Python 3 compatibility

[chanil1218]

Sugartensor is not compatible with Python 3(in my case, Python 3.5.2). Is there way to use sugartensor with Python 3?

[buriburisuri]

I have the plan to add some features in Dec.

1. Document
2. Python 3 support.
3. Docker support.

Sorry but you need python 2.7 interpreter now.

[buriburisuri]

@chanil1218

Now, we support python3 at sugartensor v0.0.2.1.

[chanil1218]

That's nice. Thank you for your efforts! 2016년 12월 6일 (화) 오후 2:17, Namju Kim notifications@github.com님이 작성:

@chanil1218 https://github.com/chanil1218

Now, we support python3 at sugartensor v0.0.2.1.

— You are receiving this because you were mentioned.

Reply to this email directly, view it on GitHub https://github.com/buriburisuri/sugartensor/issues/4#issuecomment-265063129, or mute the thread https://github.com/notifications/unsubscribe-auth/ABWrZU72-moF24Fk9QUefCf5tIF725hDks5rFO_mgaJpZM4K9IpS .

[nikparth]

I still have compatability issues with python3. Installing sugartensor 0.0.2.3 with tensorflow 0.12 I get the following error:

Traceback (most recent call last):
  File "train.py", line 51, in <module>
    .sg_periodic_shuffle(factor=2))
  File "/home/vision/anaconda3.5/lib/python3.5/site-packages/sugartensor/sg_main.py", line 151, in wrapper
    out = func(tensor, tf.sg_opt(kwargs))
  File "/home/vision/anaconda3.5/lib/python3.5/site-packages/sugartensor/sg_transform.py", line 556, in sg_periodic_shuffle
    for i in range(channel_target):
TypeError: 'float' object cannot be interpreted as an integer

I assume this has to do with the division in python 3 returning floats.

[LukasMosser]

Yes, that's correct. @buriburisuri pointed this out in another issue. Should be fixed soon. Maybe add a Pr?

[buriburisuri]

@nikparth I've patched this at ver 0.0.2.4. Sorry for late reply.