def __init__(self, offset_creator_class, weight_basis,
filters,
kernel_size,
strides=(1, 1),
padding='valid',
data_format=None,
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
**kwargs):
super(RProjLocallyConnected2D, self).__init__(offset_creator_class, weight_basis, **kwargs)
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, 2, 'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, 2, 'strides')
self.padding = conv_utils.normalize_padding(padding)
if self.padding != 'valid':
raise ValueError('Invalid border mode for LocallyConnected2D '
'(only "valid" is supported): ' + padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.input_spec = InputSpec(ndim=4)
def __init__(self, filters, kernel_size,strides=1,padding='same',dilation_rate=1,
bias_initializer='zeros',kernel_initializer='glorot_uniform', activation='linear', weights=None,
border_mode='valid', subsample_length=1,
kernel_regularizer=None, bias_regularizer=None, activity_regularizer=None,
kernel_constraint=None, bias_constraint=None,
use_bias=True, input_dim=None, input_length=None, tied_to=None,data_format='channels_last',rank=1,learnedKernel=None,layer_inner=None,
**kwargs):
if border_mode not in {'valid', 'same'}:
raise Exception('Invalid border mode for Convolution1D:', border_mode)
self.input_spec = [InputSpec(ndim=3)]
self.input_dim = input_dim
self.input_length = input_length
self.tied_to = tied_to
self.learnedKernel = np.array(learnedKernel)
#self.tied_to.set_weights([weights, bias])
self.strides = conv_utils.normalize_tuple(strides, rank, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.dilation_rate = conv_utils.normalize_tuple(dilation_rate, rank, 'dilation_rate')
self.data_format = data_format
self.filters = filters
if self.tied_to is not None:
self.kernel_size = self.tied_to.kernel_size
else:
self.kernel_size = kernel_size
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.activation = activations.get(activation)
assert border_mode in {'valid', 'same'}, 'border_mode must be in {valid, same}'
self.border_mode = border_mode
self.subsample_length = subsample_length
self.subsample = (subsample_length, 1)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.use_bias = use_bias
self.rank = 1
self.layer_inner = layer_inner
if self.input_dim:
kwargs['input_shape'] = (self.input_length, self.input_dim)
#self.layer_inner = kwargs.pop('layer_inner')
super(Convolution1D_tied, self).__init__(**kwargs)
def __init__(self,
filters,
kernel_size,
strides=(1, 1),
padding='valid',
data_format=None,
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
implementation=1,
**kwargs):
super().__init__(**kwargs)
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, 2,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides,
2,
'strides',
allow_zero=True)
self.padding = conv_utils.normalize_padding(padding)
if self.padding != 'valid' and implementation == 1:
raise ValueError(
'Invalid border mode for LocallyConnected2D '
'(only "valid" is supported if implementation is 1): ' +
padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.implementation = implementation
self.input_spec = InputSpec(ndim=4)
def __init__(self,
filters,
kernel_size,
rank=1,
strides=1,
padding='valid',
data_format='channels_last',
dilation_rate=1,
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
**kwargs):
super(Conv1D_linearphase, self).__init__(**kwargs)
self.rank = rank
self.filters = filters
self.kernel_size_ = kernel_size
if kernel_size % 2:
self.kernel_size = conv_utils.normalize_tuple(
kernel_size // 2 + 1, rank, 'kernel_size')
else:
self.kernel_size = conv_utils.normalize_tuple(
kernel_size // 2, rank, 'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, rank, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, rank, 'dilation_rate')
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.input_spec = InputSpec(ndim=self.rank + 2)
def __init__(self,
pool_function,
pool_size,
strides,
padding='valid',
data_format='channels_last',
name=None,
**kwargs):
super(Pooling3D, self).__init__(name=name, **kwargs)
if data_format is None:
data_format = backend.image_data_format()
if strides is None:
strides = pool_size
self.pool_function = pool_function
self.pool_size = conv_utils.normalize_tuple(pool_size, 3, 'pool_size')
self.strides = conv_utils.normalize_tuple(strides, 3, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.input_spec = InputSpec(ndim=5)
def __init__(self,
rank,
filters,
kernel_size,
strides=1,
padding='valid',
data_format=None,
dilation_rate=1,
activation=None,
use_bias=True,
kernel_initializer='he_complex',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
seed=None,
spectral_parametrization=False,
**kwargs):
super(_ComplexConv, self).__init__(**kwargs)
self.rank = rank
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, rank,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, rank, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = K.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, rank, 'dilation_rate')
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = cinitializers.get(kernel_initializer)
self.bias_initializer = cinitializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.input_spec = InputSpec(ndim=self.rank + 2)
self.spectral_parametrization = spectral_parametrization
