Loading deepchem/models/keras_model.py +11 −8 Original line number Diff line number Diff line Loading @@ -311,8 +311,6 @@ class KerasModel(Model): the average loss over the most recent checkpoint interval """ self._ensure_built() if restore: self.restore() if checkpoint_interval > 0: manager = tf.train.CheckpointManager(self._checkpoint, self.model_dir, max_checkpoints_to_keep) Loading @@ -324,6 +322,9 @@ class KerasModel(Model): for batch in generator: self._create_training_ops(batch) if restore: self.restore() restore = False inputs, labels, weights = self._prepare_batch(batch) self._tensorboard_step += 1 should_log = ( Loading Loading @@ -469,6 +470,8 @@ class KerasModel(Model): output_values = self._output_functions[outputs](inputs) else: output_values = self.model(inputs, training=False) if isinstance(output_values, tf.Tensor): output_values = [output_values] output_values = [t.numpy() for t in output_values] else: Loading deepchem/models/tests/test_kerasmodel.py +41 −1 Original line number Diff line number Diff line Loading @@ -60,7 +60,7 @@ class TestKerasModel(unittest.TestCase): prediction = np.squeeze(model.predict_on_batch(X)) assert np.array_equal(y, np.round(prediction)) metric = dc.metrics.Metric(dc.metrics.roc_auc_score) generator = model.default_generator(dataset) generator = model.default_generator(dataset, pad_batches=False) scores = model.evaluate_generator(generator, [metric]) assert scores[metric.name] > 0.9 Loading Loading @@ -101,6 +101,46 @@ class TestKerasModel(unittest.TestCase): with context.eager_mode(): self.test_checkpointing() def test_fit_restore(self): """Test specifying restore=True when calling fit().""" n_data_points = 10 n_features = 2 X = np.random.rand(n_data_points, n_features) y = (X[:, 0] > X[:, 1]).astype(np.float32) dataset = dc.data.NumpyDataset(X, y) # Train a model to overfit the dataset. keras_model = tf.keras.Sequential([ tf.keras.layers.Dense(10, activation='relu'), tf.keras.layers.Dense(1, activation='sigmoid') ]) model = dc.models.KerasModel( keras_model, dc.models.losses.BinaryCrossEntropy(), learning_rate=0.005) model.fit(dataset, nb_epoch=1000) prediction = np.squeeze(model.predict_on_batch(X)) assert np.array_equal(y, np.round(prediction)) # Create an identical model, do a single step of fitting with restore=True, # and make sure it got restored correctly. keras_model2 = tf.keras.Sequential([ tf.keras.layers.Dense(10, activation='relu'), tf.keras.layers.Dense(1, activation='sigmoid') ]) model2 = dc.models.KerasModel( keras_model2, dc.models.losses.BinaryCrossEntropy(), model_dir=model.model_dir) model2.fit(dataset, nb_epoch=1, restore=True) prediction = np.squeeze(model2.predict_on_batch(X)) assert np.array_equal(y, np.round(prediction)) def test_fit_restore_eager(self): """Test specifying restore=True when calling fit(), in eager mode.""" with context.eager_mode(): self.test_fit_restore() def test_uncertainty(self): """Test estimating uncertainty a KerasModel.""" n_samples = 30 Loading Loading
deepchem/models/keras_model.py +11 −8 Original line number Diff line number Diff line Loading @@ -311,8 +311,6 @@ class KerasModel(Model): the average loss over the most recent checkpoint interval """ self._ensure_built() if restore: self.restore() if checkpoint_interval > 0: manager = tf.train.CheckpointManager(self._checkpoint, self.model_dir, max_checkpoints_to_keep) Loading @@ -324,6 +322,9 @@ class KerasModel(Model): for batch in generator: self._create_training_ops(batch) if restore: self.restore() restore = False inputs, labels, weights = self._prepare_batch(batch) self._tensorboard_step += 1 should_log = ( Loading Loading @@ -469,6 +470,8 @@ class KerasModel(Model): output_values = self._output_functions[outputs](inputs) else: output_values = self.model(inputs, training=False) if isinstance(output_values, tf.Tensor): output_values = [output_values] output_values = [t.numpy() for t in output_values] else: Loading
deepchem/models/tests/test_kerasmodel.py +41 −1 Original line number Diff line number Diff line Loading @@ -60,7 +60,7 @@ class TestKerasModel(unittest.TestCase): prediction = np.squeeze(model.predict_on_batch(X)) assert np.array_equal(y, np.round(prediction)) metric = dc.metrics.Metric(dc.metrics.roc_auc_score) generator = model.default_generator(dataset) generator = model.default_generator(dataset, pad_batches=False) scores = model.evaluate_generator(generator, [metric]) assert scores[metric.name] > 0.9 Loading Loading @@ -101,6 +101,46 @@ class TestKerasModel(unittest.TestCase): with context.eager_mode(): self.test_checkpointing() def test_fit_restore(self): """Test specifying restore=True when calling fit().""" n_data_points = 10 n_features = 2 X = np.random.rand(n_data_points, n_features) y = (X[:, 0] > X[:, 1]).astype(np.float32) dataset = dc.data.NumpyDataset(X, y) # Train a model to overfit the dataset. keras_model = tf.keras.Sequential([ tf.keras.layers.Dense(10, activation='relu'), tf.keras.layers.Dense(1, activation='sigmoid') ]) model = dc.models.KerasModel( keras_model, dc.models.losses.BinaryCrossEntropy(), learning_rate=0.005) model.fit(dataset, nb_epoch=1000) prediction = np.squeeze(model.predict_on_batch(X)) assert np.array_equal(y, np.round(prediction)) # Create an identical model, do a single step of fitting with restore=True, # and make sure it got restored correctly. keras_model2 = tf.keras.Sequential([ tf.keras.layers.Dense(10, activation='relu'), tf.keras.layers.Dense(1, activation='sigmoid') ]) model2 = dc.models.KerasModel( keras_model2, dc.models.losses.BinaryCrossEntropy(), model_dir=model.model_dir) model2.fit(dataset, nb_epoch=1, restore=True) prediction = np.squeeze(model2.predict_on_batch(X)) assert np.array_equal(y, np.round(prediction)) def test_fit_restore_eager(self): """Test specifying restore=True when calling fit(), in eager mode.""" with context.eager_mode(): self.test_fit_restore() def test_uncertainty(self): """Test estimating uncertainty a KerasModel.""" n_samples = 30 Loading
