experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Properly document sequence_length parameter of dynamic_rnn

Browse Source 
Fixes #27981

PiperOrigin-RevId: 244886657
This commit is contained in:
Mihai Maruseac 2019-04-23 11:03:06 -07:00 committed by TensorFlower Gardener
parent 1154525a09
commit ca644c9ca5
1 changed files with 2 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
tensorflow/python/ops/rnn.py
View File

@ -580,9 +580,8 @@ def dynamic_rnn(cell,
tuple of Tensors each with dimensions `[batch_size, ...]`. tuple of Tensors each with dimensions `[batch_size, ...]`.
sequence_length: (optional) An int32/int64 vector sized `[batch_size]`. Used sequence_length: (optional) An int32/int64 vector sized `[batch_size]`. Used
to copy-through state and zero-out outputs when past a batch element's to copy-through state and zero-out outputs when past a batch element's
sequence length. This argument enables this API to extract sequence length. This parameter enables users to extract the last valid
the last VALID state of RNN instead of a PADDED time step, so it is state and properly padded outputs, so it is provided for correctness.
for correctness.
initial_state: (optional) An initial state for the RNN. If `cell.state_size` initial_state: (optional) An initial state for the RNN. If `cell.state_size`
is an integer, this must be a `Tensor` of appropriate type and shape is an integer, this must be a `Tensor` of appropriate type and shape
`[batch_size, cell.state_size]`. If `cell.state_size` is a tuple, this `[batch_size, cell.state_size]`. If `cell.state_size` is a tuple, this