Wei Kang
Committed by GitHub

Support customize scores for hotwords (#926)

* Support customize scores for hotwords

* Skip blank lines
... ... @@ -61,10 +61,9 @@ class ContextGraph {
}
ContextGraph(const std::vector<std::vector<int32_t>> &token_ids,
float context_score, const std::vector<float> &scores = {},
const std::vector<std::string> &phrases = {})
: ContextGraph(token_ids, context_score, 0.0f, scores, phrases,
std::vector<float>()) {}
float context_score, const std::vector<float> &scores = {})
: ContextGraph(token_ids, context_score, 0.0f, scores,
std::vector<std::string>(), std::vector<float>()) {}
std::tuple<float, const ContextState *, const ContextState *> ForwardOneStep(
const ContextState *state, int32_t token_id,
... ...
... ... @@ -145,15 +145,35 @@ class OfflineRecognizerTransducerImpl : public OfflineRecognizerImpl {
auto hws = std::regex_replace(hotwords, std::regex("/"), "\n");
std::istringstream is(hws);
std::vector<std::vector<int32_t>> current;
std::vector<float> current_scores;
if (!EncodeHotwords(is, config_.model_config.modeling_unit, symbol_table_,
bpe_encoder_.get(), &current)) {
bpe_encoder_.get(), &current, &current_scores)) {
SHERPA_ONNX_LOGE("Encode hotwords failed, skipping, hotwords are : %s",
hotwords.c_str());
}
int32_t num_default_hws = hotwords_.size();
int32_t num_hws = current.size();
current.insert(current.end(), hotwords_.begin(), hotwords_.end());
auto context_graph =
std::make_shared<ContextGraph>(current, config_.hotwords_score);
if (!current_scores.empty() && !boost_scores_.empty()) {
current_scores.insert(current_scores.end(), boost_scores_.begin(),
boost_scores_.end());
} else if (!current_scores.empty() && boost_scores_.empty()) {
current_scores.insert(current_scores.end(), num_default_hws,
config_.hotwords_score);
} else if (current_scores.empty() && !boost_scores_.empty()) {
current_scores.insert(current_scores.end(), num_hws,
config_.hotwords_score);
current_scores.insert(current_scores.end(), boost_scores_.begin(),
boost_scores_.end());
} else {
// Do nothing.
}
auto context_graph = std::make_shared<ContextGraph>(
current, config_.hotwords_score, current_scores);
return std::make_unique<OfflineStream>(config_.feat_config, context_graph);
}
... ... @@ -226,13 +246,13 @@ class OfflineRecognizerTransducerImpl : public OfflineRecognizerImpl {
}
if (!EncodeHotwords(is, config_.model_config.modeling_unit, symbol_table_,
bpe_encoder_.get(), &hotwords_)) {
bpe_encoder_.get(), &hotwords_, &boost_scores_)) {
SHERPA_ONNX_LOGE(
"Failed to encode some hotwords, skip them already, see logs above "
"for details.");
}
hotwords_graph_ =
std::make_shared<ContextGraph>(hotwords_, config_.hotwords_score);
hotwords_graph_ = std::make_shared<ContextGraph>(
hotwords_, config_.hotwords_score, boost_scores_);
}
#if __ANDROID_API__ >= 9
... ... @@ -250,13 +270,13 @@ class OfflineRecognizerTransducerImpl : public OfflineRecognizerImpl {
}
if (!EncodeHotwords(is, config_.model_config.modeling_unit, symbol_table_,
bpe_encoder_.get(), &hotwords_)) {
bpe_encoder_.get(), &hotwords_, &boost_scores_)) {
SHERPA_ONNX_LOGE(
"Failed to encode some hotwords, skip them already, see logs above "
"for details.");
}
hotwords_graph_ =
std::make_shared<ContextGraph>(hotwords_, config_.hotwords_score);
hotwords_graph_ = std::make_shared<ContextGraph>(
hotwords_, config_.hotwords_score, boost_scores_);
}
#endif
... ... @@ -264,6 +284,7 @@ class OfflineRecognizerTransducerImpl : public OfflineRecognizerImpl {
OfflineRecognizerConfig config_;
SymbolTable symbol_table_;
std::vector<std::vector<int32_t>> hotwords_;
std::vector<float> boost_scores_;
ContextGraphPtr hotwords_graph_;
std::unique_ptr<ssentencepiece::Ssentencepiece> bpe_encoder_;
