code stringlengths 3 6.57k |
|---|
use
(see [activations](../activations.md) |
applied
(ie. "linear" activation: `a(x) |
inputs
(see [initializers](../initializers.md) |
state
(see [initializers](../initializers.md) |
vector
(see [initializers](../initializers.md) |
matrix
(see [regularizer](../regularizers.md) |
matrix
(see [regularizer](../regularizers.md) |
vector
(see [regularizer](../regularizers.md) |
matrix
(see [constraints](../constraints.md) |
matrix
(see [constraints](../constraints.md) |
vector
(see [constraints](../constraints.md) |
super(SimpleRNNCell, self) |
__init__(**kwargs) |
activations.get(activation) |
initializers.get(kernel_initializer) |
initializers.get(recurrent_initializer) |
initializers.get(bias_initializer) |
regularizers.get(kernel_regularizer) |
regularizers.get(recurrent_regularizer) |
regularizers.get(bias_regularizer) |
constraints.get(kernel_constraint) |
constraints.get(recurrent_constraint) |
constraints.get(bias_constraint) |
min(1., max(0., dropout) |
min(1., max(0., recurrent_dropout) |
build(self, input_shape) |
self.add_weight(shape=(input_shape[-1], self.units) |
self.add_weight(shape=(self.units,) |
call(self, inputs, states, training=None) |
_generate_dropout_ones(inputs, K.shape(inputs) |
_generate_dropout_ones(inputs, self.units) |
K.dot(inputs * dp_mask, self.kernel) |
K.dot(inputs, self.kernel) |
K.bias_add(h, self.bias) |
K.dot(prev_output, self.recurrent_kernel) |
self.activation(output) |
get_config(self) |
activations.serialize(self.activation) |
initializers.serialize(self.kernel_initializer) |
initializers.serialize(self.recurrent_initializer) |
initializers.serialize(self.bias_initializer) |
regularizers.serialize(self.kernel_regularizer) |
regularizers.serialize(self.recurrent_regularizer) |
regularizers.serialize(self.bias_regularizer) |
constraints.serialize(self.kernel_constraint) |
constraints.serialize(self.recurrent_constraint) |
constraints.serialize(self.bias_constraint) |
super(SimpleRNNCell, self) |
get_config() |
dict(list(base_config.items() |
list(config.items() |
SimpleRNN(RNN) |
use
(see [activations](../activations.md) |
applied
(ie. "linear" activation: `a(x) |
inputs
(see [initializers](../initializers.md) |
state
(see [initializers](../initializers.md) |
vector
(see [initializers](../initializers.md) |
matrix
(see [regularizer](../regularizers.md) |
matrix
(see [regularizer](../regularizers.md) |
vector
(see [regularizer](../regularizers.md) |
layer (its "activation") |
matrix
(see [constraints](../constraints.md) |
matrix
(see [constraints](../constraints.md) |
vector
(see [constraints](../constraints.md) |
Boolean (default False) |
Boolean (default False) |
Boolean (default False) |
kwargs.pop('implementation') |
K.backend() |
super(SimpleRNN, self) |
regularizers.get(activity_regularizer) |
call(self, inputs, mask=None, training=None, initial_state=None) |
super(SimpleRNN, self) |
units(self) |
activation(self) |
use_bias(self) |
kernel_initializer(self) |
recurrent_initializer(self) |
bias_initializer(self) |
kernel_regularizer(self) |
recurrent_regularizer(self) |
bias_regularizer(self) |
kernel_constraint(self) |
recurrent_constraint(self) |
bias_constraint(self) |
dropout(self) |
recurrent_dropout(self) |
get_config(self) |
activations.serialize(self.activation) |
initializers.serialize(self.kernel_initializer) |
initializers.serialize(self.recurrent_initializer) |
initializers.serialize(self.bias_initializer) |
regularizers.serialize(self.kernel_regularizer) |
regularizers.serialize(self.recurrent_regularizer) |
regularizers.serialize(self.bias_regularizer) |
regularizers.serialize(self.activity_regularizer) |
constraints.serialize(self.kernel_constraint) |
constraints.serialize(self.recurrent_constraint) |
constraints.serialize(self.bias_constraint) |
super(SimpleRNN, self) |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.