self.seed = seed if seed is not None else np.random.randint(1, 1e6)
def __init__(self,
filters,
kernel_size,
strides=1,
padding='valid',
data_format=None,
dilation_rate=1,
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
tied_to=None,
**kwargs):
super().__init__(filters, kernel_size, **kwargs)
self.bias_constraint = constraints.get(bias_constraint)
self.supports_masking = True
self.tied_to = tied_to
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, 2,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, 2, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = K.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, 2, 'dilation_rate')
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.input_spec = InputSpec(ndim=4)
def __init__(self,
rank,
filters,
kernel_size,
strides=1,
padding='valid',
data_format=None,
dilation_rate=1,
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
Masked=False,
kernel_mask_val=None,
**kwargs):
super(MaskedConv, self).__init__(**kwargs)
self.rank = rank
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, rank,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, rank, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = K.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, rank, 'dilation_rate')
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.input_spec = InputSpec(ndim=self.rank + 2)
self.Masked = Masked
self.kernel_mask_val = kernel_mask_val
def __init__(self, filters,
kernel_size,
strides=(1, 1),
padding='valid',
group=1,
data_format=None,
dilation_rate=(1, 1),
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
**kwargs):
super(GroupConv2D, self).__init__(
filters=filters,
kernel_size=kernel_size,
strides=strides,
padding=padding,
data_format=data_format,
dilation_rate=dilation_rate,
activation=activation,
use_bias=use_bias,
kernel_initializer=kernel_initializer,
bias_initializer=bias_initializer,
kernel_regularizer=kernel_regularizer,
bias_regularizer=bias_regularizer,
activity_regularizer=activity_regularizer,
kernel_constraint=kernel_constraint,
bias_constraint=bias_constraint,
**kwargs)
self.kernel_size = conv_utils.normalize_tuple(kernel_size, 2, 'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, 2, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.dilation_rate = conv_utils.normalize_tuple(dilation_rate, 2, 'dilation_rate')
self.channel_axis = None
self.group = group
def __init__(self,
filters,
kernel_size,
strides=1,
padding='valid',
dilation_rate=1,
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
function=None,
arguments=None,
normalized_position=True,
**kwargs):
super(SemiConv2D, self).__init__(**kwargs)
self.rank = 2
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, self.rank,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, self.rank,
'strides')
self.padding = conv_utils.normalize_padding(padding)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, self.rank, 'dilation_rate')
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.input_spec = InputSpec(ndim=self.rank + 2)
self.function = function
self.arguments = arguments if arguments else {}
self.normalized_position = normalized_position
def __init__(self,
filters,
kernel_size,
strides=(1, 1),
padding='valid',
data_format=None,
dilation_rate=1,
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
template_tensor=None,
clusterid=None,
**kwargs):
super(ModifiedConv2D, self).__init__(**kwargs)
self.rank = 2
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, 2,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, 2, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, 2, 'dilation_rate')
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.input_spec = InputSpec(ndim=self.rank + 2)
self.template_tensor = template_tensor
self.clusterid = clusterid
def __init__(self,
filters,
kernel_size,
strides=1,
padding='valid',
data_format=None,
dilation_rate=1,
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
num_bits=8,
**kwargs):
super(QuantizedConv2D, self).__init__(**kwargs)
self.rank = 2
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, self.rank,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, self.rank,
'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = K.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, self.rank, 'dilation_rate')
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.num_bits = num_bits
self.quantization = _Quantization(num_bits=self.num_bits)
self.input_spec = InputSpec(ndim=self.rank + 2)
def __init__(self, rank,
filters,
kernel_size,
time_distributed,
strides=1,
padding='valid',
data_format=None,
dilation_rate=1,
activation=None,
#dynamic_activation = 'softmax',
use_bias=False,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
**kwargs):
super(adaptive_conv, self).__init__(**kwargs)
self.rank = rank
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, rank,
'kernel_size')
self.time_distributed = time_distributed
self.strides = conv_utils.normalize_tuple(strides, rank, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = K.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(dilation_rate, rank,
'dilation_rate')
self.activation = activations.get(activation)
#self.dynamic_activation = activations.relu(alpha=0.2)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
