Merge pull request #1249 from peastman/multitask (4499cd5a) · Commits · 钟慕尧 / deepchem

deepchem/dock/docking.py

+2 −2

Original line number	Diff line number	Diff line
		@@ -14,7 +14,7 @@ import os
		import tempfile
		from deepchem.data import DiskDataset
		from deepchem.models import SklearnModel
		from deepchem.models import MultiTaskRegressor
		from deepchem.models import MultitaskRegressor
		from deepchem.dock.pose_scoring import GridPoseScorer
		from deepchem.dock.pose_generation import VinaPoseGenerator
		from sklearn.ensemble import RandomForestRegressor
		@@ -91,7 +91,7 @@ class VinaGridDNNDocker(object):
		# Fit model on dataset
		pdbbind_tasks = ["-logKd/Ki"]
		n_features = 2052
		model = MultiTaskRegressor(
		model = MultitaskRegressor(
		len(pdbbind_tasks),
		n_features,
		dropouts=[.25],

deepchem/hyper/tests/test_hyperparam_opt.py

+1 −1

Original line number	Diff line number	Diff line
		@@ -144,7 +144,7 @@ class TestHyperparamOpt(unittest.TestCase):
		params_dict = {"layer_sizes": [(10,), (100,)]}

		def model_builder(model_params, model_dir):
		return dc.models.MultiTaskClassifier(
		return dc.models.MultitaskClassifier(
		len(tasks), n_features, model_dir=model_dir, **model_params)

		optimizer = dc.hyper.HyperparamOpt(model_builder)

deepchem/models/init.py

+4 −4

Original line number	Diff line number	Diff line
		@@ -10,10 +10,10 @@ from deepchem.models.xgboost_models import XGBoostModel
		from deepchem.models.multitask import SingletaskToMultitask

		from deepchem.models.tensorgraph.tensor_graph import TensorGraph
		from deepchem.models.tensorgraph.fcnet import MultiTaskRegressor
		from deepchem.models.tensorgraph.fcnet import MultiTaskClassifier
		from deepchem.models.tensorgraph.fcnet import MultiTaskFitTransformRegressor
		from deepchem.models.tensorgraph.IRV import TensorflowMultiTaskIRVClassifier
		from deepchem.models.tensorgraph.fcnet import MultitaskRegressor
		from deepchem.models.tensorgraph.fcnet import MultitaskClassifier
		from deepchem.models.tensorgraph.fcnet import MultitaskFitTransformRegressor
		from deepchem.models.tensorgraph.IRV import TensorflowMultitaskIRVClassifier
		from deepchem.models.tensorgraph.robust_multitask import RobustMultitaskClassifier
		from deepchem.models.tensorgraph.robust_multitask import RobustMultitaskRegressor
		from deepchem.models.tensorgraph.progressive_multitask import ProgressiveMultitaskRegressor, ProgressiveMultitaskClassifier

deepchem/models/tensorgraph/IRV.py

+5 −5

Original line number	Diff line number	Diff line
		@@ -144,7 +144,7 @@ class Slice(Layer):
		return out_tensor


		class TensorflowMultiTaskIRVClassifier(TensorGraph):
		class TensorflowMultitaskIRVClassifier(TensorGraph):

		def __init__(self,
		n_tasks,
		@@ -152,7 +152,7 @@ class TensorflowMultiTaskIRVClassifier(TensorGraph):
		penalty=0.0,
		mode="classification",
		**kwargs):
		"""Initialize TensorflowMultiTaskIRVClassifier
		"""Initialize TensorflowMultitaskIRVClassifier

		Parameters
		----------
		@@ -169,7 +169,7 @@ class TensorflowMultiTaskIRVClassifier(TensorGraph):
		self.n_features = 2 * self.K * self.n_tasks
		logger.info("n_features after fit_transform: %d" % int(self.n_features))
		self.penalty = penalty
		super(TensorflowMultiTaskIRVClassifier, self).__init__(**kwargs)
		super(TensorflowMultitaskIRVClassifier, self).__init__(**kwargs)
		self.build_graph()

		def build_graph(self):
		@@ -199,12 +199,12 @@ class TensorflowMultiTaskIRVClassifier(TensorGraph):
		self.add_output(outputs)

		def predict(self, dataset, transformers=[], outputs=None):
		out = super(TensorflowMultiTaskIRVClassifier, self).predict(
		out = super(TensorflowMultitaskIRVClassifier, self).predict(
		dataset, transformers=transformers, outputs=outputs)
		out = np.round(out).astype(int)
		return out

		def predict_proba(self, dataset, transformers=[], outputs=None):
		out = super(TensorflowMultiTaskIRVClassifier, self).predict_proba(
		out = super(TensorflowMultitaskIRVClassifier, self).predict_proba(
		dataset, transformers=transformers, outputs=outputs)
		return np.concatenate([1 - out, out], axis=2)

deepchem/models/tensorgraph/fcnet.py

+15 −15

Original line number	Diff line number	Diff line
		@@ -23,7 +23,7 @@ from deepchem.models.tensorgraph.layers import Feature, Label, Weights, Weighted
		logger = logging.getLogger(__name__)


		class MultiTaskClassifier(TensorGraph):
		class MultitaskClassifier(TensorGraph):

		def __init__(self,
		n_tasks,
		@@ -37,7 +37,7 @@ class MultiTaskClassifier(TensorGraph):
		activation_fns=tf.nn.relu,
		n_classes=2,
		**kwargs):
		"""Create a MultiTaskClassifier.
		"""Create a MultitaskClassifier.

