Commit 71aaf8aa authored by Ubuntu's avatar Ubuntu
Browse files

error bars and fixes

parent aeba3e66

ChangeLog

+1 −0
Original line number Diff line number Diff line 
CHANGES

=======



* for now..

* yapfed lots of stuff

* Upgrade Yapf to 0.16.2

* yapf

deepchem/models/tensorgraph/models/graph_models.py

+30 −27
Original line number Diff line number Diff line
@@ -13,6 +13,8 @@ from deepchem.models.tensorgraph.tensor_graph import TensorGraph 
from deepchem.trans import undo_transforms

from deepchem.utils.evaluate import GeneratorEvaluator

from deepchem.data import NumpyDataset

from deepchem.data.data_loader import featurize_smiles_np

from deepchem.feat.graph_features import ConvMolFeaturizer



class WeaveTensorGraph(TensorGraph):
@@ -578,6 +580,7 @@ class GraphConvTensorGraph(TensorGraph): 


    """

    self.n_tasks = n_tasks

    self.error_bars = True if 'error_bars' in kwargs and kwargs['error_bars'] else False    

    kwargs['use_queue'] = False

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

    self.build_graph()
@@ -617,6 +620,7 @@ class GraphConvTensorGraph(TensorGraph): 
        in_layers=[batch_norm3, self.degree_slice, self.membership] +

        self.deg_adjs)



    if self.error_bars == True:

      readout = Dropout(in_layers=[readout], dropout_prob=0.2)



    costs = []
@@ -717,46 +721,45 @@ class GraphConvTensorGraph(TensorGraph): 
        labels=self.my_labels,

        weights=[self.my_task_weights])



  def bayesian_predict(self, dataset, transformers=[], n_passes=4):

    max_index = dataset.shape[0]

    num_batches = max_index // self.batch_size

  def bayesian_predict(self, X, transformers=[], n_passes=4, untransform=False):

    max_index = X.shape[0] - 1

    num_batches = (max_index // self.batch_size) + 1



    mus = []

    sigmas = []

    for i in range(num_batches + 1): # think about edge cases here

    for i in range(num_batches):

      start = i * self.batch_size

      end = min( (i+1)*self.batch_size, max_index)

      batch = dataset[start:end]

      end = min((i+1)*self.batch_size, max_index + 1)

      batch = X[start:end]

      mu, sigma = self.bayesian_predict_on_batch(batch, transformers=[], n_passes=n_passes)

      mus.append(mu)

      sigmas.append(sigma)

    mu = np.concatenate(mus, axis=0)

    sigma = np.concatenate(sigmas, axis=0)

    sigma = np.concatenate(sigmas, axis=0) + 0.55

    

    if untransform:

      mu = undo_transforms(mu, transformers)

      for i in range(sigma.shape[1]):

        sigma[:,i] = sigma[:,i] * transformers[0].y_stds[i]

      

    return mu[:max_index], sigma[:max_index]

    return mu[:max_index + 1], sigma[:max_index + 1]

  

  def predict_on_smiles(self, smiles, transformers):

    max_index = len(smiles)

    num_batches = max_index // self.batch_size

  def predict_on_smiles(self, smiles, transformers=[], untransform=False):

    max_index = len(smiles) - 1

    n_tasks = len(self.outputs)

    num_batches = (max_index // self.batch_size) + 1

    featurizer = ConvMolFeaturizer()

    

    y_ = []

    for i in range(num_batches):

      smiles_batch = smiles[i * self.batch_size:(i + 1) * self.batch_size]

      y_.append(self.predict_on_smiles_batch(smiles_batch, transformers, featurizer))

    smiles_batch = smiles[num_batches * self.batch_size:max_index]

    y_.append(self.predict_on_smiles_batch(smiles_batch, transformers))



    return np.concatenate(y_, axis=1) # wrong axis?



  def predict_on_smiles_batch(self, smiles, transformers=[]):

    convmols = featurize_smiles_np(smiles, featurizer)



    n_smiles = convmols.shape[0]

    n_tasks = len(self.outputs)

      start = i * self.batch_size

      end = min((i+1)*self.batch_size, max_index + 1)

      smiles_batch = smiles[start:end]

      y_.append(self.predict_on_smiles_batch(smiles_batch, featurizer, transformers))

    y_ = np.concatenate(y_, axis=0)[:max_index + 1]

    y_ = y_.reshape(-1, n_tasks)



    dataset = NumpyDataset(X=convmols, y=None, n_tasks=n_tasks)

    generator = self.default_generator(dataset, predict=True, pad_batches=False)

    y_ = self.predict_on_generator(generator, transformers)

    if untransform:

      y_ = undo_transforms(y_, transformers)



    return y_.reshape(-1, n_tasks)[:n_smiles]
 
