[RBFOpt] GPU training and testing LSTM Networks for Optimizing Hyper-Parameters

[Giacomo Nannicini]

Hi David,
apologies for the late reply. There was a technical problem with the
mailing lists and messages were held for a very long time. Things should be
back to normal now.
Unfortunately, I am not sure I can help with your issue. It seems to be a
problem due to the neural network framework, not to RBFOpt. You should
check with the TensorFlow community. If the issue is related to RBFOpt,
then please provide more details.

Thanks

Giacomo

On Fri, Aug 24, 2018 at 12:32 PM David Jimenez <
david.jimenez.barrero at outlook.com> wrote:

> Dear community,
>
> I am experiencing a "Segmentation Fault (Core Dumped)" error when trying
> to optimize the parameters of a Long Short Term Memory Neural Network. It
> does not happen right away, but after about 4 models trained and tested. I
> wonder if someone has experienced something similar.  It happens when I use
> the GPU for training, but not when I use CPU. However, CPU is slower.
>
> Moreover, I am using Keras with TensorFlow backend. I really appreciate
> your help.
>
> Kind regards,
>
> David Jimenez
> _______________________________________________
> RBFOpt mailing list
> RBFOpt at list.coin-or.org
> https://list.coin-or.org/mailman/listinfo/rbfopt
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/rbfopt/attachments/20180824/57854510/attachment.html>