Hi,

We are developing a Transliteration machine and we have a problem taking
evaluation of Top2\Top5\Top10 results from Moses.
When we tell Moses to give us *n-best-list* we get duplication of words,
for example:

3 ||| c a r d i n a l s  ||| d: 0 -0.554311 0 0 -0.0434849 0 0 lm: -48.3492
w: -9 tm: -0.405465 -1.35896 -0.435318 -2.24059 1.99979 ||| -0.628302
3 ||| c a r d i n e l s  ||| d: 0 -0.18903 0 0 -0.100083 0 0 lm: -72.2741
w: -9 tm: -0.20067 -1.98284 -0.374693 -3.90471 1.99979 ||| -0.860206
3 ||| c a r d i n a l e s  ||| d: 0 -0.239436 0 0 -0.167054 0 0 lm:
-49.6022 w: -10 tm: -0.445845 -1.98283 -1.75803 -4.37584 1.99979 |||
-0.952976
3 ||| c a r d i n a l s  ||| d: 0 -0.250828 0 0 -0.189811 0 0 lm: -48.3492
w: -9 tm: -0.370756 -1.35896 -1.47988 -2.24058 2.99969 ||| -0.957803
3 ||| c a r d i n a l s  ||| d: 0 -0.53416 0 0 -0.0646484 0 0 lm: -48.3492
w: -9 tm: -0.245411 -1.35896 -1.2423 -2.24058 1.99979 ||| -0.958115

We've try to use the *distinct *flag, but nothing have changed.

How can we get results with different words (and not just different by the
search tree) for evaluating the Top-N words?

Thanks,
Guy.
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support

Reply via email to