Closed kastnerkyle closed 7 years ago
I also noticed this issue. Feel free to send PR if you know how to fix it. I guess the fix is simple.
I think we can check the version something like this:
import tensorflow as tf
print(tf.__version__)
Thanks for the report @kastnerkyle. Could you perhaps send a PR for your try except fix?
I will see if I get some time, if someone doesn't beat me to it
On Thu, Jun 9, 2016 at 5:03 AM, Sherjil Ozair notifications@github.com wrote:
Thanks for the report @kastnerkyle https://github.com/kastnerkyle. Could you perhaps send a PR for your try except fix?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/sherjilozair/char-rnn-tensorflow/issues/31#issuecomment-224875032, or mute the thread https://github.com/notifications/unsubscribe/ABfbHQ7hYpf1e1yGL3w_el1IW908vwDwks5qKAEHgaJpZM4IwJGQ .
I think these are under tensorflow.contrib.
This was working for me:
from tensorflow.contrib import rnn
from tensorflow.contrib import seq2seq
Everything should be working now with TF 1.0. The travis ci builds should help keep this repo up to date with latest TF.
Thanks all!
After building from HEAD of TF, I get some import errors for various things (almost every example on the web has this problem!)
I had to change imports in model.py to
Downside is that importing
tensorflow.models.rnn
raises ImportError, so something like this is probably needed to preserve backwards compat: