Update to Python3 and Windows

[BSalita]

I've modified source files for compatibility with Python3 at this repo. On Windows, Tensorflow only works with Python3, not Python2. Hence, it was necessary to update to Python3. You may wish to create a Python3 branch of IDNNs and use my mods. Cheers.

To install on Windows, install Anaconda3, then tensorflow and joblib.

conda install tensorflow-gpu joblib

Execute idnns.bat for a quick test of IDNNs.

[kir1994]

I ran your script and got error

IndexError: only integers, slices (:), ellipsis (...), numpy.newaxis (None) and integer or boolean arrays are valid indices

in file plot_figures.py, line

results = [[data[j,k][name] for k in range(data.shape[1])] for j in range(data.shape[0])]

Seems that something went wrong with data format.
Also I tried to extract the data from files in jobs directory and failed since unpickling in python3 gives just multidimensional array of numbers, not 2-dimensional array of dicts

[ravidziv]

now it should run with python 3

[BSalita]

I can now run the repos on Python3 on Windows using default parameters obtaining expected results. However, I'm having the same issue as @kir1994 when I add the parameter "-num_repeat=2". I've also tried other repeats which give the same error. Possibly all repeat values fail. It's been very easy to obtain this same error message in the previous and new versions. It would be good if you can check for the condition and add a graceful error message explaining why the condition occurs. The condition is that "IXT_vartional" is not in the dictionary of "data" but I'm unsure about the underlying cause.

[kir1994]

@BSalita Do you experience the same issues when extracting data from repo files in /jobs/ directory?

Talking about the num_repeat issue, I suppose it occurs because the array becomes 3-dim (repeat number as an extra one) and extract_array func is not designed to handle this.

[1098707338]

@kir1994 could you tell me how long time did you spend on running the main.py under the spyder 3? and do you have any more guide details about how to run these files? Thanks.

[1098707338]

@kir1994 how did you solve the problem of results = [[data[j,k][name] for k in range(data.shape[1])] for j in range(data.shape[0])], thanks

[happinesszl]

@kir1994 @BSalita @ravidziv Hi, guys. Have you solved the num_repeat issue?