def __init__(self, filters, kernel_size, strides=1, padding='valid', data_format=None, dilation_rate=1, activation=None, use_bias=True, kernel_initializer='glorot_uniform', bias_initializer='zeros', kernel_regularizer=None, bias_regularizer=None, activity_regularizer=None, kernel_constraint=None, bias_constraint=None, kl_lambda = 1e-6, drop_prob = 0.20, batch_size = 8, **kwargs): super(Conv2D_wd, self).__init__(**kwargs) self.rank = 2 self.filters = filters self.kernel_size = conv_utils.normalize_tuple(kernel_size, 2, 'kernel_size') self.strides = conv_utils.normalize_tuple(strides, 2, 'strides') self.padding = conv_utils.normalize_padding(padding) self.data_format = conv_utils.normalize_data_format(data_format) self.dilation_rate = conv_utils.normalize_tuple(dilation_rate, 2, 'dilation_rate') self.activation = activations.get(activation) self.use_bias = use_bias self.kernel_initializer = initializers.get(kernel_initializer) self.bias_initializer = initializers.get(bias_initializer) self.kernel_regularizer = regularizers.get(kernel_regularizer) self.bias_regularizer = regularizers.get(bias_regularizer) self.activity_regularizer = regularizers.get(activity_regularizer) self.kernel_constraint = constraints.get(kernel_constraint) self.bias_constraint = constraints.get(bias_constraint) self.kl_lambda = kl_lambda self.drop_prob = drop_prob self.batch_size = batch_size
def __init__(self, filters,
kernel_size,
strides=1,
padding='same',
data_format=None,
activation='tanh',
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
use_bias=True,
** kwargs):
super(self_conv1d, self).__init__(**kwargs)
self.rank = 1
self.filters = filters
self.data_format = conv_utils.normalize_data_format(data_format)
self.kernel_size = conv_utils.normalize_tuple(kernel_size, self.rank, 'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, self.rank, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.activation = activations.get(activation)
self.use_bias = use_bias
self.supports_masking = True
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
def __init__(self,
offset_creator_class,
weight_basis,
filters,
kernel_size,
strides=(1, 1),
padding='valid',
data_format=None,
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
**kwargs):
super(RProjLocallyConnected2D, self).__init__(offset_creator_class,
weight_basis, **kwargs)
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, 2,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, 2, 'strides')
self.padding = conv_utils.normalize_padding(padding)
if self.padding != 'valid':
raise ValueError('Invalid border mode for LocallyConnected2D '
'(only "valid" is supported): ' + padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.input_spec = InputSpec(ndim=4)
def __init__(self,
rank,
filters,
kernel_size,
strides=1,
padding='valid',
data_format=None,
dilation_rate=1,
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
output_merge=False,
**kwargs):
super().__init__(**kwargs)
self.rank = rank
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple((2, ) + kernel_size,
rank + 1, 'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, rank, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = K.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, rank, 'dilation_rate')
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.output_merge = output_merge
def __init__(self,
filters,
kernel_size,
strides=(1, 1),
padding='valid',
data_format='channels_last',
dilation_rate=(1, 1),
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
**kwargs):
# print(kwargs)
super(pruned_Conv2D, self).__init__(**kwargs)
self.rank = 2
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, self.rank,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, self.rank,
'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = K.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, self.rank, 'dilation_rate')
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.input_spec = InputSpec(ndim=self.rank + 2)
def __init__(self,
pool_function,
pool_size,
strides,
padding="valid",
data_format=None,
name=None,
**kwargs):
super().__init__(name=name, **kwargs)
if data_format is None:
data_format = backend.image_data_format()
if strides is None:
strides = pool_size
self.pool_function = pool_function
self.pool_size = conv_utils.normalize_tuple(pool_size, 2, "pool_size")
self.strides = conv_utils.normalize_tuple(strides,
2,
"strides",
allow_zero=True)
self.padding = conv_utils.normalize_padding(padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.input_spec = InputSpec(ndim=4)
def __init__(self,
n_replicas=1,
strides=(1, 1),
padding='valid',
dilation_rate=(1, 1),
activation='relu',
components=None,
**kwargs):
super(CosineSimilarity2D, self).__init__(**kwargs)
self.n_replicas = n_replicas
self.rank = 2
self.strides = conv_utils.normalize_tuple(strides, self.rank,
'strides')
self.padding = conv_utils.normalize_padding(padding)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, self.rank, 'dilation_rate')
self.components = components
self.activation = activations.get(activation)
if K.image_data_format() != 'channels_last':
raise ValueError("The layer requires `K.image_data_format()` to "
"be 'channels_last'.")
