def batch_norm(inputs,
decay=0.999,
center=True,
scale=False,
epsilon=0.001,
activation_fn=None,
param_initializers=None,
param_regularizers=None,
updates_collections=ops.GraphKeys.UPDATE_OPS,
is_training=True,
reuse=None,
variables_collections=None,
outputs_collections=None,
trainable=True,
batch_weights=None,
fused=False,
data_format=DATA_FORMAT_NHWC,
zero_debias_moving_mean=False,
scope=None,
renorm=False,
renorm_clipping=None,
renorm_decay=0.99):
scope = _default_scope(scope, 'BATCH_NORM', 'BatchNorm')
if data_format not in (DATA_FORMAT_NCHW, DATA_FORMAT_NHWC):
raise ValueError('data_format has to be either NCHW or NHWC.')
axis = 1 if data_format == DATA_FORMAT_NCHW else -1
with vs.variable_scope(scope, reuse=reuse) as sc:
if not param_initializers:
param_initializers = {}
beta_initializer = param_initializers.get('beta', init_ops.zeros_initializer())
gamma_initializer = param_initializers.get('gamma', init_ops.ones_initializer())
moving_mean_initializer = param_initializers.get('moving_mean', init_ops.zeros_initializer())
moving_variance_initializer = param_initializers.get('moving_variance', init_ops.ones_initializer())
if not param_regularizers:
param_regularizers = {}
beta_regularizer = param_regularizers.get('beta')
gamma_regularizer = param_regularizers.get('gamma')
return layers.batch_norm(
inputs=inputs,
axis=axis,
momentum=decay,
epsilon=epsilon,
center=center,
scale=scale,
beta_initializer=beta_initializer,
gamma_initializer=gamma_initializer,
moving_mean_initializer=moving_mean_initializer,
moving_variance_initializer=moving_variance_initializer,
beta_regularizer=beta_regularizer,
gamma_regularizer=gamma_regularizer,
trainable=trainable,
renorm=renorm,
renorm_clipping=renorm_clipping,
renorm_momentum=renorm_decay,
fused=fused,
training=is_training)
def batch_norm(inputs,
decay=0.999,
center=True,
scale=False,
epsilon=0.001,
activation_fn=None,
param_initializers=None,
param_regularizers=None,
updates_collections=ops.GraphKeys.UPDATE_OPS,
is_training=True,
reuse=None,
variables_collections=None,
outputs_collections=None,
trainable=True,
batch_weights=None,
fused=False,
data_format=DATA_FORMAT_NHWC,
zero_debias_moving_mean=False,
scope=None,
renorm=False,
renorm_clipping=None,
renorm_decay=0.99):
scope = _default_scope(scope, 'BATCH_NORM', 'BatchNorm')
if data_format not in (DATA_FORMAT_NCHW, DATA_FORMAT_NHWC):
raise ValueError('data_format has to be either NCHW or NHWC.')
axis = 1 if data_format == DATA_FORMAT_NCHW else -1
with vs.variable_scope(scope, reuse=reuse) as sc:
if not param_initializers:
param_initializers = {}
beta_initializer = param_initializers.get('beta', init_ops.zeros_initializer())
gamma_initializer = param_initializers.get('gamma', init_ops.ones_initializer())
moving_mean_initializer = param_initializers.get('moving_mean', init_ops.zeros_initializer())
moving_variance_initializer = param_initializers.get('moving_variance', init_ops.ones_initializer())
if not param_regularizers:
param_regularizers = {}
beta_regularizer = param_regularizers.get('beta')
gamma_regularizer = param_regularizers.get('gamma')
return layers.batch_norm(
inputs=inputs,
axis=axis,
momentum=decay,
epsilon=epsilon,
center=center,
scale=scale,
beta_initializer=beta_initializer,
gamma_initializer=gamma_initializer,
moving_mean_initializer=moving_mean_initializer,
moving_variance_initializer=moving_variance_initializer,
beta_regularizer=beta_regularizer,
gamma_regularizer=gamma_regularizer,
trainable=trainable,
renorm=renorm,
renorm_clipping=renorm_clipping,
renorm_momentum=renorm_decay,
fused=fused,
training=is_training)
def __init__(self,
axis=-1,
momentum=0.99,
epsilon=1e-3,
center=True,
scale=True,
beta_initializer=init_ops.zeros_initializer(),
gamma_initializer=init_ops.ones_initializer(),
moving_mean_initializer=init_ops.zeros_initializer(),
moving_variance_initializer=init_ops.ones_initializer(),
beta_regularizer=None,
gamma_regularizer=None,
renorm=False,
renorm_clipping=None,
renorm_momentum=0.99,
fused=False,
trainable=True,
name=None,
**kwargs):
super(BatchNormalization, self).__init__(trainable=trainable,
name=name,
**kwargs)
self.axis = axis
self.momentum = momentum
self.epsilon = epsilon
self.center = center
self.scale = scale
self.beta_initializer = beta_initializer
self.gamma_initializer = gamma_initializer
self.moving_mean_initializer = moving_mean_initializer
self.moving_variance_initializer = moving_variance_initializer
self.beta_regularizer = beta_regularizer
self.gamma_regularizer = gamma_regularizer
self.renorm = renorm
self.fused = fused
self.trainable = trainable
if fused:
if not center or not scale:
raise ValueError(
'fused norm requires both center and scale set to be True.'
