Tensorflow tf.reshape() seems to behave differently to numpy.reshape()

525 views
Skip to first unread message

Maik Vogt

unread,
Sep 24, 2016, 12:02:03 PM9/24/16
to Discuss

I'm trying to train a LSTM network and it trains successfully in one way, but throws an error in the other way.

In the first example I reshape the input array X using numpy reshape and in the other way I reshape it using tensorflow reshape.

I posted the code on stackoverflow, but without any success yet.
Does the fit method of tensorflow.contrib.learn.TensorFlowEstimator transform the input vector X before it is 
passed to model_fn? 

http://stackoverflow.com/questions/39671253/tensorflow-tf-reshape-seems-to-behave-differently-to-numpy-reshape


Martin Wicke

unread,
Sep 24, 2016, 12:23:03 PM9/24/16
to Maik Vogt, Discuss
TensorFlowEstimator is deprecated and will disappear, please use Estimator. Inputs are not transformed before they reach the model_fn.
--
You received this message because you are subscribed to the Google Groups "Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to discuss+u...@tensorflow.org.
To post to this group, send email to dis...@tensorflow.org.
To view this discussion on the web visit https://groups.google.com/a/tensorflow.org/d/msgid/discuss/cb33f74b-8662-4b6e-98dd-4053917ce6cf%40tensorflow.org.
Reply all
Reply to author
Forward
0 new messages