def __init__(self,
filters,
kernel_size,
strides=(1, 1),
padding='valid',
data_format=None,
dilation_rate=(1, 1),
activation=None,
use_bias=True,
kernel_initializer='he_normal',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
rank=2,
**kwargs):
super(GroupConv2D_Backend, self).__init__(**kwargs)
self.rank = rank
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, rank,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, rank, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = K.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, rank, 'dilation_rate')
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.input_spec = InputSpec(ndim=self.rank + 2)
def __init__(self,
rank,
filters,
kernel_size,
strides=1,
padding='valid',
data_format=None,
dilation_rate=1,
activation=None,
use_bias=False,
kernel_regularizer=None,
**kwargs):
super(GaussScaler, self).__init__(**kwargs)
#def __init__(self, num_filters, kernel_size, incoming_channels=1, **kwargs):
self.rank = rank
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, rank,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, rank, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, rank, 'dilation_rate')
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.input_spec = InputSpec(ndim=self.rank + 2)
#self.input_shape = input_shape
if 'input_shape' not in kwargs and 'input_dim' in kwargs:
kwargs['input_shape'] = (kwargs.pop('input_dim'))
print(kwargs)
self.kernel_size = kernel_size
self.num_filters = filters
#self.incoming_channels = incoming_channels
super(GaussScaler, self).__init__(**kwargs)
def __init__(self,
kernel_width=3,
sigma=1,
filters=1,
strides=(1, 1),
padding='valid',
data_format=None,
kernel_initializer='uniform',
kernel_regularizer=None,
kernel_constraint=None,
use_bias=True,
bias_initializer='zeros',
bias_regularizer=None,
bias_constraint=None,
activity_regularizer=None,
activation=None,
**kwargs):
# self._output_dim = output_dim
self._kernel_width = kernel_width
self.kernel_size = (kernel_width, kernel_width)
self.strides = conv_utils.normalize_tuple(strides, 2, 'strides')
self.padding = conv_utils.normalize_padding(padding)
if self.padding != 'valid':
raise ValueError('Invalid border mode for LocallyConnected2D '
'(only "valid" is supported): ' + padding)
self.activation = activations.get(activation)
self._gaussian = _gaussian(kernel_width, sigma)
self.filters = filters
self.data_format = K.normalize_data_format(data_format)
self.kernel_initializer = initializers.get(kernel_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.use_bias = use_bias
self.bias_initializer = initializers.get(bias_initializer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.bias_constraint = constraints.get(bias_constraint)
self.activity_regularizer = regularizers.get(activity_regularizer)
super(Focus, self).__init__(**kwargs)
def __init__(self,
pool_size=(2, 2),
strides=None,
padding='same',
data_format=None,
**kwargs):
super(UnPooling2D, self).__init__(**kwargs)
data_format = conv_utils.normalize_data_format(data_format)
if strides is None:
strides = pool_size
self.pool_size = conv_utils.normalize_tuple(pool_size, 2, 'pool_size')
self.strides = conv_utils.normalize_tuple(strides, 2, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.input_spec = [InputSpec(ndim=4), InputSpec(ndim=4)]
self.expected_out_shape = None
self.size = None
if padding is not 'same':
print('This Layer is implemented only\
for padding="same".')
if pool_size != strides:
print('pool_size:', pool_size, 'strides:', strides)
raise ValueError('Keep strides and pool_size equal.')
