Cleanup (be650242) · Commits · 钟慕尧 / deepchem

deepchem/data/datasets.py

+5 −0

Original line number	Diff line number	Diff line
		@@ -162,8 +162,10 @@ class Dataset(object):
		"""Get an object that iterates over the samples in the dataset.

		Example:

		>>> for x, y, w, id in dataset.itersamples():
		>>> print(x, y, w, id)

		"""
		raise NotImplementedError()

		@@ -317,8 +319,10 @@ class NumpyDataset(Dataset):
		"""Get an object that iterates over the samples in the dataset.

		Example:

		>>> for x, y, w, id in dataset.itersamples():
		>>> print(x, y, w, id)

		"""
		n_samples = self._X.shape[0]
		return ((self._X[i], self._y[i], self._w[i], self._ids[i])
		@@ -580,6 +584,7 @@ class DiskDataset(Dataset):
		"""Get an object that iterates over the samples in the dataset.

		Example:

		>>> for x, y, w, id in dataset.itersamples():
		>>> print(x, y, w, id)
		"""

deepchem/nn/copy.py

+37 −29

Original line number	Diff line number	Diff line
		@@ -492,7 +492,8 @@ class Layer(object):
		an input shape (assumes that the layer will be built
		to match that input shape).

		# Arguments
		Parameters
		----------
		input_shape: Shape tuple (tuple of integers)
		or list of shape tuples (one per output tensor of the layer).
		Shape tuples can include None for free dimensions,
		@@ -504,7 +505,8 @@ class Layer(object):
		"""Creates the layer weights.
		Must be implemented on all layers that have weights.

		# Arguments
		Parameters
		----------
		input_shape: tensor (future input to layer)
		or list/tuple of tensors to reference
		for weight shape computations.
		@@ -514,7 +516,8 @@ class Layer(object):
		def _get_node_attribute_at_index(self, node_index, attr, attr_name):
		"""Retrieves an attribute (e.g. input_tensors) from a node.

		# Arguments
		Parameters
		----------
		node_index: Integer index of the node from which
		to retrieve the attribute.
		attr: Exact node attribute name.
		@@ -735,7 +738,9 @@ class InputLayer(Layer):
		It can either wrap an existing tensor (pass an `input_tensor` argument)
		or create its a placeholder tensor (pass arguments `input_shape`
		or `batch_input_shape` as well as `input_dtype`).
		# Arguments

		Parameters
		----------
		input_shape: Shape tuple, not including the batch axis.
		batch_input_shape: Shape tuple, including the batch axis.
		input_dtype: Datatype of the input.
		@@ -832,7 +837,9 @@ def Input(shape=None, batch_shape=None,
		._keras_history: Last layer applied to the tensor.
		the entire layer graph is retrievable from that layer,
		recursively.
		# Arguments

		Parameters
		----------
		shape: A shape tuple (integer), not including the batch size.
		For instance, `shape=(32,)` indicates that the expected input
		will be batches of 32-dimensional vectors.
		@@ -846,13 +853,14 @@ def Input(shape=None, batch_shape=None,
		It will be autogenerated if it isn't provided.
		dtype: The data type expected by the input, as a string
		(`float32`, `float64`, `int32`...)
		# Example
		```python
		# this is a logistic regression in Keras
		a = Input(shape=(32,))
		b = Dense(16, activation='softmax')(a)
		model = Model(input=a, output=b)
		```

		# TODO(rbharath): Support this type of functional API.
		Example:

		>>> # this is a logistic regression in Keras
		>>> a = Input(shape=(32,))
		>>> b = Dense(16, activation='softmax')(a)
		>>> model = Model(input=a, output=b)
		"""
		if not batch_shape and tensor is None:
		assert shape, ('Please provide to Input either a `shape`'
		@@ -875,23 +883,23 @@ def Input(shape=None, batch_shape=None,
		class Dense(Layer):
		"""Just your regular densely-connected NN layer.

