PyAF_Introduction notebook hangs on train() call

[tfrojd]

Hi,

I am quite new to python so I might be screwing something up here but the same thing happened on my work and home computer.

I downloaded anaconda with python 3.6.1 and Jupiter 5.0.0 and followed the install instructions of PyAF from the GitHub repo. Running through the notebook it seems to get stuck on

lEngine.train(ozone_dataframe , 'Month' , 'Ozone', 12);

The text says it took you 7 seconds to train but for me it has been running for 15 min on a 2016 macbook and nothing happens except the output

INFO:pyaf.std:START_TRAINING 'Ozone'

I tried to run it in the ipython terminal with the same result.

Any ideas on how to figure out what is happening?

[antoinecarme]

Hi @tfrojd ,

Thanks for using PyAF and hope you'll enjoy.

Seems someone already had this type of problem (https://github.com/antoinecarme/pyaf/issues/39). Python latest version (3.6) is mentioned.

As I mentioned on the homepage., PyAF was tested on a 3.5 python. Also, sorry, not a mac user.

Can you install a python 3.5 or do you really need a 3.6 ?

Hope this helps. Don't hesitate if you have other questions.

[antoinecarme]

Just a question, when running ipython in a terminal, what happens when you interrupt the process (hit Ctrl-C) ? any logs to be copy-pasted here ?

[tfrojd]

Hi, thanks for your quick reply. I have been experimenting with grouped times series in the R forecast package a bit and PyAF seems really nice.

Gonna see if I can get Python 3.5 running instead.

If I Ctrl-C during the execution I get the following output

pyaflog.txt

[tfrojd]

Can confirm that a downgrade to Python 3.5.0 solved it.

For mac users running Anaconda it is as simple as

conda install python=3.5.0

[antoinecarme]

Thanks a lot for the fix, this will help future users.

Enjoy !!!