Cleaning up dynamic transform on fit/predict (7926492b) · Commits · 钟慕尧 / deepchem

deepchem/models/init.py

+24 −3

Original line number	Diff line number	Diff line
		@@ -147,15 +147,36 @@ class Model(object):
		batch_size = self.model_params["batch_size"]
		for (X_batch, y_batch, w_batch, ids_batch) in dataset.iterbatches(batch_size):

		# Apply fit_transformers if needed
		# HACK(JG): This is a hack to perform 10-fold averaging of y_pred on
		# a given X_batch. If fit_transformers exist, we will apply them to
		# X_batch 10 times and average the resulting y_pred before we undo
		# transforms on y_pred and y

		if self.fit_transformers:
		X_batch, y_batch, w_batch = self.transform_on_batch(X_batch, y_batch,
		w_batch, self.batch_dataset)

		y_preds = []
		for i in xrange(1):
		X_b, y_b, w_b = self.transform_on_batch(X_batch, y_batch, w_batch)
		y_pred = self.predict_on_batch(X_b)
		y_pred = np.reshape(y_pred, np.shape(y_b))
		y_preds.append(y_pred)

		#print(y_batch)
		#print(y_b)
		#print(y_preds)
		y_pred = np.array(y_preds).mean(axis=0)
		#print(y_pred)
		#break

		else:

		#print(y_batch)
		y_pred = self.predict_on_batch(X_batch)
		y_pred = np.reshape(y_pred, np.shape(y_batch))
		#print(y_pred)

		# Now undo transformations on y, y_pred

		y_raw, y_pred_raw = y_batch, y_pred
		y_batch = undo_transforms(y_batch, transformers)
		y_pred = undo_transforms(y_pred, transformers)

deepchem/models/keras_models/fcnet.py

+0 −3

Original line number	Diff line number	Diff line
		@@ -27,9 +27,6 @@ class MultiTaskDNN(KerasModel):
		verbosity=verbosity)
		if initialize_raw_model:
		sorted_tasks = sorted(task_types.keys())
		if fit_transformers:
		(n_inputs,) = model_params["init_data_shape"]
		else:
		(n_inputs,) = model_params["data_shape"]
		model = Graph()
		model.add_input(name="input", input_shape=(n_inputs,))

deepchem/transformers/init.py

+7 −21

Original line number	Diff line number	Diff line
		@@ -39,7 +39,7 @@ class Transformer(object):

		def transform_array(self, X, y, w):
		raise NotImplementedError(
		"Each Transformer is responsible for its own transform_row method.")
		"Each Transformer is responsible for its own transform_array method.")

		def untransform(self, z):
		"""Reverses stored transformation on provided data."""
		@@ -70,14 +70,7 @@ class Transformer(object):
		"""
		Transforms numpy arrays X, y, and w
		"""
		X, y, w = _transform_array(X=X, y=y, w=w, transformer=self)
		return X, y, w

		def _transform_array(X, y, w, transformer):
		"""
		Performs the numpy array transformation with a given transformer
		"""
		X, y, w = transformer.transform_array(X, y, w)
		X, y, w = self.transform_array(X, y, w)
		return X, y, w

		def _transform_row(i, df, transformer):
		@@ -282,7 +275,7 @@ class CoulombBinarizationTransformer(Transformer):

		for _, row in df.iterrows():
		X = load_from_disk(row['X-transformed'])
		self.feature_max = np.maximum(self.feature_max,X.max(axis=0))
		#self.feature_max = np.maximum(self.feature_max,X.max(axis=0))

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

		@@ -304,16 +297,6 @@ class CoulombBinarizationTransformer(Transformer):
		X_t = (Xt[i]-X_means)/X_stds
		save_to_disk(X_t, row['X-transformed'])

		def transform_on_array(self, X, y, w):

		X, y, w = super(CoulombBinarizationTransformer, self).transform_on_array(X, y, w)

		X_means = X.mean(axis=0)
		X_stds = (X-X_means).std()
		X = (X-X_means)/X_stds

		return X, y, w

		def transform_row(self, i, df):
		"""
		Binarizes data in dataset with sigmoid function
		@@ -343,7 +326,7 @@ class CoulombBinarizationTransformer(Transformer):

		X_bin = []
		if self.update_state:
		self.set_max(df)
		#self.set_max(df)
		self.update_state = False
		if self.transform_X:
		for i in range(X.shape[1]):
		@@ -351,6 +334,9 @@ class CoulombBinarizationTransformer(Transformer):
		X_bin += [np.tanh((X[:,i]-k)/self.theta)]

		X = np.array(X_bin).T
		X_means = X.mean(axis=0)
		X_stds = (X-X_means).std()
		X = (X-X_means)/X_stds

		if self.transform_y:
		print("y will not be transformed by CoulombBinarizationTransformer.")

Admin message