Accuracy is nonzero now... (83a9f0c7) · Commits · 钟慕尧 / deepchem

deepchem/models/tests/test_overfit.py

+9 −7

Original line number	Diff line number	Diff line
		@@ -673,19 +673,22 @@ class TestOverfitAPI(test_util.TensorFlowTestCase):
		# Load mini log-solubility dataset.
		splittype = "scaffold"
		featurizer = ConvMolFeaturizer()
		tasks = ["log-solubility"]
		task_type = "regression"
		tasks = ["outcome"]
		task_type = "classification"
		task_types = {task: task_type for task in tasks}
		input_file = os.path.join(self.current_dir, "example.csv")
		input_file = os.path.join(self.current_dir, "example_classification.csv")
		loader = DataLoader(tasks=tasks,
		smiles_field=self.smiles_field,
		featurizer=featurizer,
		verbosity="low")
		dataset = loader.featurize(input_file, self.data_dir)
		########################################################### DEBUG
		print("dataset.y")
		print(dataset.y)
		########################################################### DEBUG

		verbosity = "high"
		classification_metric = Metric(metrics.accuracy_score, verbosity=verbosity,
		task_averager=np.mean)
		classification_metric = Metric(metrics.accuracy_score, verbosity=verbosity)

		n_atoms = 50
		n_feat = 71
		@@ -700,11 +703,10 @@ class TestOverfitAPI(test_util.TensorFlowTestCase):
		graph_model.add(GraphGather(batch_size, activation="tanh"))

		with self.test_session() as sess:
		tensorflow_model = MultitaskGraphClassifier(
		model = MultitaskGraphClassifier(
		sess, graph_model, n_tasks, self.model_dir, learning_rate=1e-3,
		learning_rate_decay_time=1000, optimizer_type="adam", beta1=.9,
		beta2=.999, verbosity="high")
		model = TensorflowModel(tensorflow_model, self.model_dir)

		# Fit trained model
		model.fit(dataset)

deepchem/models/tf_keras_models/multitask_classifier.py

+8 −3

Original line number	Diff line number	Diff line
		@@ -9,7 +9,7 @@ from keras import initializations, activations
		from keras import backend as K
		from deepchem.datasets import pad_features
		from deepchem.utils.save import log
		from deepchem.models.tensorflow_models import TensorflowClassifier
		from deepchem.models import Model
		from deepchem.models.tensorflow_models import model_ops
		# TODO(rbharath): Find a way to get rid of this import?
		from deepchem.models.tf_keras_models.graph_topology import merge_dicts
		@@ -43,7 +43,7 @@ def get_loss_fn(final_loss):
		return tf.reduce_sum(weighted_costs)
		return loss_fn

		class MultitaskGraphClassifier(TensorflowClassifier):
		class MultitaskGraphClassifier(Model):

		def __init__(self, sess, model, n_tasks, logdir,
		final_loss='cross_entropy', learning_rate=.001,
		@@ -207,7 +207,7 @@ class MultitaskGraphClassifier(TensorflowClassifier):

		def save(self):
		"""
		No-op since models save themselves during fit()
		No-op since this model doesn't currently support saving...
		"""
		pass

		@@ -239,6 +239,11 @@ class MultitaskGraphClassifier(TensorflowClassifier):
		outputs = np.zeros((n_samples, self.n_tasks))
		for task, output in enumerate(batch_outputs):
		outputs[:, task] = np.argmax(output, axis=1)
		#################################### DEBUG
		print("predict_on_batch")
		print("outputs")
		print(outputs)
		#################################### DEBUG
		return outputs

		def predict_proba_on_batch(self, X, pad_batch=False, n_classes=2):

Admin message