nmt/attention_model.py [91:111]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    infer_mode = hparams.infer_mode

    dtype = tf.float32

    # Ensure memory is batch-major
    if self.time_major:
      memory = tf.transpose(encoder_outputs, [1, 0, 2])
    else:
      memory = encoder_outputs

    if (self.mode == tf.contrib.learn.ModeKeys.INFER and
        infer_mode == "beam_search"):
      memory, source_sequence_length, encoder_state, batch_size = (
          self._prepare_beam_search_decoder_inputs(
              hparams.beam_width, memory, source_sequence_length,
              encoder_state))
    else:
      batch_size = self.batch_size

    # Attention
    attention_mechanism = self.attention_mechanism_fn(
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



nmt/gnmt_model.py [182:200]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    infer_mode = hparams.infer_mode

    dtype = tf.float32

    if self.time_major:
      memory = tf.transpose(encoder_outputs, [1, 0, 2])
    else:
      memory = encoder_outputs

    if (self.mode == tf.contrib.learn.ModeKeys.INFER and
        infer_mode == "beam_search"):
      memory, source_sequence_length, encoder_state, batch_size = (
          self._prepare_beam_search_decoder_inputs(
              hparams.beam_width, memory, source_sequence_length,
              encoder_state))
    else:
      batch_size = self.batch_size

    attention_mechanism = self.attention_mechanism_fn(
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -