Thank you Zahurul,
it's ok, but there is no config.ini.
Predict 0
After default: -l mem_free=0.5G -hard
usage: mert-moses-new.pl input-text references decoder-executable
decoder.ini
Options:
--working-dir=mert-dir ... where all the files are created
--nbest=100 ... how big nbestlist to generate
--jobs=N ... set this to anything to run moses in parallel
--mosesparallelcmd=STRING ... use a different script instead of
moses-parallel
--queue-flags=STRING ... anything you with to pass to
qsub, eg. '-l ws06osssmt=true'
The default is
-l mem_free=0.5G -hard
To reset the parameters, please use "--queue-flags=' '" (i.e.
a space between
the quotes).
--decoder-flags=STRING ... extra parameters for the decoder
--lambdas=STRING ... default values and ranges for lambdas, a complex
string
such as 'd:1,0.5-1.5 lm:1,0.5-1.5
tm:0.3,0.25-0.75;0.2,0.25-0.75;0.2,0.25-0.75;0.3,0.25-0.75;0,-0.5-0.5
w:0,-0.5-0.5'
--allow-unknown-lambdas ... keep going even if someone supplies a new
lambda
in the lambdas option (such as 'superbmodel:1,0-1'); optimize it,
too
--continue ... continue from the last achieved state
--skip-decoder ... skip the decoder run for the first time, assuming that
we got interrupted during optimization
--shortest ... Use shortest reference length as effective reference length
(mutually exclusive with --average and --closest)
--average ... Use average reference length as effective reference length
(mutually exclusive with --shortest and --closest)
--closest ... Use closest reference length as effective reference length
(mutually exclusive with --shortest and --average)
--nocase ... Do not preserve case information; i.e. case-insensitive
evaluation (default is false)
--nonorm ... Do not use text normalization (flag is not active, i.e. text
is NOT normalized)
--filtercmd=STRING ... path to filter-model-given-input.pl
--rootdir=STRING ... where do helpers reside (if not given explicitly)
--mertdir=STRING ... path to new mert implementation
--mertargs=STRING ... extra args for mert, eg to specify scorer
--pythonpath=STRING ... where is python executable
--scorenbestcmd=STRING ... path to score-nbest.py
--old-sge ... passed to moses-parallel, assume Sun Grid Engine < 6.0
--inputtype=[0|1|2] ... Handle different input types (0 for text, 1 for
confusion network, 2 for lattices, default is 0)
--no-filter-phrase-table ... disallow filtering of phrase tables
(useful if binary phrase tables are available)
--predictable-seeds ... provide predictable seeds to mert so that random
restarts are the same on every run
--efficient_scorenbest_flag ... activate a time-efficient scoring of nbest
lists
(this method is more memory-consumptive)
--activate-features=STRING ... comma-separated list of features to work
on
(if undef work on all features)
# (others are fixed to the starting
values)
--prev-aggregate-nbestlist=INT ... number of previous step to consider
when loading data (default =-1)
-1 means all previous, i.e. from
iteration 1
0 means no previous data, i.e. from
actual iteration
1 means 1 previous data , i.e. from
the actual iteration and from the previous one
and so on
Any idea?
Best
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support
