mirror of
https://github.com/mozilla/DeepSpeech.git
synced 2025-10-26 11:19:39 +00:00
328 lines
9.8 KiB
C++
328 lines
9.8 KiB
C++
#ifdef _MSC_VER
|
|
#include <stdlib.h>
|
|
#include <io.h>
|
|
#include <windows.h>
|
|
|
|
#define R_OK 4 /* Read permission. */
|
|
#define W_OK 2 /* Write permission. */
|
|
#define F_OK 0 /* Existence. */
|
|
|
|
#define access _access
|
|
|
|
#else /* _MSC_VER */
|
|
#include <unistd.h>
|
|
#endif
|
|
|
|
#include "scorer.h"
|
|
#include <iostream>
|
|
#include <fstream>
|
|
|
|
#include "lm/config.hh"
|
|
#include "lm/model.hh"
|
|
#include "lm/state.hh"
|
|
#include "util/string_piece.hh"
|
|
|
|
#include "decoder_utils.h"
|
|
|
|
using namespace lm::ngram;
|
|
|
|
static const int32_t MAGIC = 'TRIE';
|
|
static const int32_t FILE_VERSION = 4;
|
|
|
|
int
|
|
Scorer::init(double alpha,
|
|
double beta,
|
|
const std::string& lm_path,
|
|
const std::string& trie_path,
|
|
const Alphabet& alphabet)
|
|
{
|
|
reset_params(alpha, beta);
|
|
alphabet_ = alphabet;
|
|
setup(lm_path, trie_path);
|
|
return 0;
|
|
}
|
|
|
|
int
|
|
Scorer::init(double alpha,
|
|
double beta,
|
|
const std::string& lm_path,
|
|
const std::string& trie_path,
|
|
const std::string& alphabet_config_path)
|
|
{
|
|
reset_params(alpha, beta);
|
|
int err = alphabet_.init(alphabet_config_path.c_str());
|
|
if (err != 0) {
|
|
return err;
|
|
}
|
|
setup(lm_path, trie_path);
|
|
return 0;
|
|
}
|
|
|
|
void Scorer::setup(const std::string& lm_path, const std::string& trie_path)
|
|
{
|
|
// (Re-)Initialize character map
|
|
char_map_.clear();
|
|
|
|
SPACE_ID_ = alphabet_.GetSpaceLabel();
|
|
|
|
for (int i = 0; i < alphabet_.GetSize(); i++) {
|
|
// The initial state of FST is state 0, hence the index of chars in
|
|
// the FST should start from 1 to avoid the conflict with the initial
|
|
// state, otherwise wrong decoding results would be given.
|
|
char_map_[alphabet_.StringFromLabel(i)] = i + 1;
|
|
}
|
|
|
|
// load language model
|
|
const char* filename = lm_path.c_str();
|
|
VALID_CHECK_EQ(access(filename, R_OK), 0, "Invalid language model path");
|
|
|
|
bool has_trie = trie_path.size() && access(trie_path.c_str(), R_OK) == 0;
|
|
|
|
lm::ngram::Config config;
|
|
|
|
if (!has_trie) { // no trie was specified, build it now
|
|
RetrieveStrEnumerateVocab enumerate;
|
|
config.enumerate_vocab = &enumerate;
|
|
language_model_.reset(lm::ngram::LoadVirtual(filename, config));
|
|
auto vocab = enumerate.vocabulary;
|
|
for (size_t i = 0; i < vocab.size(); ++i) {
|
|
if (is_character_based_ && vocab[i] != UNK_TOKEN &&
|
|
vocab[i] != START_TOKEN && vocab[i] != END_TOKEN &&
|
|
get_utf8_str_len(enumerate.vocabulary[i]) > 1) {
|
|
is_character_based_ = false;
|
|
}
|
|
}
|
|
// fill the dictionary for FST
|
|
if (!is_character_based()) {
|
|
fill_dictionary(vocab, true);
|
|
}
|
|
} else {
|
|
config.load_method = util::LoadMethod::LAZY;
|
|
language_model_.reset(lm::ngram::LoadVirtual(filename, config));
|
|
|
|
// Read metadata and trie from file
|
|
std::ifstream fin(trie_path, std::ios::binary);
|
|
|
|
int magic;
|
|
fin.read(reinterpret_cast<char*>(&magic), sizeof(magic));
|
|
if (magic != MAGIC) {
|
|
std::cerr << "Error: Can't parse trie file, invalid header. Try updating "
|
|
"your trie file." << std::endl;
|
|
throw 1;
|
|
}
|
|
|
|
int version;
|
|
fin.read(reinterpret_cast<char*>(&version), sizeof(version));
|
|
if (version != FILE_VERSION) {
|
|
std::cerr << "Error: Trie file version mismatch (" << version
|
|
<< " instead of expected " << FILE_VERSION
|
|
<< "). Update your trie file."
