Thanks Tom - using the Moses version of symal rather than the MGiza version
fixed it (although still not sure why it should be different to the one I
built on my desktop). I hadn't realised they were different, as the
instructions on the Moses website state you should copy all binaries from
MGiza i
I'm out of ideas.
I've used aws and azure many times so it should work
Hieu Hoang
Sent while bumping into things
On Wed, 1 Aug 2018, 20:01 James Baker, wrote:
> Alas, nothing erroneous that I can see in the logs (using
> ./train-model.perl > output.log 2>&1), and neither the memory usage nor t
Alas, nothing erroneous that I can see in the logs (using
./train-model.perl > output.log 2>&1), and neither the memory usage nor the
used disk space went over 10% during the training.
James
On Wed, 1 Aug 2018 at 08:56, Hieu Hoang wrote:
> redirect stdout and stderr into a file and grep for 'er
redirect stdout and stderr into a file and grep for 'error'
that usually turns up something
Hieu Hoang
http://statmt.org/hieu
On 1 August 2018 at 17:38, James Baker wrote:
> Thanks Hieu,
>
> I'll give that a go this morning and keep an eye on the disk space and
> RAM, although I would be surpr
Thanks Hieu,
I'll give that a go this morning and keep an eye on the disk space and RAM,
although I would be surprised if that was the problem (I've got <3GB of
training data, 64GB of RAM, and 100GB of disk space). It also wouldn't
explain why binaries built on a different machine work, but binari
