tf.contrib.seq2seq.gather_tree

Calculates the full beams from the per-step ids and parent beam ids.

On CPU, if an out of bound parent id is found, an error is returned. On GPU, if an out of bound parent id is found, a -1 is stored in the corresponding output value and the execution for that beam returns early.

For a given beam, past the time step containing the first decoded end_token all values are filled in with end_token.

Args
step_ids A Tensor. Must be one of the following types: int32. [max_time, batch_size, beam_width].
parent_ids A Tensor. Must have the same type as step_ids. [max_time, batch_size, beam_width].
max_sequence_lengths A Tensor of type int32. [batch_size].
end_token A Tensor. Must have the same type as step_ids. [].
name A name for the operation (optional).
Returns
A Tensor. Has the same type as step_ids. [max_time, batch_size, beam_width].

© 2020 The TensorFlow Authors. All rights reserved.
Licensed under the Creative Commons Attribution License 3.0.
Code samples licensed under the Apache 2.0 License.
https://www.tensorflow.org/versions/r1.15/api_docs/python/tf/contrib/seq2seq/gather_tree