Loading deepchem/models/tensorgraph/sequential.py 0 → 100644 +86 −0 Original line number Diff line number Diff line """ Convenience class for building sequential deep networks. """ from __future__ import print_function from __future__ import division from __future__ import unicode_literals import warnings import tensorflow as tf from deepchem.models.tensorgraph.tensor_graph import TensorGraph from deepchem.models.tensorgraph.layers import Feature from deepchem.models.tensorgraph.layers import Label from deepchem.models.tensorgraph.layers import SoftMaxCrossEntropy from deepchem.models.tensorgraph.layers import ReduceMean class Sequential(TensorGraph): """Sequential models are linear stacks of layers. Analogous to the Sequential model from Keras. """ def __init__(self, **kwargs): """Initializes a sequential model """ self.num_layers = 0 self._prev_layer = None if "use_queue" in kwargs: if kwargs["use_queue"]: raise ValueError("Sequential doesn't support queues.") kwargs["use_queue"] = False self.layer_list = [] super(Sequential, self).__init__(**kwargs) def add(self, layer): """Adds a new layer to model. Parameter --------- layer: Layer Adds layer to this graph. """ self.layer_list.append(layer) def fit(self, dataset, loss, **kwargs): """Fits on the specified dataset. Adds the necessary feature and placeholders. Parameters ---------- dataset: dc.data.Dataset Dataset with data loss: string Only "binary_crossentropy" for now. """ X_shape, y_shape, _, _ = dataset.get_shape() feature_shape = X_shape[1:] label_shape = y_shape[1:] # Add in features features = Feature(shape=(None,) + feature_shape) self._add_layer(features) # Add in labels labels = Label(shape=(None,) + label_shape) self._add_layer(labels) # Add in all layers prev_layer = features for ind, layer in enumerate(self.layer_list): if not len(layer.in_layers) == 0: raise ValueError("Cannot specify in_layers for Sequential.") layer.in_layers += [prev_layer] self._add_layer(layer) prev_layer = layer # The last layer is the output of the model self.outputs.append(prev_layer) if loss == "binary_crossentropy": smce = SoftMaxCrossEntropy(in_layers=[labels, prev_layer]) self._add_layer(smce) self.set_loss(ReduceMean(in_layers=[smce])) else: # TODO(rbharath): Add in support for additional losses. raise ValueError("Unsupported loss.") super(Sequential, self).fit(dataset, **kwargs) Loading
deepchem/models/tensorgraph/sequential.py 0 → 100644 +86 −0 Original line number Diff line number Diff line """ Convenience class for building sequential deep networks. """ from __future__ import print_function from __future__ import division from __future__ import unicode_literals import warnings import tensorflow as tf from deepchem.models.tensorgraph.tensor_graph import TensorGraph from deepchem.models.tensorgraph.layers import Feature from deepchem.models.tensorgraph.layers import Label from deepchem.models.tensorgraph.layers import SoftMaxCrossEntropy from deepchem.models.tensorgraph.layers import ReduceMean class Sequential(TensorGraph): """Sequential models are linear stacks of layers. Analogous to the Sequential model from Keras. """ def __init__(self, **kwargs): """Initializes a sequential model """ self.num_layers = 0 self._prev_layer = None if "use_queue" in kwargs: if kwargs["use_queue"]: raise ValueError("Sequential doesn't support queues.") kwargs["use_queue"] = False self.layer_list = [] super(Sequential, self).__init__(**kwargs) def add(self, layer): """Adds a new layer to model. Parameter --------- layer: Layer Adds layer to this graph. """ self.layer_list.append(layer) def fit(self, dataset, loss, **kwargs): """Fits on the specified dataset. Adds the necessary feature and placeholders. Parameters ---------- dataset: dc.data.Dataset Dataset with data loss: string Only "binary_crossentropy" for now. """ X_shape, y_shape, _, _ = dataset.get_shape() feature_shape = X_shape[1:] label_shape = y_shape[1:] # Add in features features = Feature(shape=(None,) + feature_shape) self._add_layer(features) # Add in labels labels = Label(shape=(None,) + label_shape) self._add_layer(labels) # Add in all layers prev_layer = features for ind, layer in enumerate(self.layer_list): if not len(layer.in_layers) == 0: raise ValueError("Cannot specify in_layers for Sequential.") layer.in_layers += [prev_layer] self._add_layer(layer) prev_layer = layer # The last layer is the output of the model self.outputs.append(prev_layer) if loss == "binary_crossentropy": smce = SoftMaxCrossEntropy(in_layers=[labels, prev_layer]) self._add_layer(smce) self.set_loss(ReduceMean(in_layers=[smce])) else: # TODO(rbharath): Add in support for additional losses. raise ValueError("Unsupported loss.") super(Sequential, self).fit(dataset, **kwargs)