		# Example
		TODO(rbharath): Make this functional in deepchem

		```python
		# as first layer in a sequential model:
		model = Sequential()
		model.add(Dense(32, input_dim=16))
		# now the model will take as input arrays of shape (*, 16)
		# and output arrays of shape (*, 32)
		Example:

		# this is equivalent to the above:
		model = Sequential()
		model.add(Dense(32, input_shape=(16,)))
		>>> # as first layer in a sequential model:
		>>> model = Sequential()
		>>> model.add(Dense(32, input_dim=16))
		>>> # now the model will take as input arrays of shape (*, 16)
		>>> # and output arrays of shape (*, 32)

		# after the first layer, you don't need to specify
		# the size of the input anymore:
		model.add(Dense(32))
		```
		>>> # this is equivalent to the above:
		>>> model = Sequential()
		>>> model.add(Dense(32, input_shape=(16,)))

		>>> # after the first layer, you don't need to specify
		>>> # the size of the input anymore:
		>>> model.add(Dense(32))

		Parameters
		----------
		@@ -1056,15 +1064,15 @@ class BatchNormalization(Layer):
		beta_regularizer: instance of [WeightRegularizer](../regularizers.md),
		applied to the beta vector.

		# Input shape
		Input shape:
		Arbitrary. Use the keyword argument `input_shape`
		(tuple of integers, does not include the samples axis)
		when using this layer as the first layer in a model.

		# Output shape
		Output shape:
		Same shape as input.

		# References
		References:
		- [Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift](https://arxiv.org/abs/1502.03167)
		"""

deepchem/nn/model_ops.py

+74 −73

Original line number	Diff line number	Diff line
		"""Ops for graph construction.

		Large amounts of code borrowed from Keras.
		Large amounts of code borrowed from Keras. Will try to incorporate into
		DeepChem properly.
		"""
		from __future__ import print_function
		from __future__ import division
		@@ -23,27 +24,6 @@ _UID_PREFIXES = defaultdict(int)
		# either train mode (learning_phase == 1) or test mode (learning_phase == 0).
		_GRAPH_LEARNING_PHASES = {}

		# TODO(rbharath): Can this be improved.
		def _convert_string_dtype(dtype):
		if dtype == 'float16':
		return tf.float16
		if dtype == 'float32':
		return tf.float32
		elif dtype == 'float64':
		return tf.float64
		elif dtype == 'int16':
		return tf.int16
		elif dtype == 'int32':
		return tf.int32
		elif dtype == 'int64':
		return tf.int64
		elif dtype == 'uint8':
		return tf.int8
		elif dtype == 'uint16':
		return tf.uint16
		else:
		raise ValueError('Unsupported dtype:', dtype)

		def _to_tensor(x, dtype):
		x = tf.convert_to_tensor(x)
		if x.dtype != dtype:
		@@ -59,8 +39,7 @@ def learning_phase():
		"""
		graph = tf.get_default_graph()
		if graph not in _GRAPH_LEARNING_PHASES:
		phase = tf.placeholder(dtype='bool',
		name='keras_learning_phase')
		phase = tf.placeholder(dtype='bool', name='keras_learning_phase')
		_GRAPH_LEARNING_PHASES[graph] = phase
		return _GRAPH_LEARNING_PHASES[graph]

		@@ -68,7 +47,8 @@ def in_train_phase(x, alt):
		"""Selects `x` in train phase, and `alt` otherwise.
		Note that `alt` should have the same shape as `x`.

		# Returns
		Returns
		-------
		Either `x` or `alt` based on `K.learning_phase`.
		"""
		if learning_phase() is 1:
		@@ -150,7 +130,7 @@ def ones(shape, dtype=None, name=None):
		Parameters
		----------
		shape: Tuple of integers, shape of returned Keras variable.
		dtype: String, data type of returned Keras variable.
		dtype: Tensorflow dtype
		name: String, name of returned Keras variable.

		Returns
		@@ -160,8 +140,7 @@ def ones(shape, dtype=None, name=None):
		if dtype is None:
		dtype = tf.float32
		shape = tuple(map(int, shape))
		tf_dtype = _convert_string_dtype(dtype)
		return variable(tf.constant_initializer(1., dtype=tf_dtype)(shape),
		return variable(tf.constant_initializer(1., dtype=dtype)(shape),
		dtype, name)

		def cast_to_floatx(x):
		@@ -309,10 +288,12 @@ def dot(x, y):
		def get_ndim(x):
		"""Returns the number of axes in a tensor, as an integer.

		# Arguments
		Parameters
		----------
		x: Tensor or variable.

		# Returns
		Returns
		-------
		Integer (scalar), number of axes.
		"""
		dims = x.get_shape()._dims
		@@ -323,10 +304,12 @@ def get_ndim(x):
		def get_dtype(x):
		"""Returns the dtype of a Keras tensor or variable, as a string.

