code stringlengths 3 6.57k |
|---|
format(self.state_spec, self.cell.state_size) |
InputSpec(shape=(None, dim) |
self.reset_states() |
get_initial_state(self, inputs) |
shape (samples, output_dim) |
K.zeros_like(inputs) |
K.sum(initial_state, axis=(1, 2) |
K.expand_dims(initial_state) |
hasattr(self.cell.state_size, '__len__') |
K.tile(initial_state, [1, dim]) |
K.tile(initial_state, [1, self.cell.state_size]) |
__call__(self, inputs, initial_state=None, constants=None, **kwargs) |
super(RNN, self) |
__call__(inputs, **kwargs) |
InputSpec(shape=K.int_shape(state) |
InputSpec(shape=K.int_shape(constant) |
len(constants) |
hasattr(additional_inputs[0], '_keras_history') |
hasattr(tensor, '_keras_history') |
super(RNN, self) |
__call__(full_input, **kwargs) |
super(RNN, self) |
__call__(inputs, **kwargs) |
time (padded with zeros) |
build() |
isinstance(inputs, list) |
self.get_initial_state(inputs) |
isinstance(mask, list) |
len(initial_state) |
len(self.states) |
ValueError('Layer has ' + str(len(self.states) |
str(len(initial_state) |
K.int_shape(inputs) |
has_arg(self.cell.call, 'training') |
has_arg(self.cell.call, 'constants') |
ValueError('RNN cell does not support constants') |
step(inputs, states) |
step(inputs, states) |
self.cell.call(inputs, states, **kwargs) |
range(len(states) |
updates.append((self.states[i], states[i]) |
self.add_update(updates, inputs) |
getattr(last_output, '_uses_learning_phase', False) |
isinstance(states, (list, tuple) |
list(states) |
_standardize_args(self, inputs, initial_state, constants) |
tensors (or None) |
isinstance(inputs, list) |
len(inputs) |
to_list_or_none(x) |
isinstance(x, list) |
isinstance(x, tuple) |
list(x) |
to_list_or_none(initial_state) |
to_list_or_none(constants) |
reset_states(self, states=None) |
AttributeError('Layer must be stateful.') |
hasattr(self.cell.state_size, '__len__') |
K.zeros((batch_size, dim) |
K.zeros((batch_size, self.cell.state_size) |
hasattr(self.cell.state_size, '__len__') |
zip(self.states, self.cell.state_size) |
K.set_value(state, np.zeros((batch_size, dim) |
np.zeros((batch_size, self.cell.state_size) |
isinstance(states, (list, tuple) |
len(states) |
len(self.states) |
str(len(self.states) |
str(len(states) |
str(states) |
enumerate(zip(states, self.states) |
hasattr(self.cell.state_size, '__len__') |
ValueError('State ' + str(index) |
str((batch_size, dim) |
str(value.shape) |
K.set_value(state, value) |
get_config(self) |
self.cell.get_config() |
super(RNN, self) |
get_config() |
dict(list(base_config.items() |
list(config.items() |
from_config(cls, config, custom_objects=None) |
deserialize_layer(config.pop('cell') |
config.pop('num_constants', None) |
cls(cell, **config) |
trainable_weights(self) |
isinstance(self.cell, Layer) |
non_trainable_weights(self) |
isinstance(self.cell, Layer) |
losses(self) |
isinstance(self.cell, Layer) |
get_losses_for(self, inputs=None) |
isinstance(self.cell, Layer) |
self.cell.get_losses_for(inputs) |
super(RNN, self) |
get_losses_for(inputs) |
super(RNN, self) |
get_losses_for(inputs) |
SimpleRNNCell(Layer) |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.