These are chat archives for soumith/convnet-benchmarks

4th
Mar 2016
Ekraam
@Ekraam
Mar 04 2016 18:49
Hi, has anyone tried multi-gpu implementation of LSTM? I'm trying Element-Research's implementation of LSTM in torch ( https://github.com/Element-Research/rnn ) and can't get all the gpu's to run in simultaneously. Thanks.
Elad Hoffer
@eladhoffer
Mar 04 2016 19:09
I managed to run on multiple gpus using https://github.com/eladhoffer/recurrent.torch
Ekraam
@Ekraam
Mar 04 2016 19:19
Thanks for the link