moses-smt/mosesdecoder
1
1
Fork 0
mirror of https://github.com/moses-smt/mosesdecoder.git synced 2024-09-20 07:42:21 +03:00
Code Issues Packages Projects Releases Wiki Activity

State, almost correct or correct

Browse Source
This commit is contained in:
XapaJIaMnu 2014-08-29 19:32:19 +01:00 committed by Paul Baltescu
parent c225fb27d1
commit 4309903d6e
1 changed files with 23 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

42
moses/LM/bilingual-lm/BilingualLM.cpp
View File

@ -59,6 +59,8 @@ int BilingualLM::getNeuralLMId(const Factor * factor) const{
}
//Populates words with amount words from the targetPhrase from the previous hypothesis where
//words[0] is the last word of the previous hypothesis, words[1] is the second last etc...
void BilingualLM::requestPrevTargetNgrams(const Hypothesis &cur_hypo, int amount, std::vector<int> &words) const {
const Hypothesis * prev_hyp = cur_hypo.GetPrevHypo();
int found = 0;
@ -273,36 +275,38 @@ size_t BilingualLM::getState(const Hypothesis& cur_hypo) const {
const TargetPhrase &targetPhrase = cur_hypo.GetCurrTargetPhrase();
size_t hashCode = 0;
//Check if we need to look at previous target phrases
bool previous_phrase_required;
int additional_needed = targetPhrase.GetSize() - target_ngrams;
if (additional_needed < 0) {
previous_phrase_required = true;
additional_needed = -additional_needed;
}
std::vector<int> prev_words(additional_needed);
requestPrevTargetNgrams(cur_hypo, additional_needed, prev_words);
for (int i=0; i<additional_needed; i++){ //This needs to be reverted.
boost::hash_combine(hashCode, prev_words[i]);
}
//Get the rest of the phrases needed
for (int i = 0; i < targetPhrase.GetSize(); i++) {
int neuralLM_wordID;
//Get last n-1 target phrases
size_t hashCode = 0;
for (int i = targetPhrase.GetSize() - target_ngrams; i < targetPhrase.GetSize(); i++) {
int neuralLM_wordID;
if (i < 0) {
break;
} else {
const Word& word = targetPhrase.GetWord(i);
const Factor* factor = word.GetFactor(0); //Parameter here is m_factorType, hard coded to 0
neuralLM_wordID = getNeuralLMId(factor);
}
boost::hash_combine(hashCode, neuralLM_wordID);
}
//In case we need to look at previous target phrases, request them here.
if (previous_phrase_required) {
std::vector<int> prev_words(additional_needed);
requestPrevTargetNgrams(cur_hypo, additional_needed, prev_words);
for (int i=0; i<additional_needed; i++){
boost::hash_combine(hashCode, prev_words[i]);
boost::hash_combine(hashCode, neuralLM_wordID);
}
} else {
for (int i = targetPhrase.GetSize() - target_ngrams; i < targetPhrase.GetSize(); i++) {
int neuralLM_wordID;
const Word& word = targetPhrase.GetWord(i);
const Factor* factor = word.GetFactor(0); //Parameter here is m_factorType, hard coded to 0
neuralLM_wordID = getNeuralLMId(factor);
boost::hash_combine(hashCode, neuralLM_wordID);
}
}
return hashCode;