		# Arguments
		Parameters
		----------
		x: Tensor or variable.

		# Returns
		Returns
		-------
		String, dtype of `x`.
		"""
		return x.dtype.name
		@@ -348,7 +331,8 @@ def epsilon():
		"""Returns the value of the fuzz
		factor used in numeric expressions.

		# Returns
		Returns
		-------
		A float.
		"""
		return 1e-7
		@@ -356,15 +340,17 @@ def epsilon():
		def variable(value, dtype=tf.float32, name=None):
		"""Instantiates a variable and returns it.

		# Arguments
		Parameters
		----------
		value: Numpy array, initial value of the tensor.
		dtype: Tensor type.
		name: Optional name string for the tensor.

		# Returns
		Returns
		-------
		A variable instance (with Keras metadata included).
		"""
		v = tf.Variable(value, dtype=_convert_string_dtype(dtype), name=name)
		v = tf.Variable(value, dtype=dtype, name=name)
		if hasattr(value, 'get_shape'):
		v._keras_shape = tuple(map(int, value.get_shape()))
		v._uses_learning_phase = False
		@@ -375,24 +361,25 @@ def random_uniform_variable(shape, low, high, dtype=tf.float32,
		"""Instantiates an Keras variable filled with
		samples drawn from a uniform distribution and returns it.

		# Arguments
		Parameters
		----------
		shape: Tuple of integers, shape of returned Keras variable.
		low: Float, lower boundary of the output inteval.
		high: Float, upper boundary of the output interval.
		dtype: String, dtype of returned Keras variable.
		dtype: Tensorflow dtype
		name: String, name of returned Keras variable.
		seed: Integer, random seed.

		# Returns
		A Keras variable, filled with drawn samples.
		Returns
		-------
		A tf.Variable, filled with drawn samples.
		"""
		shape = tuple(map(int, shape))
		tf_dtype = _convert_string_dtype(dtype)
		if seed is None:
		# ensure that randomness is conditioned by the Numpy RNG
		seed = np.random.randint(10e8)
		value = tf.random_uniform_initializer(
		low, high, dtype=tf_dtype, seed=seed)(shape)
		low, high, dtype=dtype, seed=seed)(shape)
		return variable(value, dtype=dtype, name=name)

		def random_normal_variable(shape, mean, scale, dtype=tf.float32,
		@@ -400,30 +387,32 @@ def random_normal_variable(shape, mean, scale, dtype=tf.float32,
		"""Instantiates an Keras variable filled with
		samples drawn from a normal distribution and returns it.

		# Arguments
		Parameters
		----------
		shape: Tuple of integers, shape of returned Keras variable.
		mean: Float, mean of the normal distribution.
		scale: Float, standard deviation of the normal distribution.
		dtype: String, dtype of returned Keras variable.
		dtype: Tensorflow dtype
		name: String, name of returned Keras variable.
		seed: Integer, random seed.

		# Returns
		A Keras variable, filled with drawn samples.
		Returns
		-------
		A tf.Variable, filled with drawn samples.
		"""
		shape = tuple(map(int, shape))
		tf_dtype = _convert_string_dtype(dtype)
		if seed is None:
		# ensure that randomness is conditioned by the Numpy RNG
		seed = np.random.randint(10e8)
		value = tf.random_normal_initializer(
		mean, scale, dtype=tf_dtype, seed=seed)(shape)
		mean, scale, dtype=dtype, seed=seed)(shape)
		return variable(value, dtype=dtype, name=name)

		def max(x, axis=None, keepdims=False):
		"""Maximum value in a tensor.

		# Arguments
		Parameters
		----------
		x: A tensor or variable.
		axis: An integer, the axis to find maximum values.
		keepdims: A boolean, whether to keep the dimensions or not.
		@@ -431,7 +420,8 @@ def max(x, axis=None, keepdims=False):
		by 1. If `keepdims` is `True`,
		the reduced dimension is retained with length 1.

		# Returns
		Returns
		-------
		A tensor with maximum values of `x`.
		"""
		axis = _normalize_axis(axis, get_ndim(x))
		@@ -440,16 +430,18 @@ def max(x, axis=None, keepdims=False):
		def sum(x, axis=None, keepdims=False):
		"""Sum of the values in a tensor, alongside the specified axis.