		In addition to the following arguments, this class also accepts
		all the keyword arguments from TensorGraph.
		@@ -76,7 +76,7 @@ class MultiTaskClassifier(TensorGraph):
		n_classes: int
		the number of classes
		"""
		super(MultiTaskClassifier, self).__init__(**kwargs)
		super(MultitaskClassifier, self).__init__(**kwargs)
		self.n_tasks = n_tasks
		self.n_features = n_features
		self.n_classes = n_classes
		@@ -169,7 +169,7 @@ class MultiTaskClassifier(TensorGraph):
		return tensors

		def predict_proba(self, dataset, transformers=[], outputs=None):
		return super(MultiTaskClassifier, self).predict(dataset, transformers,
		return super(MultitaskClassifier, self).predict(dataset, transformers,
		outputs)

		def predict(self, dataset, transformers=[], outputs=None):
		@@ -193,13 +193,13 @@ class MultiTaskClassifier(TensorGraph):
		y_pred: numpy ndarray or list of numpy ndarrays
		"""
		# Results is of shape (n_samples, n_tasks, n_classes)
		retval = super(MultiTaskClassifier, self).predict(dataset, transformers,
		retval = super(MultitaskClassifier, self).predict(dataset, transformers,
		outputs)
		# retval is of shape (n_samples, n_tasks)
		return np.argmax(retval, axis=2)


		class MultiTaskRegressor(TensorGraph):
		class MultitaskRegressor(TensorGraph):

		def __init__(self,
		n_tasks,
		@@ -213,7 +213,7 @@ class MultiTaskRegressor(TensorGraph):
		activation_fns=tf.nn.relu,
		uncertainty=False,
		**kwargs):
		"""Create a MultiTaskRegressor.
		"""Create a MultitaskRegressor.

		In addition to the following arguments, this class also accepts all the keywork arguments
		from TensorGraph.
		@@ -249,7 +249,7 @@ class MultiTaskRegressor(TensorGraph):
		if True, include extra outputs and loss terms to enable the uncertainty
		in outputs to be predicted
		"""
		super(MultiTaskRegressor, self).__init__(**kwargs)
		super(MultitaskRegressor, self).__init__(**kwargs)
		self.n_tasks = n_tasks
		self.n_features = n_features
		n_layers = len(layer_sizes)
		@@ -333,8 +333,8 @@ class MultiTaskRegressor(TensorGraph):
		self.set_loss(weighted_loss)


		class MultiTaskFitTransformRegressor(MultiTaskRegressor):
		"""Implements a MultiTaskRegressor that performs on-the-fly transformation during fit/predict.
		class MultitaskFitTransformRegressor(MultitaskRegressor):
		"""Implements a MultitaskRegressor that performs on-the-fly transformation during fit/predict.

		Example:

		@@ -347,7 +347,7 @@ class MultiTaskFitTransformRegressor(MultiTaskRegressor):
		>>> w = np.ones((n_samples, n_tasks))
		>>> dataset = dc.data.NumpyDataset(X, y, w, ids)
		>>> fit_transformers = [dc.trans.CoulombFitTransformer(dataset)]
		>>> model = dc.models.MultiTaskFitTransformRegressor(n_tasks, [n_features, n_features],
		>>> model = dc.models.MultitaskFitTransformRegressor(n_tasks, [n_features, n_features],
		... dropouts=[0.], learning_rate=0.003, weight_init_stddevs=[np.sqrt(6)/np.sqrt(1000)],
		... batch_size=n_samples, fit_transformers=fit_transformers, n_evals=1)
		>>> model.n_features
		@@ -361,10 +361,10 @@ class MultiTaskFitTransformRegressor(MultiTaskRegressor):
		n_evals=1,
		batch_size=50,
		**kwargs):
		"""Create a MultiTaskFitTransformRegressor.
		"""Create a MultitaskFitTransformRegressor.

		In addition to the following arguments, this class also accepts all the keywork arguments
		from MultiTaskRegressor.
		from MultitaskRegressor.

		Parameters
		----------
		@@ -392,7 +392,7 @@ class MultiTaskFitTransformRegressor(MultiTaskRegressor):
		X_b = transformer.X_transform(X_b)
		n_features = X_b.shape[1]
		logger.info("n_features after fit_transform: %d", int(n_features))
		super(MultiTaskFitTransformRegressor, self).__init__(
		super(MultitaskFitTransformRegressor, self).__init__(
		n_tasks, n_features, batch_size=batch_size, **kwargs)

		def default_generator(self,
		@@ -430,5 +430,5 @@ class MultiTaskFitTransformRegressor(MultiTaskRegressor):
		feed_dict[self.features[0]] = X_t
		yield feed_dict

		return super(MultiTaskFitTransformRegressor, self).predict_on_generator(
		return super(MultitaskFitTransformRegressor, self).predict_on_generator(
		transform_generator(), transformers, outputs)

Admin message