@@ -117,6 +117,7 @@ class WeaveTensorGraph(TensorGraph):
		similar to deepchem.models.tf_new_models.graph_topology.AlternateWeaveTopology.batch_to_feed_dict
		"""
		for epoch in range(epochs):
		if not predict:
		print('Starting epoch %i' % epoch)
		for (X_b, y_b, w_b, ids_b) in dataset.iterbatches(
		batch_size=self.batch_size,
		@@ -266,6 +267,7 @@ class DTNNTensorGraph(TensorGraph):
		similar to deepchem.models.tf_new_models.graph_topology.DTNNGraphTopology.batch_to_feed_dict
		"""
		for epoch in range(epochs):
		if not predict:
		print('Starting epoch %i' % epoch)
		for (X_b, y_b, w_b, ids_b) in dataset.iterbatches(
		batch_size=self.batch_size,
		@@ -411,6 +413,7 @@ class DAGTensorGraph(TensorGraph):
		similar to deepchem.models.tf_new_models.graph_topology.DAGGraphTopology.batch_to_feed_dict
		"""
		for epoch in range(epochs):
		if not predict:
		print('Starting epoch %i' % epoch)
		for (X_b, y_b, w_b, ids_b) in dataset.iterbatches(
		batch_size=self.batch_size,
		@@ -543,6 +546,7 @@ class GraphConvTensorGraph(TensorGraph):
		predict=False,
		pad_batches=True):
		for epoch in range(epochs):
		if not predict:
		print('Starting epoch %i' % epoch)
		for ind, (X_b, y_b, w_b, ids_b) in enumerate(
		dataset.iterbatches(

deepchem/models/tensorgraph/models/symmetry_function_regression.py

+20 −19

Original line number	Diff line number	Diff line
		@@ -94,6 +94,8 @@ class BPSymmetryFunctionRegression(TensorGraph):
		predict=False,
		pad_batches=True):
		for epoch in range(epochs):
		if not predict:
		print('Starting epoch %i' % epoch)
		for (X_b, y_b, w_b, ids_b) in dataset.iterbatches(
		batch_size=self.batch_size,
		deterministic=True,
		@@ -118,8 +120,8 @@ class ANIRegression(TensorGraph):
		def __init__(self,
		n_tasks,
		max_atoms,
		n_feat=1120,
		n_hidden=40,
		n_feat,
		layer_structures=[128, 64],
		atom_number_cases=[1, 6, 7, 8, 16],
		**kwargs):
		"""
		@@ -134,8 +136,8 @@ class ANIRegression(TensorGraph):
		"""
		self.n_tasks = n_tasks
		self.max_atoms = max_atoms
		self.n_hidden = n_hidden
		self.n_feat = n_feat
		self.layer_structures = layer_structures
		self.atom_number_cases = atom_number_cases
		super(ANIRegression, self).__init__(**kwargs)
		self.build_graph()
		@@ -144,26 +146,24 @@ class ANIRegression(TensorGraph):
		self.atom_numbers = Feature(shape=(None, self.max_atoms), dtype=tf.int32)
		self.atom_flags = Feature(shape=(None, self.max_atoms, self.max_atoms))
		self.atom_feats = Feature(shape=(None, self.max_atoms, self.n_feat))
		previous_layer = self.atom_feats

		Hiddens = []
		for n_hidden in self.layer_structures:
		Hidden = AtomicDifferentiatedDense(
		self.max_atoms,
		self.n_hidden,
		n_hidden,
		self.atom_number_cases,
		activation='tanh',
		in_layers=[self.atom_feats, self.atom_numbers])

		Hidden2 = AtomicDifferentiatedDense(
		self.max_atoms,
		self.n_hidden,
		self.atom_number_cases,
		activation='tanh',
		in_layers=[Hidden, self.atom_numbers])
		in_layers=[previous_layer, self.atom_numbers])
		Hiddens.append(Hidden)
		previous_layer = Hiddens[-1]

		costs = []
		self.labels_fd = []
		for task in range(self.n_tasks):
		regression = Dense(
		out_channels=1, activation_fn=None, in_layers=[Hidden2])
		out_channels=1, activation_fn=None, in_layers=[Hiddens[-1]])
		output = BPGather(self.max_atoms, in_layers=[regression, self.atom_flags])
		self.add_output(output)

		@@ -183,6 +183,7 @@ class ANIRegression(TensorGraph):
		predict=False,
		pad_batches=True):
		for epoch in range(epochs):
		if not predict:
		print('Starting epoch %i' % epoch)
		for (X_b, y_b, w_b, ids_b) in dataset.iterbatches(
		batch_size=self.batch_size,

examples/qm7/qm7_ANI.py

+4 −7

Original line number	Diff line number	Diff line
		@@ -18,9 +18,8 @@ train_dataset, valid_dataset, test_dataset = datasets

		# Batch size of models
		max_atoms = 23
		n_hidden = 40
		n_embedding = 0
		batch_size = 16
		layer_structures = [128, 128, 64]
		atom_number_cases = [1, 6, 7, 8, 16]

		ANItransformer = dc.trans.ANITransformer(max_atoms=max_atoms,
		@@ -28,8 +27,6 @@ ANItransformer = dc.trans.ANITransformer(max_atoms=max_atoms,
		train_dataset = ANItransformer.transform(train_dataset)
		valid_dataset = ANItransformer.transform(valid_dataset)
		test_dataset = ANItransformer.transform(test_dataset)

		# The first column is atom numbers
		n_feat = ANItransformer.get_num_feats() - 1

		# Fit models
		@@ -42,8 +39,8 @@ metric = [
		model = dc.models.ANIRegression(
		len(tasks),
		max_atoms,
		n_feat=n_feat,
		n_hidden=n_hidden,
		n_feat,
		layer_structures=layer_structures,
		atom_number_cases=atom_number_cases,
		batch_size=batch_size,
		learning_rate=0.001,
		@@ -51,7 +48,7 @@ model = dc.models.ANIRegression(
		mode="regression")

		# Fit trained model
		model.fit(train_dataset, nb_epoch=100, checkpoint_interval=100)
		model.fit(train_dataset, nb_epoch=3000, checkpoint_interval=100)

		print("Evaluating model")
		train_scores = model.evaluate(train_dataset, metric, transformers)

Admin message