Hi Hiue,
Unfortunately, I use the new scripts to create phrase tables.
The hierarchical model has the new format.
[X][X] 五 百 [X] ||| [X][X] five hundred [X] ||| 0.944295
0.44408 0.595991 0.478547 2.718 ||| 0-0 ||| 12.3925 19.6349
The phrase-based model also has the new
hi yonee
1. i didn't see any problems with it. can you give some more details of
where it conditonally branched because of the #define
2. when line is empty, line[0] == 0x. therefore, it's still valid. i think.
had me worried there
On 11/08/2010 02:44, Yonee Yang wrote:
> hi:
> 1.
> 96th line
CALL FOR PAPERS
---
Workshop on Annotation and Exploitation of Parallel Corpora (AEPC)
December 2, 2010, University of Tartu, Estonia
http://math.ut.ee/tlt9/aepc/
Co-located with TLT9 - The Ninth International Workshop on Treebanks
and Lingu
hi:
1.
96th line in "PhraseDictionary.cpp "
assert(pdm->Load(m_input
, m_output
, m_filePath
, m_weight
Hi Mark,
I have checked that the weight of the laguage model is -0.000345 in
the tuned moses.ini,
which is almost the same as you anaylzed :-)
I have replaced the source langauge model in moses.ini file generated
by the training step with the in-domain target langauge model, and
then went on to
Hi Hwidong
Can you send me a few lines of the phrase table for the hierarchical model.
The 1 change that we've done in the last few weeks which is breaking
previous experiments is the phrase table format has changed from
source ||| target ||| [alignment] ||| scores ||| [count]
to
sou