)
if renorm:
raise ValueError('renorm is currently not supported.')
def batch_normalization(
inputs,
axis=-1,
momentum=0.99,
epsilon=1e-3,
center=True,
scale=True,
beta_initializer=init_ops.zeros_initializer(),
gamma_initializer=init_ops.ones_initializer(),
moving_mean_initializer=init_ops.zeros_initializer(),
moving_variance_initializer=init_ops.ones_initializer(),
beta_regularizer=None,
gamma_regularizer=None,
training=False,
trainable=True,
name=None,
reuse=None,
renorm=False,
renorm_clipping=None,
renorm_momentum=0.99,
fused=False):
layer = BatchNormalization(
axis=axis,
momentum=momentum,
epsilon=epsilon,
center=center,
scale=scale,
beta_initializer=beta_initializer,
gamma_initializer=gamma_initializer,
moving_mean_initializer=moving_mean_initializer,
moving_variance_initializer=moving_variance_initializer,
beta_regularizer=beta_regularizer,
gamma_regularizer=gamma_regularizer,
renorm=renorm,
renorm_clipping=renorm_clipping,
renorm_momentum=renorm_momentum,
fused=fused,
trainable=trainable,
name=name,
_reuse=reuse,
_scope=name)
return layer.apply(inputs, training=training)
def batch_normalization(inputs,
axis=-1,
momentum=0.99,
epsilon=1e-3,
center=True,
scale=True,
beta_initializer=init_ops.zeros_initializer(),
gamma_initializer=init_ops.ones_initializer(),
moving_mean_initializer=init_ops.zeros_initializer(),
moving_variance_initializer=init_ops.ones_initializer(),
beta_regularizer=None,
gamma_regularizer=None,
training=False,
trainable=True,
name=None,
reuse=None,
renorm=False,
renorm_clipping=None,
renorm_momentum=0.99,
fused=False):
layer = BatchNormalization(
axis=axis,
momentum=momentum,
epsilon=epsilon,
center=center,
scale=scale,
beta_initializer=beta_initializer,
gamma_initializer=gamma_initializer,
moving_mean_initializer=moving_mean_initializer,
moving_variance_initializer=moving_variance_initializer,
beta_regularizer=beta_regularizer,
gamma_regularizer=gamma_regularizer,
renorm=renorm,
renorm_clipping=renorm_clipping,
renorm_momentum=renorm_momentum,
fused=fused,
trainable=trainable,
name=name,
_reuse=reuse,
_scope=name)
return layer.apply(inputs, training=training)
def __init__(self,
axis=-1,
momentum=0.99,
epsilon=1e-3,
center=True,
scale=True,
beta_initializer=init_ops.zeros_initializer(),
gamma_initializer=init_ops.ones_initializer(),
moving_mean_initializer=init_ops.zeros_initializer(),
moving_variance_initializer=init_ops.ones_initializer(),
beta_regularizer=None,
gamma_regularizer=None,
renorm=False,
renorm_clipping=None,
renorm_momentum=0.99,
fused=False,
trainable=True,
name=None,
**kwargs):
super(BatchNormalization, self).__init__(trainable=trainable, name=name, **kwargs)
self.axis = axis
self.momentum = momentum
self.epsilon = epsilon
self.center = center
self.scale = scale
self.beta_initializer = beta_initializer
self.gamma_initializer = gamma_initializer
self.moving_mean_initializer = moving_mean_initializer
self.moving_variance_initializer = moving_variance_initializer
self.beta_regularizer = beta_regularizer
self.gamma_regularizer = gamma_regularizer
self.renorm = renorm
self.fused = fused
self.trainable = trainable
if fused:
if not center or not scale:
raise ValueError('fused norm requires both center and scale set to be True.')
if renorm:
raise ValueError('renorm is currently not supported.')
