Hello, I have very strange behavior of train-model.perl. Here it is the tail of training.out:
Entire Viterbi H333444 Training took: 3.43885e+06 seconds ========================================================== Entire Training took: 4.54946e+06 seconds Program Finished at: Thu Feb 12 18:14:07 2015 ========================================================== Merging A3.final.part* tables Executing: /opt/moses/mgiza/merge_alignment.py /root/working/train/giza.es-en/es-en.A3.final.part*>/root/working/train/giza.es-en/es-en.A3.final Combined 16 files, totally 89889896 sents Executing: rm -f /root/working/train/giza.es-en/es-en.A3.final.gz Executing: gzip /root/working/train/giza.es-en/es-en.A3.final The problem is that we have gzip process. $ ps ax | grep gzip 11087 pts/0 S+ 0:00 grep --color=auto gzip Looks like train-model.perl died silently. Why it could occur? Can I restart process from the moment when train-model.perl died or I need to start from scratch? It will be a disaster, frankly speaking.
_______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