|
|
<< std::endl;
|
|
throw 1;
|
|
}
|
|
|
|
fin.read(reinterpret_cast<char*>(&is_character_based_), sizeof(is_character_based_));
|
|
|
|
if (!is_character_based_) {
|
|
fst::FstReadOptions opt;
|
|
opt.mode = fst::FstReadOptions::MAP;
|
|
opt.source = trie_path;
|
|
dictionary.reset(FstType::Read(fin, opt));
|
|
}
|
|
}
|
|
|
|
max_order_ = language_model_->Order();
|
|
}
|
|
|
|
void Scorer::save_dictionary(const std::string& path)
|
|
{
|
|
std::ofstream fout(path, std::ios::binary);
|
|
fout.write(reinterpret_cast<const char*>(&MAGIC), sizeof(MAGIC));
|
|
fout.write(reinterpret_cast<const char*>(&FILE_VERSION), sizeof(FILE_VERSION));
|
|
fout.write(reinterpret_cast<const char*>(&is_character_based_), sizeof(is_character_based_));
|
|
if (!is_character_based_) {
|
|
fst::FstWriteOptions opt;
|
|
opt.align = true;
|
|
opt.source = path;
|
|
dictionary->Write(fout, opt);
|
|
}
|
|
}
|
|
|
|
double Scorer::get_log_cond_prob(const std::vector<std::string>& words,
|
|
bool bos,
|
|
bool eos)
|
|
{
|
|
return get_log_cond_prob(words.begin(), words.end(), bos, eos);
|
|
}
|
|
|
|
double Scorer::get_log_cond_prob(const std::vector<std::string>::const_iterator& begin,
|
|
const std::vector<std::string>::const_iterator& end,
|
|
bool bos,
|
|
bool eos)
|
|
{
|
|
const auto& vocab = language_model_->BaseVocabulary();
|
|
lm::ngram::State state_vec[2];
|
|
lm::ngram::State *in_state = &state_vec[0];
|
|
lm::ngram::State *out_state = &state_vec[1];
|
|
|
|
if (bos) {
|
|
language_model_->BeginSentenceWrite(in_state);
|
|
} else {
|
|
language_model_->NullContextWrite(in_state);
|
|
}
|
|
|
|
double cond_prob = 0.0;
|
|
for (auto it = begin; it != end; ++it) {
|
|
lm::WordIndex word_index = vocab.Index(*it);
|
|
|
|
// encounter OOV
|
|
if (word_index == lm::kUNK) {
|
|
return OOV_SCORE;
|
|
}
|
|
|
|
cond_prob = language_model_->BaseScore(in_state, word_index, out_state);
|
|
std::swap(in_state, out_state);
|
|
}
|
|
|
|
if (eos) {
|
|
cond_prob = language_model_->BaseScore(in_state, vocab.EndSentence(), out_state);
|
|
}
|
|
|
|
// return loge prob
|
|
return cond_prob/NUM_FLT_LOGE;
|
|
}
|
|
|
|
double Scorer::get_sent_log_prob(const std::vector<std::string>& words)
|
|
{
|
|
// For a given sentence (`words`), return sum of LM scores over windows on
|
|
// sentence. For example, given the sentence:
|
|
//
|
|
// there once was an ugly barnacle
|
|
//
|
|
// And a language model with max_order_ = 3, this function will return the sum
|
|
// of the following scores:
|
|
//
|
|
// there | <s>
|
|
// there once | <s>
|
|
// there once was
|
|
// once was an
|
|
// was an ugly
|
|
// an ugly barnacle
|
|
// ugly barnacle </s>
|
|
//
|
|
// This is used in the decoding process to compute the LM contribution for a
|
|
// given beam's accumulated score, so that it can be removed and only the
|
|
// acoustic model contribution can be returned as a confidence score for the
|
|
// transcription. See DecoderState::decode.
