News for hackers. Stuff that matters.

A comment on Eager Execution: An imperative, define-by-run interface to TensorFlow. Parent comment

Re:Eager Execution: An imperative, define-by-run interface t... by ychujo2 on Tuesday October 31, 2017 @04:09PM (#15596132)
I think you mean Google is following the leadership of Chainer, like Facebook already does? PyTorch started as a Chainer fork. Its dynamic graph internals are all from Chainer.
[ Parent | on HN ]
Re:Eager Execution: An imperative, define-by-run interface t... by bradleyjg on Tuesday October 31, 2017 @04:30PM (#15596279)
This isn't art. There are no points for originality. If open source projects borrow the best parts from each other, that's a good thing.
[ Parent | on HN ]
Re:Eager Execution: An imperative, define-by-run interface t... by ychujo2 on Tuesday October 31, 2017 @07:08PM (#15597454)
It's not a bad thing. It's good for users. But give credit to the leaders in the field. If you make an iPod clone, you call it an iPod clone, not a clone of the Zume HD.

Chainer started it, was around years earlier, and it still has more users. So Google is not copying PyTorch, it's copying Chainer.

[ Parent | on HN ]
Re:Eager Execution: An imperative, define-by-run interface t... by arbie on Wednesday November 01, 2017 @06:34AM (#15600018)
Do you have a source for Chainer currently having more users than TensorFlow?
[ Parent | on HN ]
Re:Eager Execution: An imperative, define-by-run interface t... by Narew on Friday November 03, 2017 @04:55AM (#15617647)
Not more users than Tensorflow. But maybe more user than other dynamic deeplearning framework (PyTorch, Gluon, DyNet...)
[ Parent | on HN ]
Re:Eager Execution: An imperative, define-by-run interface t... by chrisprobert on Tuesday October 31, 2017 @04:53PM (#15596455)
Totally agree!
[ Parent | on HN ]
Re:Eager Execution: An imperative, define-by-run interface t... by ma2rten on Wednesday November 01, 2017 @12:11AM (#15598749)
This is the first time I am hearing this. I though pytorch was based on torch (like the name implies). Do you have a reference or more information?
[ Parent | on HN ]
Re:Eager Execution: An imperative, define-by-run interface t... by Narew on Friday November 03, 2017 @05:00AM (#15617665)
PyTorch use the same backend as torch (cutorch for GPU ...) But PyTorch use almost the same Python API than Chainer. On this point we can say PyTorch "copy" Chainer.
[ Parent | on HN ]