martin-gorner/tensorflow-rnn-shakespeare

ValueError in my_txtutils.py

Opened this issue · 4 comments

yingdongucas commented

When run rnn_train.py, I got the following error:

Traceback (most recent call last):
File "/tensorflow-rnn-shakespeare/rnn_train.py", line 148, in
txt.print_learning_learned_comparison(x, y, l, bookranges, bl, acc, epoch_size, step, epoch)
File "/tensorflow-rnn-shakespeare/my_txtutils.py", line 180, in print_learning_learned_comparison
footer = format_string.format('INDEX', 'BOOK NAME', 'TRAINING SEQUENCE', 'PREDICTED SEQUENCE', 'LOSS')
ValueError: Invalid conversion specification

I use Python 2.7.6 and tensorflow 1.1.0 on Ubuntu 14.04. How can I fix this? Any reply will be very much appreciated.

martin-gorner commented

This looks like a Python 2 / Python 3 discrepancy but theoretically "".format() works in Pyhon 2.7. i do not have an immediate answer. I will have to test under Python 2.7.6.

  • 👍1
mahmood431226 commented

Hi Martin
I study many script and code in tensorflow and python. It is obvious in all of them they use tf.reduce _mean() function after tf.nn.softmax_cross_entropy_with_logits to make loss but you don't, my question is Why????????

hcchengithub commented

see issue #15 , no problem on python3, windows 10, tensorflow 1.1.0.

martin-gorner commented

@mahmood431226 it is because on a mini-batch of input images, you get one loss per image. The loss is supposed to be a single scalar though so you average them.