When using static_state_saving_rnn(..) in the following manner
authorA. Unique TensorFlower <gardener@tensorflow.org>
Wed, 9 May 2018 20:55:20 +0000 (13:55 -0700)
committerTensorFlower Gardener <gardener@tensorflow.org>
Wed, 9 May 2018 21:13:57 +0000 (14:13 -0700)
commitec0ef29835563b762ec9443a3c194c5c904fd6be
treecf6db99b4760f9214dedc9879dba6830827a8d2a
parent705550357fb9f1955207b5953779e8a382744f30
When using static_state_saving_rnn(..) in the following manner

    _, state = tf.nn.static_state_saving_rnn(..)

the runtime will be blocked after some time, because the save_state method of the state_saver object won't be executed as a part of the graph (that part depends only on output node in the current implementation).
Now it should depend on state as well, so the above implementation won't be blocked.

PiperOrigin-RevId: 196024050
tensorflow/contrib/rnn/python/kernel_tests/core_rnn_test.py
tensorflow/python/ops/rnn.py