本文整理汇总了Python中chainer.links.LayerNormalization方法的典型用法代码示例。如果您正苦于以下问题：Python links.LayerNormalization方法的具体用法？Python links.LayerNormalization怎么用？Python links.LayerNormalization使用的例子？那么, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类chainer.links的用法示例。
在下文中一共展示了links.LayerNormalization方法的9个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: layer_normalization
# 需要导入模块: from chainer import links [as 别名]
# 或者: from chainer.links import LayerNormalization [as 别名]
def layer_normalization(x, gamma, beta, eps=1e-5):
"""Layer normalization.
This function implements a "layer normalization"
which normalizes the input units by statistics
that are computed along the second axis,
scales and shifts them.
Args:
x (~chainer.Variable): Batch vectors.
Shape of this value must be `(batch_size, unit_size)`,
e.g., the output of :func:`~chainer.functions.linear`.
gamma (~chainer.Variable): Scaling vectors.
beta (~chainer.Variable): Shifting vectors.
Returns:
~chainer.Variable: The output variable which has the same shape
as :math:`x`.
See: `Layer Normalization <https://arxiv.org/abs/1607.06450>`_
"""
return LayerNormalization(eps)(x, gamma, beta) 示例2: __init__
# 需要导入模块: from chainer import links [as 别名]
# 或者: from chainer.links import LayerNormalization [as 别名]
def __init__(self, idim, n_layers, n_units,
e_units=2048, h=8, dropout=0.1):
super(TransformerEncoder, self).__init__()
with self.init_scope():
self.linear_in = L.Linear(idim, n_units)
self.lnorm_in = L.LayerNormalization(n_units)
self.pos_enc = PositionalEncoding(n_units, dropout, 5000)
self.n_layers = n_layers
self.dropout = dropout
for i in range(n_layers):
setattr(self, '{}{:d}'.format("lnorm1_", i),
L.LayerNormalization(n_units))
setattr(self, '{}{:d}'.format("self_att_", i),
MultiHeadSelfAttention(n_units, h))
setattr(self, '{}{:d}'.format("lnorm2_", i),
L.LayerNormalization(n_units))
setattr(self, '{}{:d}'.format("ff_", i),
PositionwiseFeedForward(n_units, e_units, dropout))
self.lnorm_out = L.LayerNormalization(n_units) 示例3: __init__
# 需要导入模块: from chainer import links [as 别名]
# 或者: from chainer.links import LayerNormalization [as 别名]
def __init__(self, dropout=None, residual_mode="normal", no_normalize=False):
super(DropoutAndAddAndNormalize, self).__init__()
if not no_normalize:
LayerNormalization = get_layer_normalization_class()
self.add_link("normalizing_layer", LayerNormalization())
assert residual_mode in "normal none after".split()
self.residual_mode = residual_mode
self.no_normalize = no_normalize
self.dropout = dropout 示例4: get_layer_normalization_class
# 需要导入模块: from chainer import links [as 别名]
# 或者: from chainer.links import LayerNormalization [as 别名]
def get_layer_normalization_class():
global use_chainer_layer_normalization
if use_chainer_layer_normalization:
return L.LayerNormalization
else:
log.info("using faster LayerNormalization")
return LayerNormalizationLink 示例5: _create_ln
# 需要导入模块: from chainer import links [as 别名]
# 或者: from chainer.links import LayerNormalization [as 别名]
def _create_ln(*args, **kwargs):
flag = chainer.disable_experimental_feature_warning
chainer.disable_experimental_feature_warning = True
try:
return links.LayerNormalization(*args, **kwargs)
finally:
chainer.disable_experimental_feature_warning = flag 示例6: __init__
# 需要导入模块: from chainer import links [as 别名]
# 或者: from chainer.links import LayerNormalization [as 别名]
def __init__(self, layer, size, dropout_ratio=0.1):
super().__init__()
self.dropout_ratio = dropout_ratio
with self.init_scope():
self.layer = layer
self.norm = L.LayerNormalization(size) 示例7: __init__
# 需要导入模块: from chainer import links [as 别名]
# 或者: from chainer.links import LayerNormalization [as 别名]
def __init__(self, sublayer, N):
super().__init__()
with self.init_scope():
self.sub_layers = sublayer.repeat(N, mode='copy')
self.norm = L.LayerNormalization(sublayer.size) 示例8: __init__
# 需要导入模块: from chainer import links [as 别名]
# 或者: from chainer.links import LayerNormalization [as 别名]
def __init__(self, hidden_channels=16, n_edge_types=5,
activation=functions.relu):
super(GNNFiLMUpdate, self).__init__()
self.n_edge_types = n_edge_types
self.activation = activation
with self.init_scope():
self.W_linear = GraphLinear(
in_size=None, out_size=self.n_edge_types * hidden_channels,
nobias=True) # W_l in eq. (6)
self.W_g = GraphLinear(
in_size=None, out_size=self.n_edge_types * hidden_channels * 2,
nobias=True) # g in eq. (6)
self.norm_layer = links.LayerNormalization() # l in eq. (6) 示例9: __init__
# 需要导入模块: from chainer import links [as 别名]
# 或者: from chainer.links import LayerNormalization [as 别名]
def __init__(self, ch0, ch1, \
nn='conv', \
norm='bn', \
activation=F.relu, \
dropout=False, \
noise=None, \
w_init=None, \
k_size = 3, \
normalize_input=False ):
self.norm = norm
self.normalize_input = normalize_input
self.activation = activation
self.dropout = dropout
self.noise = noise
self.nn = nn
layers = {}
if w_init == None:
w = chainer.initializers.GlorotNormal()
else:
w = w_init
if nn == 'down_conv':
layers['c'] = L.Convolution2D(ch0, ch1, 4, 2, 1, initialW=w)
elif nn == 'up_deconv':
layers['c'] = L.Deconvolution2D(ch0, ch1, 4, 2, 1, initialW=w)
elif nn == 'up_subpixel':
pad = k_size//2
layers['c'] = L.Convolution2D(ch0, ch1*4, k_size, 1, pad, initialW=w)
elif nn=='conv' or nn=='up_unpooling':
pad = k_size//2
layers['c'] = L.Convolution2D(ch0, ch1, k_size, 1, pad, initialW=w)
elif nn=='linear':
layers['c'] = L.Linear(ch0, ch1, initialW=w)
else:
raise Exception("Cannot find method %s" % nn)
if self.norm == 'bn':
if self.noise:
layers['n'] = L.BatchNormalization(ch1, use_gamma=False)
else:
layers['n'] = L.BatchNormalization(ch1)
elif self.norm == 'ln':
layers['n'] = L.LayerNormalization(ch1)
super(NNBlock, self).__init__(**layers)