Loading deepchem/models/__init__.py +1 −1 Original line number Diff line number Diff line Loading @@ -29,4 +29,4 @@ from deepchem.models.tensorflow_models.progressive_joint import ProgressiveJoint from deepchem.models.tensorflow_models.IRV import TensorflowMultiTaskIRVClassifier from deepchem.models.tensorgraph.tensor_graph import TensorGraph from deepchem.models.tensorgraph.models.graph_models import WeaveTensorGraph, DTNNTensorGraph, DAGTensorGraph, GraphConvTensorGraph from deepchem.models.tensorgraph.models.symmetry_function_regression import BPSymmetryFunctionRegression from deepchem.models.tensorgraph.models.symmetry_function_regression import BPSymmetryFunctionRegression, ANIRegression deepchem/models/tensorgraph/models/symmetry_function_regression.py +120 −1 Original line number Diff line number Diff line Loading @@ -13,7 +13,8 @@ from deepchem.models.tensorgraph.layers import L2Loss, Label, Weights, Feature from deepchem.models.tensorgraph.tensor_graph import TensorGraph from deepchem.models.tensorgraph.graph_layers import DTNNEmbedding from deepchem.models.tensorgraph.symmetry_functions import DistanceMatrix, \ DistanceCutoff, RadialSymmetry, AngularSymmetry, BPFeatureMerge, BPGather DistanceCutoff, RadialSymmetry, AngularSymmetry, AngularSymmetryMod, \ BPFeatureMerge, BPGather, AtomicDifferentiatedDense class BPSymmetryFunctionRegression(TensorGraph): Loading Loading @@ -111,3 +112,121 @@ class BPSymmetryFunctionRegression(TensorGraph): feed_dict[self.atom_numbers] = np.array(X_b[:, :, 0], dtype=int) feed_dict[self.atom_coordinates] = np.array(X_b[:, :, 1:], dtype=float) yield feed_dict class ANIRegression(TensorGraph): def __init__(self, n_tasks, max_atoms, n_hidden=40, n_embedding=10, atom_number_cases=[1, 6, 7, 8], **kwargs): """ Parameters ---------- n_tasks: int Number of tasks max_atoms: int Maximum number of atoms in the dataset n_hidden: int, optional Number of hidden units in the readout function """ self.n_tasks = n_tasks self.max_atoms = max_atoms self.n_hidden = n_hidden self.n_embedding = n_embedding self.atom_number_cases = atom_number_cases super(ANIRegression, self).__init__(**kwargs) self.build_graph() def build_graph(self): 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_coordinates = Feature(shape=(None, self.max_atoms, 3)) distance_matrix = DistanceMatrix( self.max_atoms, in_layers=[self.atom_coordinates, self.atom_flags]) distance_cutoff_radial = DistanceCutoff( self.max_atoms, cutoff=4.6 / 0.52917721092, in_layers=[distance_matrix, self.atom_flags]) distance_cutoff_angular = DistanceCutoff( self.max_atoms, cutoff=3.1 / 0.52917721092, in_layers=[distance_matrix, self.atom_flags]) radial_symmetry = RadialSymmetry( self.max_atoms, atomic_number_differentiated=True, atom_numbers=self.atom_number_cases, in_layers=[distance_cutoff_radial, distance_matrix, self.atom_numbers]) angular_symmetry = AngularSymmetryMod( self.max_atoms, atomic_number_differentiated=True, atom_numbers=self.atom_number_cases, in_layers=[distance_cutoff_angular, distance_matrix, self.atom_coordinates, self.atom_numbers]) atom_embedding = DTNNEmbedding( n_embedding=0, in_layers=[self.atom_numbers]) feature_merge = BPFeatureMerge( self.max_atoms, in_layers=[ atom_embedding, radial_symmetry, angular_symmetry, self.atom_flags ]) Hidden = AtomicDifferentiatedDense( self.max_atoms, self.n_hidden, self.atom_number_cases, activation='tanh', in_layers=[feature_merge, self.atom_numbers]) Hidden2 = AtomicDifferentiatedDense( self.max_atoms, self.n_hidden, self.atom_number_cases, activation='tanh', in_layers=[Hidden, self.atom_numbers]) costs = [] self.labels_fd = [] for task in range(self.n_tasks): regression = Dense( out_channels=1, activation_fn=None, in_layers=[Hidden2]) output = BPGather(self.max_atoms, in_layers=[regression, self.atom_flags]) self.add_output(output) label = Label(shape=(None, 1)) self.labels_fd.append(label) cost = L2Loss(in_layers=[label, output]) costs.append(cost) all_cost = Concat(in_layers=costs, axis=0) self.weights = Weights(shape=(None, self.n_tasks)) loss = WeightedError(in_layers=[all_cost, self.weights]) self.set_loss(loss) def default_generator(self, dataset, epochs=1, predict=False, pad_batches=True): for epoch in range(epochs): for (X_b, y_b, w_b, ids_b) in dataset.iterbatches( batch_size=self.batch_size, deterministic=True, pad_batches=pad_batches): feed_dict = dict() if y_b is not None and not predict: for index, label in enumerate(self.labels_fd): feed_dict[label] = y_b[:, index:index + 1] if w_b is not None and not predict: feed_dict[self.weights] = w_b flags = np.sign(np.array(X_b[:, :, 0])) feed_dict[self.atom_flags] = np.stack([flags]*self.max_atoms, axis=2)*\ np.stack([flags]*self.max_atoms, axis=1) feed_dict[self.atom_numbers] = np.array(X_b[:, :, 0], dtype=int) feed_dict[self.atom_coordinates] = np.array(X_b[:, :, 1:], dtype=float) yield feed_dict No newline at end of file deepchem/models/tensorgraph/symmetry_functions.py +42 −2 Original line number Diff line number Diff line Loading @@ -116,7 +116,7 @@ class RadialSymmetry(Layer): Rs_init, ita_init = np.meshgrid(self.Rs_init, self.ita_init) self.Rs = tf.constant(Rs_init.flatten(), dtype=tf.float32) self.ita = tf.constant(ita_init.flatten(), dtype=tf.float32) self.atom_number_embedding = tf.eye(max(self.atom_number_cases)) self.atom_number_embedding = tf.eye(max(self.atom_number_cases)+1) def create_tensor(self, in_layers=None, set_tensors=True, **kwargs): """ Generate Radial Symmetry Function """ Loading Loading @@ -289,7 +289,7 @@ class AngularSymmetryMod(Layer): self.zeta = tf.constant(zeta_init.flatten(), dtype=tf.float32) self.Rs = tf.constant(Rs_init.flatten(), dtype=tf.float32) self.thetas = tf.constant(thetas_init.flatten(), dtype=tf.float32) self.atom_number_embedding = tf.eye(max(self.atom_number_cases)) self.atom_number_embedding = tf.eye(max(self.atom_number_cases)+1) def create_tensor(self, in_layers=None, set_tensors=True, **kwargs): """ Generate Angular Symmetry Function """ Loading Loading @@ -389,3 +389,43 @@ class BPGather(Layer): out_tensor = tf.reduce_sum(out_tensor * flags[:, :, 0:1], axis=1) self.out_tensor = out_tensor class AtomicDifferentiatedDense(Layer): """ Separate Dense module for different atoms """ def __init__(self, max_atoms, out_channels, atom_number_cases=[1, 6, 7, 8], init='glorot_uniform', activation='relu', **kwargs): self.init = initializations.get(init) # Set weight initialization self.activation = activations.get(activation) # Get activations self.max_atoms = max_atoms self.out_channels = out_channels self.atom_number_cases = atom_number_cases super(AtomicDifferentiatedDense, self).__init__(**kwargs) def create_tensor(self, in_layers=None, set_tensors=True, **kwargs): """ Generate Radial Symmetry Function """ if in_layers is None: in_layers = self.in_layers in_layers = convert_to_layers(in_layers) inputs = in_layers[0].out_tensor atom_numbers = in_layers[1].out_tensor in_channels = inputs.get_shape().as_list()[-1] in_dimension = len(inputs.get_shape()) self.W = self.init([max(self.atom_number_cases), in_channels, self.out_channels]) self.b = model_ops.zeros((max(self.atom_number_cases), self.out_channels)) W = tf.nn.embedding_lookup(self.W, atom_numbers) b = tf.nn.embedding_lookup(self.b, atom_numbers) outputs = tf.expand_dims(inputs, in_dimension) * W outputs = tf.reduce_sum(outputs, in_dimension-1) + b outputs = self.activation(outputs) self.out_tensor = outputs examples/qm7/qm7_ANI.py 0 → 100644 +54 −0 Original line number Diff line number Diff line """ Script that trains ANI models on qm7 dataset. """ from __future__ import print_function from __future__ import division from __future__ import unicode_literals import numpy as np np.random.seed(123) import tensorflow as tf tf.set_random_seed(123) import deepchem as dc # Load Tox21 dataset tasks, datasets, transformers = dc.molnet.load_qm7_from_mat( featurizer='BPSymmetryFunction') train_dataset, valid_dataset, test_dataset = datasets # Fit models metric = [ dc.metrics.Metric(dc.metrics.mean_absolute_error, mode="regression"), dc.metrics.Metric(dc.metrics.pearson_r2_score, mode="regression") ] # Batch size of models max_atoms = 23 n_hidden = 40 n_embedding = 0 batch_size = 16 atom_number_cases = [1, 6, 7, 8] model = dc.models.ANIRegression( len(tasks), max_atoms, n_hidden=n_hidden, n_embedding=n_embedding, atom_number_cases=atom_number_cases, batch_size=batch_size, learning_rate=0.001, use_queue=False, mode="regression") # Fit trained model model.fit(train_dataset, nb_epoch=20, checkpoint_interval=1000) print("Evaluating model") train_scores = model.evaluate(train_dataset, metric, transformers) valid_scores = model.evaluate(valid_dataset, metric, transformers) print("Train scores") print(train_scores) print("Validation scores") print(valid_scores) examples/qm7/qm7_BPSymmetryFunction.py +1 −3 Original line number Diff line number Diff line """ Script that trains DTNN models on qm7 dataset. """ from __future__ import print_function from __future__ import division from __future__ import unicode_literals Loading Loading
deepchem/models/__init__.py +1 −1 Original line number Diff line number Diff line Loading @@ -29,4 +29,4 @@ from deepchem.models.tensorflow_models.progressive_joint import ProgressiveJoint from deepchem.models.tensorflow_models.IRV import TensorflowMultiTaskIRVClassifier from deepchem.models.tensorgraph.tensor_graph import TensorGraph from deepchem.models.tensorgraph.models.graph_models import WeaveTensorGraph, DTNNTensorGraph, DAGTensorGraph, GraphConvTensorGraph from deepchem.models.tensorgraph.models.symmetry_function_regression import BPSymmetryFunctionRegression from deepchem.models.tensorgraph.models.symmetry_function_regression import BPSymmetryFunctionRegression, ANIRegression
deepchem/models/tensorgraph/models/symmetry_function_regression.py +120 −1 Original line number Diff line number Diff line Loading @@ -13,7 +13,8 @@ from deepchem.models.tensorgraph.layers import L2Loss, Label, Weights, Feature from deepchem.models.tensorgraph.tensor_graph import TensorGraph from deepchem.models.tensorgraph.graph_layers import DTNNEmbedding from deepchem.models.tensorgraph.symmetry_functions import DistanceMatrix, \ DistanceCutoff, RadialSymmetry, AngularSymmetry, BPFeatureMerge, BPGather DistanceCutoff, RadialSymmetry, AngularSymmetry, AngularSymmetryMod, \ BPFeatureMerge, BPGather, AtomicDifferentiatedDense class BPSymmetryFunctionRegression(TensorGraph): Loading Loading @@ -111,3 +112,121 @@ class BPSymmetryFunctionRegression(TensorGraph): feed_dict[self.atom_numbers] = np.array(X_b[:, :, 0], dtype=int) feed_dict[self.atom_coordinates] = np.array(X_b[:, :, 1:], dtype=float) yield feed_dict class ANIRegression(TensorGraph): def __init__(self, n_tasks, max_atoms, n_hidden=40, n_embedding=10, atom_number_cases=[1, 6, 7, 8], **kwargs): """ Parameters ---------- n_tasks: int Number of tasks max_atoms: int Maximum number of atoms in the dataset n_hidden: int, optional Number of hidden units in the readout function """ self.n_tasks = n_tasks self.max_atoms = max_atoms self.n_hidden = n_hidden self.n_embedding = n_embedding self.atom_number_cases = atom_number_cases super(ANIRegression, self).