Commit e12af1ec authored by Bharath Ramsundar's avatar Bharath Ramsundar
Browse files

Fixed queue issues (hopefully)

parent dd8ae4e4

deepchem/models/tensorgraph/layers.py

+3 −2
Original line number Diff line number Diff line
@@ -76,8 +76,9 @@ class Layer(object): 
class TensorWrapper(Layer):

  """Used to wrap a tensorflow tensor."""



  def __init__(self, out_tensor):

  def __init__(self, out_tensor, **kwargs):

    self.out_tensor = out_tensor

    super(TensorWrapper, self).__init__(**kwargs)



  def create_tensor(self, in_layers=None):

    """Take no actions."""
@@ -606,7 +607,7 @@ class InputFifoQueue(Layer): 
      self.dtypes = [tf.float32] * len(self.shapes)

    self.queue = tf.FIFOQueue(

        self.capacity, self.dtypes, shapes=self.shapes, names=self.names)

    feed_dict = {x.out_tensor.name: x.out_tensor for x in in_layers}

    feed_dict = {x.name: x.out_tensor for x in in_layers}

    self.out_tensor = self.queue.enqueue(feed_dict)

    self.close_op = self.queue.close()

    self.out_tensors = self.queue.dequeue()

deepchem/models/tensorgraph/models/graph_conv.py

+1 −1
Original line number Diff line number Diff line 
import tensorflow as tf

from deepchem.models.tensorgraph.tensor_graph import TensorGraph

from deepchem.models.tensorgraph.layers import Input, Dense, Concat, SoftMax, SoftMaxCrossEntropy, Layer, \

  GraphConv, BatchNormLayer, GraphPoolLayer, GraphGather, WeightedError

  GraphConv, BatchNorm, GraphPool, GraphGather, WeightedError

from deepchem.metrics import to_one_hot

from deepchem.feat.mol_graphs import ConvMol

import time

deepchem/models/tensorgraph/models/graph_models.py

+3 −3
Original line number Diff line number Diff line
@@ -99,7 +99,7 @@ class WeaveTensorGraph(TensorGraph): 


        label = Label(shape=(None, 1))

        self.labels_fd.append(label)

        cost = L2LossLayer(in_layers=[label, regression])

        cost = L2Loss(in_layers=[label, regression])

        costs.append(cost)



    all_cost = Concat(in_layers=costs)
@@ -283,7 +283,7 @@ class DTNNTensorGraph(TensorGraph): 


      label = Label(shape=(None, 1))

      self.labels_fd.append(label)

      cost = L2LossLayer(in_layers=[label, regression])

      cost = L2Loss(in_layers=[label, regression])

      costs.append(cost)



    all_cost = Concat(in_layers=costs)
@@ -464,7 +464,7 @@ class DAGTensorGraph(TensorGraph): 


        label = Label(shape=(None, 1))

        self.labels_fd.append(label)

        cost = L2LossLayer(in_layers=[label, regression])

        cost = L2Loss(in_layers=[label, regression])

        costs.append(cost)



    all_cost = Concat(in_layers=costs)

deepchem/models/tensorgraph/tests/test_layers.py

+5 −3
Original line number Diff line number Diff line
@@ -38,6 +38,7 @@ from deepchem.models.tensorgraph.layers import SoftMax 
from deepchem.models.tensorgraph.layers import WeightedError

from deepchem.models.tensorgraph.layers import VinaFreeEnergy

from deepchem.models.tensorgraph.layers import WeightedLinearCombo

from deepchem.models.tensorgraph.layers import TensorWrapper



import deepchem as dc
@@ -324,10 +325,11 @@ class TestLayers(test_util.TensorFlowTestCase): 
    batch_size = 10

    n_features = 5

    in_tensor = np.random.rand(batch_size, n_features)

    tf.reset_default_graph()

    with self.test_session() as sess:

      in_tensor = tf.convert_to_tensor(

          in_tensor, dtype=tf.float32, name="input")

      InputFifoQueue([(batch_size, n_features)], ["input:0"])(in_tensor)

      in_tensor = TensorWrapper(

          tf.convert_to_tensor(in_tensor, dtype=tf.float32), name="input")

      InputFifoQueue([(batch_size, n_features)], ["input"])(in_tensor)



  def test_graph_conv(self):

    """Test that GraphConv can be invoked."""

deepchem/models/tensorgraph/tests/test_tensor_graph.py

+36 −41
Original line number Diff line number Diff line
@@ -12,9 +12,6 @@ from deepchem.models.tensorgraph.layers import Feature, Label 
from deepchem.models.tensorgraph.layers import ReduceSquareDifference

from deepchem.models.tensorgraph.tensor_graph import TensorGraph



# TODO(rbharath): Queues were causing strange test failures.

# Turned off all queues for now.





class TestTensorGraph(unittest.TestCase):

