Convolutional layers support initializers (640c86bd) · Commits · 钟慕尧 / deepchem

deepchem/models/tensorgraph/layers.py

+62 −5

Original line number	Diff line number	Diff line
		@@ -376,6 +376,8 @@ class Conv1D(Layer):
		stride=1,
		padding='SAME',
		activation_fn=tf.nn.relu,
		biases_initializer=tf.random_normal_initializer,
		weights_initializer=tf.random_normal_initializer,
		**kwargs):
		"""Create a Conv1D layer.

		@@ -391,12 +393,19 @@ class Conv1D(Layer):
		the padding method to use, either 'SAME' or 'VALID'
		activation_fn: object
		the Tensorflow activation function to apply to the output
		biases_initializer: callable object
		the initializer for bias values. This may be None, in which case the layer
		will not include biases.
		weights_initializer: callable object
		the initializer for weight values
		"""
		self.width = width
		self.out_channels = out_channels
		self.stride = stride
		self.padding = padding
		self.activation_fn = activation_fn
		self.weights_initializer = weights_initializer
		self.biases_initializer = biases_initializer
		self.out_tensor = None
		super(Conv1D, self).__init__(**kwargs)
		try:
		@@ -416,11 +425,15 @@ class Conv1D(Layer):
		raise ValueError("Parent tensor must be (batch, width, channel)")
		parent_shape = parent.get_shape()
		parent_channel_size = parent_shape[2].value
		f = tf.Variable(
		tf.random_normal([self.width, parent_channel_size, self.out_channels]))
		b = tf.Variable(tf.random_normal([self.out_channels]))
		f = tf.Variable(self.weights_initializer()
		([self.width, parent_channel_size, self.out_channels]))
		t = tf.nn.conv1d(parent, f, stride=self.stride, padding=self.padding)
		if self.biases_initializer is not None:
		b = tf.Variable(self.biases_initializer()([self.out_channels]))
		t = tf.nn.bias_add(t, b)
		if self.activation_fn is None:
		out_tensor = t
		else:
		out_tensor = self.activation_fn(t)
		if set_tensors:
		self._record_variable_scope(self.name)
		@@ -1464,6 +1477,8 @@ class Conv2D(SharedVariableScope):
		padding='SAME',
		activation_fn=tf.nn.relu,
		normalizer_fn=None,
		biases_initializer=tf.zeros_initializer,
		weights_initializer=tf.contrib.layers.xavier_initializer,
		scope_name=None,
		**kwargs):
		"""Create a Conv2D layer.
		@@ -1486,6 +1501,11 @@ class Conv2D(SharedVariableScope):
		the Tensorflow activation function to apply to the output
		normalizer_fn: object
		the Tensorflow normalizer function to apply to the output
		biases_initializer: callable object
		the initializer for bias values. This may be None, in which case the layer
		will not include biases.
		weights_initializer: callable object
		the initializer for weight values
		"""
		self.num_outputs = num_outputs
		self.kernel_size = kernel_size
		@@ -1493,6 +1513,8 @@ class Conv2D(SharedVariableScope):
		self.padding = padding
		self.activation_fn = activation_fn
		self.normalizer_fn = normalizer_fn
		self.weights_initializer = weights_initializer
		self.biases_initializer = biases_initializer
		super(Conv2D, self).__init__(**kwargs)
		if scope_name is None:
		scope_name = self.name
		@@ -1522,6 +1544,8 @@ class Conv2D(SharedVariableScope):
		padding=self.padding,
		activation_fn=self.activation_fn,
		normalizer_fn=self.normalizer_fn,
		biases_initializer=self.biases_initializer(),
		weights_initializer=self.weights_initializer(),
		scope=self._get_scope_name(),
		reuse=reuse)
		break
		@@ -1552,6 +1576,8 @@ class Conv3D(SharedVariableScope):
		padding='SAME',
		activation_fn=tf.nn.relu,
		normalizer_fn=None,
		biases_initializer=tf.zeros_initializer,
		weights_initializer=tf.contrib.layers.xavier_initializer,
		scope_name=None,
		**kwargs):
		"""Create a Conv3D layer.
		@@ -1574,6 +1600,11 @@ class Conv3D(SharedVariableScope):
		the Tensorflow activation function to apply to the output
		normalizer_fn: object
		the Tensorflow normalizer function to apply to the output
		biases_initializer: callable object
		the initializer for bias values. This may be None, in which case the layer
		will not include biases.
		weights_initializer: callable object
		the initializer for weight values
		"""
		self.num_outputs = num_outputs
		self.kernel_size = kernel_size
		@@ -1581,6 +1612,8 @@ class Conv3D(SharedVariableScope):
		self.padding = padding
		self.activation_fn = activation_fn
		self.normalizer_fn = normalizer_fn
		self.weights_initializer = weights_initializer
		self.biases_initializer = biases_initializer
		super(Conv3D, self).__init__(**kwargs)
		if scope_name is None:
		scope_name = self.name
		@@ -1611,6 +1644,8 @@ class Conv3D(SharedVariableScope):
		padding=self.padding,
		activation=self.activation_fn,
		activity_regularizer=self.normalizer_fn,
		bias_initializer=self.biases_initializer(),
		kernel_initializer=self.weights_initializer(),
		name=self._get_scope_name(),
		reuse=reuse)
		break
		@@ -1642,6 +1677,8 @@ class Conv2DTranspose(SharedVariableScope):
		padding='SAME',
		activation_fn=tf.nn.relu,
		normalizer_fn=None,
		biases_initializer=tf.zeros_initializer,
		weights_initializer=tf.contrib.layers.xavier_initializer,
		scope_name=None,
		**kwargs):
		"""Create a Conv2DTranspose layer.
		@@ -1664,6 +1701,11 @@ class Conv2DTranspose(SharedVariableScope):
		the Tensorflow activation function to apply to the output
		normalizer_fn: object
		the Tensorflow normalizer function to apply to the output
		biases_initializer: callable object
		the initializer for bias values. This may be None, in which case the layer
		will not include biases.
		weights_initializer: callable object
		the initializer for weight values
		"""
		self.num_outputs = num_outputs
		self.kernel_size = kernel_size
		@@ -1671,6 +1713,8 @@ class Conv2DTranspose(SharedVariableScope):
		self.padding = padding
		self.activation_fn = activation_fn
		self.normalizer_fn = normalizer_fn
		self.weights_initializer = weights_initializer
		self.biases_initializer = biases_initializer
		super(Conv2DTranspose, self).__init__(**kwargs)
		if scope_name is None:
		scope_name = self.name
		@@ -1700,6 +1744,8 @@ class Conv2DTranspose(SharedVariableScope):
		padding=self.padding,
		activation_fn=self.activation_fn,
		normalizer_fn=self.normalizer_fn,
		biases_initializer=self.biases_initializer(),
		weights_initializer=self.weights_initializer(),
		scope=self._get_scope_name(),
		reuse=reuse)
		break
		@@ -1730,6 +1776,8 @@ class Conv3DTranspose(SharedVariableScope):
		padding='SAME',
		activation_fn=tf.nn.relu,
		normalizer_fn=None,
		biases_initializer=tf.zeros_initializer,
		weights_initializer=tf.contrib.layers.xavier_initializer,
		scope_name=None,
		**kwargs):
		"""Create a Conv3DTranspose layer.
		@@ -1752,6 +1800,11 @@ class Conv3DTranspose(SharedVariableScope):
		the Tensorflow activation function to apply to the output
		normalizer_fn: object
		the Tensorflow normalizer function to apply to the output
		biases_initializer: callable object
		the initializer for bias values. This may be None, in which case the layer
		will not include biases.
		weights_initializer: callable object
		the initializer for weight values
		"""
		self.num_outputs = num_outputs
		self.kernel_size = kernel_size
		@@ -1759,6 +1812,8 @@ class Conv3DTranspose(SharedVariableScope):
		self.padding = padding
		self.activation_fn = activation_fn
		self.normalizer_fn = normalizer_fn
		self.weights_initializer = weights_initializer
		self.biases_initializer = biases_initializer
		super(Conv3DTranspose, self).__init__(**kwargs)
		if scope_name is None:
		scope_name = self.name
		@@ -1789,6 +1844,8 @@ class Conv3DTranspose(SharedVariableScope):
		padding=self.padding,
		activation=self.activation_fn,
		activity_regularizer=self.normalizer_fn,
		bias_initializer=self.biases_initializer(),
		kernel_initializer=self.weights_initializer(),
		name=self._get_scope_name(),
		reuse=reuse)
		break

Admin message