__init__(**kwargs) self.build_graph() def build_graph(self): 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_coordinates = Feature(shape=(None, self.max_atoms, 3)) distance_matrix = DistanceMatrix( self.max_atoms, in_layers=[self.atom_coordinates, self.atom_flags]) distance_cutoff_radial = DistanceCutoff( self.max_atoms, cutoff=4.6 / 0.52917721092, in_layers=[distance_matrix, self.atom_flags]) distance_cutoff_angular = DistanceCutoff( self.max_atoms, cutoff=3.1 / 0.52917721092, in_layers=[distance_matrix, self.atom_flags]) radial_symmetry = RadialSymmetry( self.max_atoms, atomic_number_differentiated=True, atom_numbers=self.atom_number_cases, in_layers=[distance_cutoff_radial, distance_matrix, self.atom_numbers]) angular_symmetry = AngularSymmetryMod( self.max_atoms, atomic_number_differentiated=True, atom_numbers=self.atom_number_cases, in_layers=[distance_cutoff_angular, distance_matrix, self.atom_coordinates, self.atom_numbers]) atom_embedding = DTNNEmbedding( n_embedding=0, in_layers=[self.atom_numbers]) feature_merge = BPFeatureMerge( self.max_atoms, in_layers=[ atom_embedding, radial_symmetry, angular_symmetry, self.atom_flags ]) Hidden = AtomicDifferentiatedDense( self.max_atoms, self.n_hidden, self.atom_number_cases, activation='tanh', in_layers=[feature_merge, self.atom_numbers]) Hidden2 = AtomicDifferentiatedDense( self.max_atoms, self.n_hidden, self.atom_number_cases, activation='tanh', in_layers=[Hidden, self.atom_numbers]) costs = [] self.labels_fd = [] for task in range(self.n_tasks): regression = Dense( out_channels=1, activation_fn=None, in_layers=[Hidden2]) output = BPGather(self.max_atoms, in_layers=[regression, self.atom_flags]) self.add_output(output) label = Label(shape=(None, 1)) self.labels_fd.append(label) cost = L2Loss(in_layers=[label, output]) costs.append(cost) all_cost = Concat(in_layers=costs, axis=0) self.weights = Weights(shape=(None, self.n_tasks)) loss = WeightedError(in_layers=[all_cost, self.weights]) self.set_loss(loss) def default_generator(self, dataset, epochs=1, predict=False, pad_batches=True): for epoch in range(epochs): for (X_b, y_b, w_b, ids_b) in dataset.iterbatches( batch_size=self.batch_size, deterministic=True, pad_batches=pad_batches): feed_dict = dict() if y_b is not None and not predict: for index, label in enumerate(self.labels_fd): feed_dict[label] = y_b[:, index:index + 1] if w_b is not None and not predict: feed_dict[self.weights] = w_b flags = np.sign(np.array(X_b[:, :, 0])) feed_dict[self.atom_flags] = np.stack([flags]*self.max_atoms, axis=2)*\ np.stack([flags]*self.max_atoms, axis=1) feed_dict[self.atom_numbers] = np.array(X_b[:, :, 0], dtype=int) feed_dict[self.atom_coordinates] = np.array(X_b[:, :, 1:], dtype=float) yield feed_dict No newline at end of file
