python 2.7 can't install tensorflow now

[DuckDuckBug]

Hello, I am very glad to find this project. Nowadays, I am looking for a pratical method to detect web-based  attacks and your idea is so cool and creative. But I meet some problems when trying to run this project.
First,  I have installed anaconda with python 3.6 in my pc (its os is win 10), I attended to use python 2.7 to run your project, but python 2.7 can't install tensorflow now.
Second,  now I changed your code to python 3, and I can run it. But I found a bug in your code. In the function string_to_int of class Vocabulary in vocab.py, you wrote "text = text.decode('utf-8')", which makes characters become ['<UNK>']. I don't know how to fix this problem, so I delete this line. Finally, the project can run, but it is stuck in the training step.
Finally, the post of this mentioned the training time is very long, how can I shorten the training time if I want to see the accurancy and adjest the model quickly?

[amurzina]

Hi!
To avoid all problems with python version and dependencies just to try to run a docker image.
docker run -it -p 8888:8888 montekki/seq2seq-web-attack-detection:latest jupyter notebook --ip=0.0.0.0

It downloads an image and runs a python notebook with all needed requirements.
You can skip the training phase and start with parameters settings. It loads saved model checkpoints and you can modify the code to check what you need.

I guess this is the fastest way to check what you want.