brannondorsey / PassGAN

A Deep Learning Approach for Password Guessing (https://arxiv.org/abs/1709.00440)
MIT License
1.77k stars 360 forks source link

Why use Tensorflow==1.4.1? #11

Closed bpartin2009 closed 5 years ago

bpartin2009 commented 6 years ago

Is there a reason requirments.txt says tensorflow==1.4.1 and not the latest like 1.7.0 compatible with CUDA 9.1? Is it just what you had/have at the time or is there other library or dependency issues that pose an issue?

brannondorsey commented 6 years ago

I'm using tensorflow==1.4.1 because that is what the original authors of https://github.com/igul222/improved_wgan_training used, and I'm leveraging most of their code. I haven't tried with any other versions. It might work great, or it might need some love. PRs welcome ;)

brannondorsey commented 6 years ago

I've confirmed 1.4.1 to work, hence why it is in the requirements.txt.