def __init__(self,
rank,
patch_size,
strides=1,
padding='valid',
data_format=None,
dilation_rate=1,
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
activity_regularizer=None,
kernel_constraint=None,
**kwargs):
super(PatchExtracter, self).__init__(**kwargs)
self.rank = rank
self.patch_size = conv_utils.normalize_tuple(patch_size, rank,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, rank, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, rank, 'dilation_rate')
self.input_spec = InputSpec(ndim=self.rank + 2)
def __init__(self, fwh_projector, rank,
filters,
kernel_size,
strides=1,
padding='valid',
data_format=None,
dilation_rate=1,
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
**kwargs):
# embed()
super(_RProjFWHConv, self).__init__(**kwargs)
self.fwh_projector = fwh_projector
self.rank = rank
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, rank, 'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, rank, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(dilation_rate, rank, 'dilation_rate')
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.input_spec = InputSpec(ndim=self.rank + 2)
def __init__(self,
pool_size=2,
strides=None,
padding='valid',
default_value=0.0,
terminate_mask=False,
**kwargs):
assert padding in ['valid',
'same'], "Invalid value for `padding`: " + padding
super().__init__(**kwargs)
if strides is None:
strides = pool_size
self.pool_size = conv_utils.normalize_tuple(pool_size, 1, 'pool_size')
self.strides = conv_utils.normalize_tuple(strides, 1, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.default_value = default_value
self.terminate_mask = terminate_mask
self.input_spec = None # removed ndim restriction: InputSpec(ndim=3)
return
def __init__(self,
filters,
kernel_size,
strides=1,
padding='valid',
data_format=None,
dilation_rate=1,
rank=2,
**kwargs):
super(GradientConv2D, self).__init__(**kwargs)
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, rank,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, rank, 'strides')
self.data_format = conv_utils.normalize_data_format(data_format)
self.padding = conv_utils.normalize_padding(padding)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, rank, 'dilation_rate')
self.sx = np.array([[-1, 0, 1], [-2, 0, 2], [-1, 0, 1]])
# self.sx = K.constant(sx, shape=(3, 3))
self.sy = np.array([[1, 2, 1], [0, 0, 0], [-1, -2, -1]])
def test_invalid_padding():
with pytest.raises(ValueError):
conv_utils.normalize_padding('diagonal')
def __init__(self,
num_filters=1,
kernel_size=(1, 1),
bin_sizes=[1, 2, 3, 6],
pool_mode='avg',
padding='valid',
activation=None,
use_bias=True,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
**kwargs):
"""
Initialize a new Pyramid Pooling Module.
Args:
num_filters: the number of filters per convolutional unit
bin_sizes: sizes for pooling bins
pool_mode: pooling mode to use
padding: One of `"valid"` or `"same"` (case-insensitive).
activation: Activation function to use
use_bias: whether layer uses a bias vector
kernel_initializer: Initializer for kernel weights
bias_initializer: Initializer for bias vector
kernel_regularizer: Regularizer function applied to kernel weights
bias_regularizer: Regularizer function applied to bias vector
activity_regularizer: Regularizer function applied to output
kernel_constraint: Constraint function applied to kernel
bias_constraint: Constraint function applied to bias vector
kwargs: keyword arguments for Layer super constructor
Returns:
None
"""
if padding != 'same' and any(x > 1 for x in kernel_size):
raise ValueError(
"padding should be 'same' if the kernel size is larger than (1, 1)"
)
# setup instance variables
self.input_spec = InputSpec(ndim=4)
self.num_filters = num_filters
self.kernel_size = kernel_size
self.bin_sizes = bin_sizes
self.pool_mode = pool_mode
self.padding = conv_utils.normalize_padding(padding)
self.data_format = 'channels_last'
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
# initialize the kernels and biases
self.kernels = None
self.biases = None
# call the super constructor
super(PyramidPoolingModule, self).__init__(**kwargs)
def __init__(self, filters,
kernel_size,
retention_ratio=None,
learn_retention_ratio=False,
strides=(1, 1),
padding='valid',
data_format=None,
dilation_rate=(1, 1),
activation='tanh',
recurrent_activation='hard_sigmoid',
use_bias=True,
kernel_initializer='glorot_uniform',
recurrent_initializer='orthogonal',
bias_initializer='zeros',
unit_forget_bias=True,
kernel_regularizer=None,
recurrent_regularizer=None,
bias_regularizer=None,
kernel_constraint=None,
recurrent_constraint=None,
bias_constraint=None,
dropout=0.,
recurrent_dropout=0.,
**kwargs):
super(ConvlpRNN2DCell, self).__init__(**kwargs)
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, 2, 'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, 2, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(dilation_rate, 2, 'dilation_rate')
self.activation = activations.get(activation)
self.recurrent_activation = activations.get(recurrent_activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.recurrent_initializer = initializers.get(recurrent_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.unit_forget_bias = unit_forget_bias
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.recurrent_regularizer = regularizers.get(recurrent_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.recurrent_constraint = constraints.get(recurrent_constraint)
self.bias_constraint = constraints.get(bias_constraint)
if K.backend() == 'theano' and (dropout or recurrent_dropout):
warnings.warn(
'RNN dropout is no longer supported with the Theano backend '
'due to technical limitations. '
'You can either set `dropout` and `recurrent_dropout` to 0, '
'or use the TensorFlow backend.')
dropout = 0.
recurrent_dropout = 0.
self.dropout = min(1., max(0., dropout))
self.recurrent_dropout = min(1., max(0., recurrent_dropout))
self.state_size = (self.filters, self.filters)
self._dropout_mask = None
self._recurrent_dropout_mask = None
self._learn_retention_ratio = learn_retention_ratio
self._num_units = filters
self.set_retention_ratio = retention_ratio
def __init__(
self,
rank,
filters,
kernel_size,
strides=1,
padding='valid',
data_format=None,
dilation_rate=1,
activation=None, # key of activation
use_bias=True,
normalize_weight=False,
kernel_initializer='he_complex',
bias_initializer='zeros',
gamma_diag_initializer=sqrt_init,
gamma_off_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
gamma_diag_regularizer=None,
gamma_off_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
gamma_diag_constraint=None,
gamma_off_constraint=None,
init_criterion='he',
seed=None,
spectral_parametrization=False,
transposed=False,
epsilon=1e-7,
**kwargs):
super(_ComplexConv, self).__init__(**kwargs)
self.rank = rank
self.filters = filters
self.kernel_size = conv_utils.normalize_tuple(kernel_size, rank,
'kernel_size')
self.strides = conv_utils.normalize_tuple(strides, rank, 'strides')
self.padding = conv_utils.normalize_padding(padding)
self.data_format = normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, rank, 'dilation_rate')
self.activation = activation
self.use_bias = use_bias
self.normalize_weight = normalize_weight
self.init_criterion = init_criterion
self.spectral_parametrization = spectral_parametrization
self.transposed = transposed
self.epsilon = epsilon
self.kernel_initializer = sanitizedInitGet(kernel_initializer)
self.bias_initializer = sanitizedInitGet(bias_initializer)
self.gamma_diag_initializer = sanitizedInitGet(gamma_diag_initializer)
self.gamma_off_initializer = sanitizedInitGet(gamma_off_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.gamma_diag_regularizer = regularizers.get(gamma_diag_regularizer)
self.gamma_off_regularizer = regularizers.get(gamma_off_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.gamma_diag_constraint = constraints.get(gamma_diag_constraint)
self.gamma_off_constraint = constraints.get(gamma_off_constraint)
self.seed = seed if seed is not None else np.random.randint(1, 1e6)
self.input_spec = InputSpec(ndim=self.rank + 2)
# The following are initialized later
self.kernel_shape = None
self.kernel = None
self.gamma_rr = None
self.gamma_ii = None
self.gamma_ri = None
self.bias = None
def __init__(
self,
rank,
filters,
kernel_size,
strides=1,
padding="valid",
data_format=None,
dilation_rate=1,
groups=1,
activation=None,
use_bias=True,
kernel_initializer="glorot_uniform",
bias_initializer="zeros",
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
trainable=True,
name=None,
conv_op=None,
**kwargs,
):
super().__init__(
trainable=trainable,
name=name,
activity_regularizer=regularizers.get(activity_regularizer),
**kwargs,
)
self.rank = rank
if isinstance(filters, float):
filters = int(filters)
if filters is not None and filters <= 0:
raise ValueError("Invalid value for argument `filters`. "
"Expected a strictly positive value. "
f"Received filters={filters}.")
self.filters = filters
self.groups = groups or 1
self.kernel_size = conv_utils.normalize_tuple(kernel_size, rank,
"kernel_size")
self.strides = conv_utils.normalize_tuple(strides,
rank,
"strides",
allow_zero=True)
self.padding = conv_utils.normalize_padding(padding)
self.data_format = conv_utils.normalize_data_format(data_format)
self.dilation_rate = conv_utils.normalize_tuple(
dilation_rate, rank, "dilation_rate")
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.input_spec = InputSpec(min_ndim=self.rank + 2)
self._validate_init()
self._is_causal = self.padding == "causal"
self._channels_first = self.data_format == "channels_first"
self._tf_data_format = conv_utils.convert_data_format(
self.data_format, self.rank + 2)
def test_invalid_padding():
with pytest.raises(ValueError):
conv_utils.normalize_padding('diagonal')
