Skip to content

Latest commit

 

History

History
27 lines (19 loc) · 1.16 KB

bulk_advance_from_oracle.md

File metadata and controls

27 lines (19 loc) · 1.16 KB

dragnn_ops.bulk_advance_from_oracle(handle, component=None, name=None)

dragnn_ops.bulk_advance_from_oracle(handle, component=None, name=None)

Defined in tensorflow/dragnn/core/ops/gen_dragnn_bulk_ops.py.

Given a handle to a ComputeSession, advances until all states are final. Note

that, unlike AdvanceFromOracle, this op does mutate the master state, by advancing all of its states until they are final.

Args:

  • handle: A Tensor of type string. handle to a ComputeSession.
  • component: An optional string. Defaults to "".
  • name: A name for the operation (optional).

Returns:

A tuple of Tensor objects (output_handle, gold_labels). * output_handle: A Tensor of type string. handle to updated ComputeSession. * gold_labels: A Tensor of type int32. [batch_size

  • beam_size * max_num_steps] vector of oracle actions, where max_num_steps is the maximum number of steps in the oracle action sequences for every state in the batch of beams. Each sub-segment of length max_num_steps provides the oracle action sequence for the corresponding state in the batch of beams, padded with trailing -1s.