deepchem/models/tensorgraph/symmetry_functions.py +42 −2 Original line number Diff line number Diff line Loading @@ -116,7 +116,7 @@ class RadialSymmetry(Layer): Rs_init, ita_init = np.meshgrid(self.Rs_init, self.ita_init) self.Rs = tf.constant(Rs_init.flatten(), dtype=tf.float32) self.ita = tf.constant(ita_init.flatten(), dtype=tf.float32) self.atom_number_embedding = tf.eye(max(self.atom_number_cases)) self.atom_number_embedding = tf.eye(max(self.atom_number_cases)+1) def create_tensor(self, in_layers=None, set_tensors=True, **kwargs): """ Generate Radial Symmetry Function """ Loading Loading @@ -289,7 +289,7 @@ class AngularSymmetryMod(Layer): self.zeta = tf.constant(zeta_init.flatten(), dtype=tf.float32) self.Rs = tf.constant(Rs_init.flatten(), dtype=tf.float32) self.thetas = tf.constant(thetas_init.flatten(), dtype=tf.float32) self.atom_number_embedding = tf.eye(max(self.atom_number_cases)) self.atom_number_embedding = tf.eye(max(self.atom_number_cases)+1) def create_tensor(self, in_layers=None, set_tensors=True, **kwargs): """ Generate Angular Symmetry Function """ Loading Loading @@ -389,3 +389,43 @@ class BPGather(Layer): out_tensor = tf.reduce_sum(out_tensor * flags[:, :, 0:1], axis=1) self.out_tensor = out_tensor class AtomicDifferentiatedDense(Layer): """ Separate Dense module for different atoms """ def __init__(self, max_atoms, out_channels, atom_number_cases=[1, 6, 7, 8], init='glorot_uniform', activation='relu', **kwargs): self.init = initializations.get(init) # Set weight initialization self.activation = activations.get(activation) # Get activations self.max_atoms = max_atoms self.out_channels = out_channels self.atom_number_cases = atom_number_cases super(AtomicDifferentiatedDense, self).__init__(**kwargs) def create_tensor(self, in_layers=None, set_tensors=True, **kwargs): """ Generate Radial Symmetry Function """ if in_layers is None: in_layers = self.in_layers in_layers = convert_to_layers(in_layers) inputs = in_layers[0].out_tensor atom_numbers = in_layers[1].out_tensor in_channels = inputs.get_shape().as_list()[-1] in_dimension = len(inputs.get_shape()) self.W = self.init([max(self.atom_number_cases), in_channels, self.out_channels]) self.b = model_ops.zeros((max(self.atom_number_cases), self.out_channels)) W = tf.nn.embedding_lookup(self.W, atom_numbers) b = tf.nn.embedding_lookup(self.b, atom_numbers) outputs = tf.expand_dims(inputs, in_dimension) * W outputs = tf.reduce_sum(outputs, in_dimension-1) + b outputs = self.activation(outputs) self.out_tensor = outputs
examples/qm7/qm7_ANI.py 0 → 100644 +54 −0 Original line number Diff line number Diff line """ Script that trains ANI models on qm7 dataset. """ from __future__ import print_function from __future__ import division from __future__ import unicode_literals import numpy as np np.random.seed(123) import tensorflow as tf tf.set_random_seed(123) import deepchem as dc # Load Tox21 dataset tasks, datasets, transformers = dc.molnet.load_qm7_from_mat( featurizer='BPSymmetryFunction') train_dataset, valid_dataset, test_dataset = datasets # Fit models metric = [ dc.metrics.Metric(dc.metrics.mean_absolute_error, mode="regression"), dc.metrics.Metric(dc.metrics.pearson_r2_score, mode="regression") ] # Batch size of models max_atoms = 23 n_hidden = 40 n_embedding = 0 batch_size = 16 atom_number_cases = [1, 6, 7, 8] model = dc.models.ANIRegression( len(tasks), max_atoms, n_hidden=n_hidden, n_embedding=n_embedding, atom_number_cases=atom_number_cases, batch_size=batch_size, learning_rate=0.001, use_queue=False, mode="regression") # Fit trained model model.fit(train_dataset, nb_epoch=20, checkpoint_interval=1000) print("Evaluating model") train_scores = model.evaluate(train_dataset, metric, transformers) valid_scores = model.evaluate(valid_dataset, metric, transformers) print("Train scores") print(train_scores) print("Validation scores") print(valid_scores)
examples/qm7/qm7_BPSymmetryFunction.py +1 −3 Original line number Diff line number Diff line """ Script that trains DTNN models on qm7 dataset. """ from __future__ import print_function from __future__ import division from __future__ import unicode_literals Loading
