MCPcopy Create free account
hub / github.com/tensorflow/lingvo / _PostprocessSample

Method _PostprocessSample

lingvo/core/base_decoder.py:387–436  ·  view source on GitHub ↗

Add topk_hyps, topk_ids, topk_lens, topk_scores tensors to `sample`. These features are required by `.BeamSearchDecodeOutput`. Args: sample: a NestedMap with `id`, `paddings`, and `logits` fields. is_tpu: whether inference is being run on TPU. Returns: sample with ad

(self, sample, is_tpu)

Source from the content-addressed store, hash-verified

source not stored for this graph (policy: none)

Callers 2

SampleSequenceDecodeMethod · 0.80
SampleSequenceDecodeMethod · 0.80

Calls

no outgoing calls

Tested by

no test coverage detected