问题
What is the upside of using tf.nn.rnn instead of tf.nn.dynamic_rnn?
The documentation says :
[
dynamic_rnn
] is functionally identical to the functionrnn
above, but performs fully dynamic unrolling of inputs.
Is there any case where one might prefer to use tf.nn.rnn
instead of tf.nn.dynamic_rnn
?
来源:https://stackoverflow.com/questions/42356027/what-is-the-upside-of-using-tf-nn-rnn-instead-of-tf-nn-dynamic-rnn-in-tensor