		# Arguments
		Parameters
		----------
		x: A tensor or variable.
		axis: An integer, the axis to sum over.
		keepdims: A boolean, whether to keep the dimensions or not.
		If `keepdims` is `False`, the rank of the tensor is reduced
		by 1. If `keepdims` is `True`,
		If keepdims is False, the rank of the tensor is reduced
		by 1. If keepdims is True,
		the reduced dimension is retained with length 1.

		# Returns
		A tensor with sum of `x`.
		Returns
		-------
		A tensor with sum of x.
		"""
		axis = _normalize_axis(axis, get_ndim(x))
		return tf.reduce_sum(x, reduction_indices=axis, keep_dims=keepdims)
		@@ -462,7 +454,7 @@ def zeros(shape, dtype=tf.float32, name=None):
		Parameters
		----------
		shape: Tuple of integers, shape of returned Keras variable
		dtype: String, data type of returned Keras variable
		dtype: Tensorflow dtype
		name: String, name of returned Keras variable

		Returns
		@@ -470,8 +462,7 @@ def zeros(shape, dtype=tf.float32, name=None):
		A variable (including Keras metadata), filled with `0.0`.
		"""
		shape = tuple(map(int, shape))
		tf_dtype = _convert_string_dtype(dtype)
		return variable(tf.constant_initializer(0., dtype=tf_dtype)(shape),
		return variable(tf.constant_initializer(0., dtype=dtype)(shape),
		dtype, name)

		def cosine_distances(test, support):
		@@ -504,11 +495,13 @@ def cosine_distances(test, support):
		def elu(x, alpha=1.):
		"""Exponential linear unit.

		# Arguments
		x: A tenor or variable to compute the activation function for.
		Parameters
		----------
		x: A tensor or variable to compute the activation function for.
		alpha: A scalar, slope of positive section.

		# Returns
		Returns
		-------
		A tensor.
		"""
		res = tf.nn.elu(x)
		@@ -521,12 +514,14 @@ def relu(x, alpha=0., max_value=None):
		"""Rectified linear unit.
		With default values, it returns element-wise `max(x, 0)`.

		# Arguments
		Parameters
		----------
		x: A tensor or variable.
		alpha: A scalar, slope of negative section (default=`0.`).
		max_value: Saturation threshold.

		# Returns
		Returns
		-------
		A tensor.
		"""
		if alpha != 0.:
		@@ -544,13 +539,15 @@ def relu(x, alpha=0., max_value=None):
		def hard_sigmoid(x):
		"""Segment-wise linear approximation of sigmoid.
		Faster than sigmoid.
		Returns `0.` if `x < -2.5`, `1.` if `x > 2.5`.
		In `-2.5 <= x <= 2.5`, returns `0.2 * x + 0.5`.
		Returns 0. if x < -2.5, 1. if x > 2.5.
		In -2.5 <= x <= 2.5, returns 0.2 * x + 0.5.

		# Arguments
		Parameters
		----------
		x: A tensor or variable.

		# Returns
		Returns
		-------
		A tensor.
		"""
		x = (0.2 * x) + 0.5
		@@ -562,10 +559,12 @@ def hard_sigmoid(x):
		def sqrt(x):
		"""Element-wise square root.

		# Arguments
		Parameters
		----------
		x: input tensor.

		# Returns
		Returns
		-------
		A tensor.
		"""
		zero = _to_tensor(0., x.dtype.base_dtype)
		@@ -576,15 +575,17 @@ def sqrt(x):
		def var(x, axis=None, keepdims=False):
		"""Variance of a tensor, alongside the specified axis.

		# Arguments
		Parameters
		----------
		x: A tensor or variable.
		axis: An integer, the axis to compute the variance.
		keepdims: A boolean, whether to keep the dimensions or not.
		If `keepdims` is `False`, the rank of the tensor is reduced
		by 1. If `keepdims` is `True`,
		If keepdims is False, the rank of the tensor is reduced
		by 1. If keepdims is True,
		the reduced dimension is retained with length 1.

		# Returns
		Returns
		-------
		A tensor with the variance of elements of `x`.
		"""
		axis = _normalize_axis(axis, get_ndim(x))

Admin message