  """
@@ -33,7 +30,7 @@ class TestTensorGraph(unittest.TestCase): 
    label = Label(shape=(None, 2))

    smce = SoftMaxCrossEntropy(in_layers=[label, dense])

    loss = ReduceMean(in_layers=[smce])

    tg = dc.models.TensorGraph(learning_rate=0.1, use_queue=False)

    tg = dc.models.TensorGraph(learning_rate=0.1)

    tg.add_output(output)

    tg.set_loss(loss)

    tg.fit(dataset, nb_epoch=10)
@@ -69,7 +66,7 @@ class TestTensorGraph(unittest.TestCase): 


    total_loss = ReduceMean(in_layers=entropies)



    tg = dc.models.TensorGraph(learning_rate=0.1, use_queue=False)

    tg = dc.models.TensorGraph(learning_rate=0.1)

    for output in outputs:

      tg.add_output(output)

    tg.set_loss(total_loss)
@@ -93,7 +90,7 @@ class TestTensorGraph(unittest.TestCase): 
    dense = Dense(out_channels=1, in_layers=[features])

    label = Label(shape=(None, 1))

    loss = ReduceSquareDifference(in_layers=[dense, label])

    tg = dc.models.TensorGraph(learning_rate=0.1, use_queue=False)

    tg = dc.models.TensorGraph(learning_rate=0.1)

    tg.add_output(dense)

    tg.set_loss(loss)

    tg.fit(dataset, nb_epoch=10)
@@ -128,7 +125,7 @@ class TestTensorGraph(unittest.TestCase): 


    total_loss = ReduceMean(in_layers=losses)



    tg = dc.models.TensorGraph(learning_rate=0.1, use_queue=False)

    tg = dc.models.TensorGraph(learning_rate=0.1)

    for output in outputs:

      tg.add_output(output)

    tg.set_loss(total_loss)
@@ -174,7 +171,6 @@ class TestTensorGraph(unittest.TestCase): 
    smce = SoftMaxCrossEntropy(in_layers=[label, dense])

    loss = ReduceMean(in_layers=[smce])

    tg = dc.models.TensorGraph(

        use_queue=False,

        tensorboard=True,

        tensorboard_log_frequency=1,

        learning_rate=0.1,
@@ -201,7 +197,7 @@ class TestTensorGraph(unittest.TestCase): 
    label = Label(shape=(None, 2))

    smce = SoftMaxCrossEntropy(in_layers=[label, dense])

    loss = ReduceMean(in_layers=[smce])

    tg = dc.models.TensorGraph(learning_rate=0.1, use_queue=False)

    tg = dc.models.TensorGraph(learning_rate=0.1)

    tg.add_output(output)

    tg.set_loss(loss)

    tg.fit(dataset, nb_epoch=1)
@@ -212,44 +208,43 @@ class TestTensorGraph(unittest.TestCase): 
    prediction2 = np.squeeze(tg1.predict_proba_on_batch(X))

    assert_true(np.all(np.isclose(prediction, prediction2, atol=0.01)))



  # TODO(rbharath): Failing with strange bugs. Turn test back on!

  #def test_shared_layer(self):

  #  n_data_points = 20

  #  n_features = 2

  def test_shared_layer(self):

    n_data_points = 20

    n_features = 2



  #  X = np.random.rand(n_data_points, n_features)

  #  y1 = np.array([[0, 1] for x in range(n_data_points)])

  #  X = NumpyDataset(X)

  #  ys = [NumpyDataset(y1)]

    X = np.random.rand(n_data_points, n_features)

    y1 = np.array([[0, 1] for x in range(n_data_points)])

    X = NumpyDataset(X)

    ys = [NumpyDataset(y1)]



  #  databag = Databag()

    databag = Databag()



  #  features = Feature(shape=(None, n_features))

  #  databag.add_dataset(features, X)

    features = Feature(shape=(None, n_features))

    databag.add_dataset(features, X)



  #  outputs = []

    outputs = []



  #  label = Label(shape=(None, 2))

  #  dense1 = Dense(out_channels=2, in_layers=[features])

  #  dense2 = dense1.shared(in_layers=[features])

  #  output1 = SoftMax(in_layers=[dense1])

  #  output2 = SoftMax(in_layers=[dense2])

  #  smce = SoftMaxCrossEntropy(in_layers=[label, dense1])

    label = Label(shape=(None, 2))

    dense1 = Dense(out_channels=2, in_layers=[features])

    dense2 = dense1.shared(in_layers=[features])

    output1 = SoftMax(in_layers=[dense1])

    output2 = SoftMax(in_layers=[dense2])

    smce = SoftMaxCrossEntropy(in_layers=[label, dense1])



  #  outputs.append(output1)

  #  outputs.append(output2)

  #  databag.add_dataset(label, ys[0])

    outputs.append(output1)

    outputs.append(output2)

    databag.add_dataset(label, ys[0])



  #  total_loss = ReduceMean(in_layers=[smce])

    total_loss = ReduceMean(in_layers=[smce])



  #  tg = dc.models.TensorGraph(learning_rate=0.1, use_queue=False)

  #  for output in outputs:

  #    tg.add_output(output)

  #  tg.set_loss(total_loss)

    tg = dc.models.TensorGraph(learning_rate=0.1)

    for output in outputs:

      tg.add_output(output)

    tg.set_loss(total_loss)



  #  tg.fit_generator(

  #      databag.iterbatches(

  #          epochs=1, batch_size=tg.batch_size, pad_batches=True))

  #  prediction = tg.predict_proba_on_generator(databag.iterbatches())

  #  assert_true(

  #      np.all(np.isclose(prediction[:, 0], prediction[:, 1], atol=0.01)))
 
    tg.fit_generator(

        databag.iterbatches(

            epochs=1, batch_size=tg.batch_size, pad_batches=True))

    prediction = tg.predict_proba_on_generator(databag.iterbatches())

    assert_true(

        np.all(np.isclose(prediction[:, 0], prediction[:, 1], atol=0.01)))

CRA Git | Maintained and supported by SUSTech CRA and CCSE