hub / github.com/tensorflow/lingvo / BaseZeroState

Method BaseZeroState

lingvo/tasks/asr/decoder.py:464–491 · view source on GitHub ↗

Returns initial state of RNNs, and attention.

(self,
                    theta,
                    encoder_outputs,
                    bs,
                    misc_zero_states,
                    per_step_source_padding=None)

Source from the content-addressed store, hash-verified

source not stored for this graph (policy: none)

Callers 1

DecoderStepZeroStateMethod · 0.95

Calls 6

_InitAttentionMethod · 0.95

ZerosMethod · 0.80

zero_stateMethod · 0.45

ZeroAttentionStateMethod · 0.45

ComputeContextVectorWithSourceMethod · 0.45

ZeroAttentionMethod · 0.45

Tested by

no test coverage detected