1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
|
#ifndef _DTRAIN_KBESTGET_H_
#define _DTRAIN_KBESTGET_H_
#include "kbest.h"
namespace dtrain
{
struct Samples
{
vector<SparseVector<double> > feats;
vector<vector<WordID> > sents;
vector<double> model_scores;
vector<double> scores;
size_t GetSize() { return sents.size(); }
};
struct HypoSampler : public DecoderObserver
{
virtual Samples* GetSamples() {}
};
struct KBestGetter : public HypoSampler
{
const size_t k_;
const string filter_type;
Samples s;
KBestGetter(const size_t k, const string filter_type) :
k_(k), filter_type(filter_type) {}
virtual void
NotifyTranslationForest(const SentenceMetadata& smeta, Hypergraph* hg)
{
KBest(*hg);
}
Samples* GetSamples() { return &s; }
void
KBest(const Hypergraph& forest)
{
if (filter_type == "unique") {
KBestUnique(forest);
} else if (filter_type == "no") {
KBestNoFilter(forest);
}
}
void
KBestUnique(const Hypergraph& forest)
{
s.sents.clear();
s.feats.clear();
s.model_scores.clear();
s.scores.clear();
KBest::KBestDerivations<vector<WordID>, ESentenceTraversal, KBest::FilterUnique, prob_t, EdgeProb> kbest(forest, k_);
for (size_t i = 0; i < k_; ++i) {
const KBest::KBestDerivations<vector<WordID>, ESentenceTraversal, KBest::FilterUnique, prob_t, EdgeProb>::Derivation* d =
kbest.LazyKthBest(forest.nodes_.size() - 1, i);
if (!d) break;
s.sents.push_back(d->yield);
s.feats.push_back(d->feature_values);
s.model_scores.push_back(log(d->score));
}
}
void
KBestNoFilter(const Hypergraph& forest)
{
s.sents.clear();
s.feats.clear();
s.model_scores.clear();
s.scores.clear();
KBest::KBestDerivations<vector<WordID>, ESentenceTraversal> kbest(forest, k_);
for (size_t i = 0; i < k_; ++i) {
const KBest::KBestDerivations<vector<WordID>, ESentenceTraversal>::Derivation* d =
kbest.LazyKthBest(forest.nodes_.size() - 1, i);
if (!d) break;
s.sents.push_back(d->yield);
s.feats.push_back(d->feature_values);
s.model_scores.push_back(log(d->score));
}
}
};
} // namespace
#endif
|
