caffe2.python.rnn_cell.MultiRNNCell Class Reference

Inheritance diagram for caffe2.python.rnn_cell.MultiRNNCell:

Public Member Functions
def	__init__ (self, cells, residual_output_layers=None, kwargs)
def	layer_scoper (self, layer_id)
def	prepare_input (self, model, input_blob)
def	get_state_names (self)
def	get_output_state_index (self)
Public Member Functions inherited from caffe2.python.rnn_cell.RNNCell
def	__init__ (self, name=None, forward_only=False, initializer=None)
def	initializer (self)
def	initializer (self, value)
def	scope (self, name)
def	apply_over_sequence (self, model, inputs, seq_lengths=None, initial_states=None, outputs_with_grads=None)
def	apply (self, model, input_t, seq_lengths, states, timestep)
def	apply_override (self, model, input_t, seq_lengths, timestep, extra_inputs=None)
def	prepare_input (self, model, input_blob)
def	get_output_state_index (self)
def	get_state_names (self)
def	get_state_names_override (self)
def	get_output_dim (self)

Public Attributes
	cells
	residual_output_layers
	output_connected_layers
	output_indices
	state_names
	initializer
Public Attributes inherited from caffe2.python.rnn_cell.RNNCell
	name
	recompute_blobs
	forward_only

Detailed Description

Multilayer RNN via the composition of RNNCell instance.

It is the resposibility of calling code to ensure the compatibility
of the successive layers in terms of input/output dimensiality, etc.,
and to ensure that their blobs do not have name conflicts, typically by
creating the cells with names that specify layer number.

Assumes first state (recurrent output) for each layer should be the input
to the next layer.

Definition at line 911 of file rnn_cell.py.

Constructor & Destructor Documentation

def caffe2.python.rnn_cell.MultiRNNCell.__init__	(		self,
			cells,
			residual_output_layers = None,
			kwargs
	)

cells: list of RNNCell instances, from input to output side.

name: string designating network component (for scoping)

residual_output_layers: list of indices of layers whose input will
be added elementwise to their output elementwise. (It is the
responsibility of the client code to ensure shape compatibility.)
Note that layer 0 (zero) cannot have residual output because of the
timing of prepare_input().

forward_only: used to construct inference-only network.

Definition at line 924 of file rnn_cell.py.

---

The documentation for this class was generated from the following file:

• caffe2/python/rnn_cell.py