f90/Wave-U-Net

How to Install Tensorflow on Python 2.7 ?

Closed this issue · 7 comments

brunobulgaron commented

Hi there.
I'm having this error while trying to install Tensorflow on Python 2.7

"Could not find a version that satisfies the requirement tensorflow-gpu==1.8.0 (from -r requirements.txt (line 3)) (from versions: )
No matching distribution found for tensorflow-gpu==1.8.0 (from -r requirements.txt (line 3))"

I'm not very familiar with these things, is there anything that I'm missing before installing this?

Thanks!

f90 commented

Hey, which system are you on? MacOS, Windows, Linux?
Check that you are actually using a Python 2.7 environment with python --version. Also make sure your pip is up-to-date:

pip install --upgrade pip
brunobulgaron commented

I'm using Windows 7 64bit, Python 2.7.13 and Pip's version is 19.1

brunobulgaron commented

I've tried running the command "python Predict.py with cfg.full_44KHz" and got this error (screenshot attached)
Sem título

f90 commented

Yes that looks like Tensorflow is not properly installed yet. If you run python, and then execute something like

import tensorflow as tf
sess = tf.Session()

does that work or also throw an error? If that gives you an error as well, it is definitely not a problem on my end/in my Wave-U-Net code, so in that case please consult online sources to properly install Tensorflow on your system first. A quick search on Google brought up a lot of Windows people having trouble with pip installation and many suggested solutions, so that would be a good start for you.
Good luck!

f90 commented

Did you make it work in the end?

HIN0209 commented

"TensorFlow supports Python 3.5.x and 3.6.x on Windows." Hence you can not use tensorflow with Python 2.7 on Windows.
https://stackoverflow.com/questions/50044649/how-to-install-tensorflow-for-python-2-7#

f90 commented

Thanks for the information on that, @HIN0209
Going to close this since OP needs to switch to Linux/MacOS. Alternatively one can try Python 3, which I think my code is fully or almost fully compatible with, but I dont give any guarantees on correctness there. If someone wants to port it properly and submit a pull request however, I am happy to consider switching the codebase over to Python 3!