|
|
const int sent_len = words.size();
|
|
|
|
double score = 0.0;
|
|
for (int win_start = 0, win_end = 1; win_end <= sent_len+1; ++win_end) {
|
|
const int win_size = win_end - win_start;
|
|
bool bos = win_size < max_order_;
|
|
bool eos = win_end == (sent_len + 1);
|
|
|
|
// The last window goes one past the end of the words vector as passing the
|
|
// EOS=true flag counts towards the length of the scored sentence, so we
|
|
// adjust the win_end index here to not go over bounds.
|
|
score += get_log_cond_prob(words.begin() + win_start,
|
|
words.begin() + (eos ? win_end - 1 : win_end),
|
|
bos,
|
|
eos);
|
|
|
|
// Only increment window start position after we have a full window
|
|
if (win_size == max_order_) {
|
|
win_start++;
|
|
}
|
|
}
|
|
|
|
return score / NUM_FLT_LOGE;
|
|
}
|
|
|
|
void Scorer::reset_params(float alpha, float beta)
|
|
{
|
|
this->alpha = alpha;
|
|
this->beta = beta;
|
|
}
|
|
|
|
std::vector<std::string> Scorer::split_labels(const std::vector<int>& labels)
|
|
{
|
|
if (labels.empty()) return {};
|
|
|
|
std::string s = alphabet_.LabelsToString(labels);
|
|
std::vector<std::string> words;
|
|
if (is_character_based_) {
|
|
words = split_utf8_str(s);
|
|
} else {
|
|
words = split_str(s, " ");
|
|
}
|
|
return words;
|
|
}
|
|
|
|
std::vector<std::string> Scorer::make_ngram(PathTrie* prefix)
|
|
{
|
|
std::vector<std::string> ngram;
|
|
PathTrie* current_node = prefix;
|
|
PathTrie* new_node = nullptr;
|
|
|
|
for (int order = 0; order < max_order_; order++) {
|
|
std::vector<int> prefix_vec;
|
|
std::vector<int> prefix_steps;
|
|
|
|
if (is_character_based_) {
|
|
new_node = current_node->get_path_vec(prefix_vec, prefix_steps, SPACE_ID_, 1);
|
|
current_node = new_node;
|
|
} else {
|
|
new_node = current_node->get_path_vec(prefix_vec, prefix_steps, SPACE_ID_);
|
|
current_node = new_node->parent; // Skipping spaces
|
|
}
|
|
|
|
// reconstruct word
|
|
std::string word = alphabet_.LabelsToString(prefix_vec);
|
|
ngram.push_back(word);
|
|
|
|
if (new_node->character == -1) {
|
|
break;
|
|
}
|
|
}
|
|
std::reverse(ngram.begin(), ngram.end());
|
|
return ngram;
|
|
}
|
|
|
|
void Scorer::fill_dictionary(const std::vector<std::string>& vocabulary, bool add_space)
|
|
{
|
|
// ConstFst is immutable, so we need to use a MutableFst to create the trie,
|
|
// and then we convert to a ConstFst for the decoder and for storing on disk.
|
|
fst::StdVectorFst dictionary;
|
|
// For each unigram convert to ints and put in trie
|
|
for (const auto& word : vocabulary) {
|
|
add_word_to_dictionary(word, char_map_, add_space, SPACE_ID_ + 1, &dictionary);
|
|
}
|
|
|
|
/* Simplify FST
|
|
|
|
* This gets rid of "epsilon" transitions in the FST.
|
|
* These are transitions that don't require a string input to be taken.
|
|
* Getting rid of them is necessary to make the FST deterministic, but
|
|
* can greatly increase the size of the FST
|
|
*/
|
|
fst::RmEpsilon(&dictionary);
|
|
std::unique_ptr<fst::StdVectorFst> new_dict(new fst::StdVectorFst);
|
|
|
|
/* This makes the FST deterministic, meaning for any string input there's
|
|
* only one possible state the FST could be in. It is assumed our
|
|
* dictionary is deterministic when using it.
|
|
* (lest we'd have to check for multiple transitions at each state)
|
|
*/
|
|
fst::Determinize(dictionary, new_dict.get());
|
|
|
|
/* Finds the simplest equivalent fst. This is unnecessary but decreases
|
|
* memory usage of the dictionary
|
|
*/
|
|
fst::Minimize(new_dict.get());
|
|
|
|
// Now we convert the MutableFst to a ConstFst (Scorer::FstType) via its ctor
|
|
std::unique_ptr<FstType> converted(new FstType(*new_dict));
|
|
this->dictionary = std::move(converted);
|
|
}
|