Hopefully this is not stupid. In the Tensorflow tutorial, the states of the RNN are up

I am also interested in the problem. maybe base on experience? <a class="user-mentio

<a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/us

bidirectional RNN states never updated? about tensorflow-examples HOT 6 OPEN

aymericdamien commented on April 28, 2024

bidirectional RNN states never updated?

from tensorflow-examples.

Comments (6)

aymericdamien commented on April 28, 2024 1

The old version of Tensorflow didn't returned the states, so output[-1] is fine (see https://github.com/tensorflow/tensorflow/blob/r0.7/tensorflow/python/ops/rnn.py#L323).
However in newer version, it returns a tuple instead, so as you said, you need to change
outputs = rnn.bidirectionnal_rnn(... by outputs, _, _ = rnn.bidirectionnal_rnn(...

from tensorflow-examples.

coderLong commented on April 28, 2024 1

I am also interested in the problem. maybe base on experience? @zandaqo

from tensorflow-examples.

aymericdamien commented on April 28, 2024 1

There is no particular reason, it depends on your data. A recommendation is to use multiple of 32 as it may speedup a little the computation when using float32.

from tensorflow-examples.

zandaqo commented on April 28, 2024

@aymericdamien Speaking of BDLSTM example, is there any particular reason for n_hidden to be 128? I'm hesitant to open an issue for this question since it's likely a triviality, but the number puzzles me.

from tensorflow-examples.