3 ms·
Ask HN: Keras on PyPy?
Wondering if non-standard interpreters like pypy and cpython can be used to speed up training.
- pmiller2 7y agoIt seems like there's some progress, including a merged PR to tensorflow that should end up making this work, but I'm confused as to the exact status:https://github.com/tensorflow/tensorflow/issues/252 https://github.com/tensorflow/tensorflow/issues/252
- ghostbust555 7y agoAsked about this before. I was told no one cared since the real work is done in C and would not benefit from pypy. So the python level optimizations would not affect much.
- entilzha 7y agoNot entirely true. For good utilization you need both GPU/TPU ops to be fast (written in C), but that won’t get you far if your input pipeline (possibly written in python) is slow. I could imagine if all the TF calls work in PyPy, that it would help with throughout by speeding up the input pipeline and keeping the GPU/TPU saturated more effectively. One solution is to write everything to TFRecords, but for experimentation that’s kinda annoying.
- RocketSyntax 7y agoAh, I figured it was all Python. Now it makes sense why tensor objects are such a pain to work with. C++ 60.6% Python 30.6%
- garybake 7y agoI got stuck on this problem a while ago. I was doing reinforcement learning and the environment needed to run in pypy to make it fast enough. I ended up using Pyro4 to bridge pypy (environment) and CPython (keras).
- psv1 7y ago> non-standard interpreters like pypy and cpython Isn't CPython the default and most standard interpreter?