Function recurrent_net

caffe2/python/recurrent.py:10–285

net: the main net operator should be added to cell_net: cell_net which is executed in a recurrent fasion inputs: sequences to be fed into the recurrent net. Currently only one input is supported. It has to be in a format T x N x (D1...Dk) where T is lengths of the sequence. N

(
        net, cell_net, inputs, initial_cell_inputs,
        links, timestep=None, scope=None, outputs_with_grads=(0,),
        recompute_blobs_on_backward=None, forward_only=False,
)

from caffe2.python import core, workspace

def recurrent_net(
        net, cell_net, inputs, initial_cell_inputs,
        links, timestep=None, scope=None, outputs_with_grads=(0,),
        recompute_blobs_on_backward=None, forward_only=False,
):
    '''
    net: the main net operator should be added to

    cell_net: cell_net which is executed in a recurrent fasion

    inputs: sequences to be fed into the recurrent net. Currently only one input
    is supported. It has to be in a format T x N x (D1...Dk) where T is lengths
    of the sequence. N is a batch size and (D1...Dk) are the rest of dimentions

    initial_cell_inputs: inputs of the cell_net for the 0 timestamp.
    Format for each input is:
        (cell_net_input_name, external_blob_with_data)

    links: a dictionary from cell_net input names in moment t+1 and
    output names of moment t. Currently we assume that each output becomes
    an input for the next timestep.

    timestep: name of the timestep blob to be used. If not provided "timestep"
    is used.

    scope: Internal blobs are going to be scoped in a format
    <scope_name>/<blob_name>
    If not provided we generate a scope name automatically

    outputs_with_grads : position indices of output blobs which will receive
    error gradient (from outside recurrent network) during backpropagation

    recompute_blobs_on_backward: specify a list of blobs that will be
                 recomputed for backward pass, and thus need not to be
                 stored for each forward timestep.

    forward_only: if True, only forward steps are executed
    ''&#x27;
    assert len(inputs) == 1, "Only one input blob is supported so far"

    input_blobs = [str(i[0]) for i in inputs]
    initial_input_blobs = [str(x[1]) for x in initial_cell_inputs]
    op_name = net.NextName('recurrent')

    def s(name):
        # We have to manually scope due to our internal/external blob
        # relationships.
        scope_name = op_name if scope is None else scope
        return "{}/{}".format(str(scope_name), str(name))

    # determine inputs that are considered to be references
    # it is those that are not referred to in inputs or initial_cell_inputs
    known_inputs = [str(b) for b in input_blobs + initial_input_blobs]
    known_inputs += [str(x[0]) for x in initial_cell_inputs]
    if timestep is not None:
        known_inputs.append(str(timestep))
    references = [
        core.BlobReference(b) for b in cell_net.Proto().external_input