@fchollet Too bad the compilation time blows up.
@edersantana was looking into speeding it up with PyPy though?
Keras 1.0 will also be getting unrolled (scan-free) RNNs in Theano, made optional. Seems to yield a decent speedup (25% of so).
-
-
-
@mElantkowski@fchollet not actually :( I thought I'd have time for it T_T on the other hand 0.8 did some awesome optimization job already - Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.