    return y_

deepchem/models/tensorgraph/tensor_graph.py

+12 −5
Original line number Diff line number Diff line
@@ -15,7 +15,8 @@ from deepchem.models.models import Model 
from deepchem.models.tensorgraph.layers import InputFifoQueue, Label, Feature, Weights

from deepchem.trans import undo_transforms

from deepchem.utils.evaluate import GeneratorEvaluator



from deepchem.feat.graph_features import ConvMolFeaturizer

from deepchem.data.data_loader import featurize_smiles_np



class TensorGraph(Model):
@@ -278,11 +279,11 @@ class TensorGraph(Model): 
          results.append(result)

        return np.concatenate(results, axis=0)



  def bayesian_predict_on_batch(self, X, sess=None, transformers=[], n_passes=4):

  def bayesian_predict_on_batch(self, X, transformers=[], n_passes=4):

    """

    Returns:

      mu: SHAPE

      sigma: SHAPE

      mu: numpy ndarray of shape (n_samples, n_tasks)

      sigma: numpy ndarray of shape (n_samples, n_tasks)

    """

    dataset = NumpyDataset(X=X, y=None, n_tasks=len(self.outputs))

    y_ = []
@@ -296,6 +297,13 @@ class TensorGraph(Model): 
    

    return mu, sigma



  def predict_on_smiles_batch(self, smiles, featurizer, n_tasks, transformers=[]):

    convmols = featurize_smiles_np(smiles, featurizer)



    dataset = NumpyDataset(X=convmols, y=None, n_tasks=len(self.outputs))

    generator = self.default_generator(dataset, predict=True, pad_batches=True)

    return self.predict_on_generator(generator, transformers)



  def predict_on_batch(self, X, sess=None, transformers=[]):

    """Generates output predictions for the input samples,

      processing the samples in a batched way.
@@ -308,7 +316,6 @@ class TensorGraph(Model): 
    # Returns

        A Numpy array of predictions.

    """

    print('inside tensorgraph predict_on_batch')

    dataset = NumpyDataset(X=X, y=None)

    generator = self.default_generator(dataset, predict=True, pad_batches=False)

    return self.predict_on_generator(generator, transformers)

deepchem/trans/transformers.py

+6 −5
Original line number Diff line number Diff line
@@ -113,11 +113,6 @@ class NormalizationTransformer(Transformer): 
               dataset=None,

               transform_gradients=False):

    """Initialize normalization transformation."""

    super(NormalizationTransformer, self).__init__(

        transform_X=transform_X,

        transform_y=transform_y,

        transform_w=transform_w,

        dataset=dataset)

    if transform_X:

      X_means, X_stds = dataset.get_statistics(X_stats=True, y_stats=False)

      self.X_means = X_means
@@ -135,6 +130,12 @@ class NormalizationTransformer(Transformer): 
      self.grad = np.reshape(true_grad, (true_grad.shape[0], -1, 3))

      self.ydely_means = ydely_means



    super(NormalizationTransformer, self).__init__(

              transform_X=transform_X,

              transform_y=transform_y,

              transform_w=transform_w,

              dataset=dataset)

      

  def transform(self, dataset, parallel=False):

    return super(NormalizationTransformer, self).transform(

        dataset, parallel=parallel)

examples/delaney/#delaney_tensorgraph_graphconv.py#

deleted100644 → 0
+0 −51
Original line number Diff line number Diff line 
"""

Script that trains graphconv models on delaney 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



from sklearn.metrics import r2_score



# Load Delaney dataset

delaney_tasks, delaney_datasets, transformers = dc.molnet.load_lipo(

    featurizer='GraphConv', split='scaffold')

train_dataset, valid_dataset, test_dataset = delaney_datasets



# Fit models

metric = dc.metrics.Metric(dc.metrics.pearson_r2_score, np.mean)



n_atom_feat = 75

n_pair_feat = 14

# Batch size of models

batch_size = 48

n_feat = 128



model = dc.models.GraphConvTensorGraph(

    len(delaney_tasks),

    batch_size=batch_size,

    learning_rate=1e-3,

    use_queue=False,

    mode='regression')



for i in xrange(0, 50):

  model.fit(train_dataset, nb_epoch=1)

  valid_scores = model.evaluate(valid_dataset, [metric], transformers)

  mu, sigma = model.bayesian_predict(valid_dataset.X, transformers)

  y = valid_dataset.y

  print(r2_score(y, mu))



  tmp = sigma

  amax = np.amax(tmp.reshape(-1, 1))

  amin = np.amin(tmp.reshape(-1, 1))

  print('max uncrt [%.4f] min uncrt [%.4f]' % (amin, amax))



  mu = mu.reshape(-1, 1).tolist()

  if i, estimate in enumerate(mu):

    
 No newline at end of file

CRA Git | Maintained and supported by SUSTech CRA and CCSE