#! /usr/bin/python
# -*- coding: utf-8 -*-
import tensorflow as tf
from tensorlayer.layers.core import Layer
from tensorlayer import logging
from tensorlayer.decorators import deprecated_alias
__all__ = [
'GaussianNoiseLayer',
]
[docs]class GaussianNoiseLayer(Layer):
"""
The :class:`GaussianNoiseLayer` class is noise layer that adding noise with
gaussian distribution to the activation.
Parameters
------------
prev_layer : :class:`Layer`
Previous layer.
mean : float
The mean. Default is 0.
stddev : float
The standard deviation. Default is 1.
is_train : boolean
Is trainable layer. If False, skip this layer. default is True.
seed : int or None
The seed for random noise.
name : str
A unique layer name.
Examples
----------
>>> import tensorflow as tf
>>> import tensorlayer as tl
>>> x = tf.placeholder(tf.float32, shape=(100, 784))
>>> net = tl.layers.InputLayer(x, name='input')
>>> net = tl.layers.DenseLayer(net, n_units=100, act=tf.nn.relu, name='dense3')
>>> net = tl.layers.GaussianNoiseLayer(net, name='gaussian')
(64, 100)
"""
@deprecated_alias(layer='prev_layer', end_support_version=1.9) # TODO remove this line for the 1.9 release
def __init__(
self,
prev_layer,
mean=0.0,
stddev=1.0,
is_train=True,
seed=None,
name='gaussian_noise_layer',
):
super(GaussianNoiseLayer, self).__init__(prev_layer=prev_layer, name=name)
if is_train is False:
logging.info(" skip GaussianNoiseLayer")
self.outputs = prev_layer.outputs
else:
logging.info("GaussianNoiseLayer %s: mean: %f stddev: %f" % (self.name, mean, stddev))
with tf.variable_scope(name):
# noise = np.random.normal(0.0 , sigma , tf.to_int64(self.inputs).get_shape())
noise = tf.random_normal(shape=self.inputs.get_shape(), mean=mean, stddev=stddev, seed=seed)
self.outputs = self.inputs + noise
self._add_layers(self.outputs)