Is there anyone convert it into py3.x version?

[Stevewsw]

it is grateful if anyone rewrite it by py3.x .

[Zyjacya-In-love]

I have the same requirement. Have you solved it

[genaris]

Hi all, I'm working on a migration to Python 3 and Tensorflow 1.x. Seems to be working fine under Python 3.5.4 and tensorflow-gpu-1.15.2. Check my fork.

Anyway I'll send a pull request to the authors so they can test it and confirm it is fine (there were several changes to the RNN API since this original work has been done, and still some others to do if looking forward to a migration to tf 2.x).

[wangzhang021]

Hi all, I'm working on a migration to Python 3 and Tensorflow 1.x. Seems to be working fine under Python 3.5.4 and tensorflow-gpu-1.15.2. Check my fork.

Anyway I'll send a pull request to the authors so they can test it and confirm it is fine (there were several changes to the RNN API since this original work has been done, and still some others to do if looking forward to a migration to tf 2.x).

Hi, @genaris!
I am trying to reproduce this project, and get in trouble with the bugs.
When executing your work, I still can't succeed.
I failed in prepare_physionet.py:
_, _, ann = list(zip(*reader_ann.records()))

Returns
TypeError: catching classes that do not inherit from BaseException is not allowed

I reinstalled the dependency mne version to 0.15.2, but it does not work.
And my Python is 3.6
Could you please help me out? Thank you!

[genaris]

Hi @wangzhang021 , you are right. prepare_physionet.py was not updated (I've downloaded data with original py2.7 setup), so the repo was not in a consistent state.

I've just commited some changes: it should work now on python3 with mne 0.2.0 (latest version). However, it is still work in progress, as prepare_physionet.py code is using a deprecated parameter (scaling_time) that will not be available on mne 0.2.1. Hopefully I'll update it in the following days.