3 ms·

All stable versions of the big ML frameworks are still 2.7.

• mastazi 10y ago
  Even if your statement was true (it's not)[1][2][3][4], how would that be a sign that Python is dying? [1] http://www.deeplearning.net/software/theano/install.html http://www.deeplearning.net/software/theano/install.html [2] https://www.tensorflow.org/versions/r0.11/get_started/os_setup.html https://www.tensorflow.org/versions/r0.11/get_started/os_set... [3] https://keras.io/#you-have-just-found-keras https://keras.io/#you-have-just-found-keras [4] https://lasagne.readthedocs.io/en/latest/user/installation.html https://lasagne.readthedocs.io/en/latest/user/installation.h...
  • aminorex 10y ago
    I think the author's point was not that python3 is dying. His point was that it deserved death, that 2.7 is flourishing, while 3 is failing to thrive, and reasons for this were postulated.
• p10_user 10y ago
  Are there many ML frameworks that only work in 2.7? Scipy and numpy work in both 2.7 and 3.x, so I'd imagine (without being heavy in the field) that any new development using these scientific computing packages could be Python version agnostic as well. Is this generally not the case?
• AstralStorm 10y ago
  Not true. Please specify which ML library authors are obtuse. About the only thing that is dead like that seems to be Enthought project for drawing and creating UI for graphs.
  • hartator 10y ago
    Sorry, should have said run better. There is support for latest 3.x branch, but everyone uses 2.7 as it's less bugguy.