std::unique_ptr<OfflineTransducerModel> model_;
... ...
... ... @@ -182,14 +182,35 @@ class OnlineRecognizerTransducerImpl : public OnlineRecognizerImpl {
auto hws = std::regex_replace(hotwords, std::regex("/"), "\n");
std::istringstream is(hws);
std::vector<std::vector<int32_t>> current;
std::vector<float> current_scores;
if (!EncodeHotwords(is, config_.model_config.modeling_unit, sym_,
bpe_encoder_.get(), &current)) {
bpe_encoder_.get(), &current, &current_scores)) {
SHERPA_ONNX_LOGE("Encode hotwords failed, skipping, hotwords are : %s",
hotwords.c_str());
}
int32_t num_default_hws = hotwords_.size();
int32_t num_hws = current.size();
current.insert(current.end(), hotwords_.begin(), hotwords_.end());
auto context_graph =
std::make_shared<ContextGraph>(current, config_.hotwords_score);
if (!current_scores.empty() && !boost_scores_.empty()) {
current_scores.insert(current_scores.end(), boost_scores_.begin(),
boost_scores_.end());
} else if (!current_scores.empty() && boost_scores_.empty()) {
current_scores.insert(current_scores.end(), num_default_hws,
config_.hotwords_score);
} else if (current_scores.empty() && !boost_scores_.empty()) {
current_scores.insert(current_scores.end(), num_hws,
config_.hotwords_score);
current_scores.insert(current_scores.end(), boost_scores_.begin(),
boost_scores_.end());
} else {
// Do nothing.
}
auto context_graph = std::make_shared<ContextGraph>(
current, config_.hotwords_score, current_scores);
auto stream =
std::make_unique<OnlineStream>(config_.feat_config, context_graph);
InitOnlineStream(stream.get());
... ... @@ -376,13 +397,13 @@ class OnlineRecognizerTransducerImpl : public OnlineRecognizerImpl {
}
if (!EncodeHotwords(is, config_.model_config.modeling_unit, sym_,
bpe_encoder_.get(), &hotwords_)) {
bpe_encoder_.get(), &hotwords_, &boost_scores_)) {
SHERPA_ONNX_LOGE(
"Failed to encode some hotwords, skip them already, see logs above "
"for details.");
}
hotwords_graph_ =
std::make_shared<ContextGraph>(hotwords_, config_.hotwords_score);
hotwords_graph_ = std::make_shared<ContextGraph>(
hotwords_, config_.hotwords_score, boost_scores_);
}
#if __ANDROID_API__ >= 9
... ... @@ -400,13 +421,13 @@ class OnlineRecognizerTransducerImpl : public OnlineRecognizerImpl {
}
if (!EncodeHotwords(is, config_.model_config.modeling_unit, sym_,
bpe_encoder_.get(), &hotwords_)) {
bpe_encoder_.get(), &hotwords_, &boost_scores_)) {
SHERPA_ONNX_LOGE(
"Failed to encode some hotwords, skip them already, see logs above "
"for details.");
}
hotwords_graph_ =
std::make_shared<ContextGraph>(hotwords_, config_.hotwords_score);
hotwords_graph_ = std::make_shared<ContextGraph>(
hotwords_, config_.hotwords_score, boost_scores_);
}
#endif
... ... @@ -428,6 +449,7 @@ class OnlineRecognizerTransducerImpl : public OnlineRecognizerImpl {
private:
OnlineRecognizerConfig config_;
std::vector<std::vector<int32_t>> hotwords_;
std::vector<float> boost_scores_;
ContextGraphPtr hotwords_graph_;
std::unique_ptr<ssentencepiece::Ssentencepiece> bpe_encoder_;
std::unique_ptr<OnlineTransducerModel> model_;
... ...
... ... @@ -35,17 +35,21 @@ TEST(TEXT2TOKEN, TEST_cjkchar) {
auto sym_table = SymbolTable(tokens);
std::string text = "世界人民大团结\n中国 V S 美国";
std::string text =
"世界人民大团结\n中国 V S 美国\n\n"; // Test blank lines also
std::istringstream iss(text);
std::vector<std::vector<int32_t>> ids;
std::vector<float> scores;
auto r = EncodeHotwords(iss, "cjkchar", sym_table, nullptr, &ids);
auto r = EncodeHotwords(iss, "cjkchar", sym_table, nullptr, &ids, &scores);
std::vector<std::vector<int32_t>> expected_ids(
{{379, 380, 72, 874, 93, 1251, 489}, {262, 147, 3423, 2476, 21, 147}});
EXPECT_EQ(ids, expected_ids);
EXPECT_EQ(scores.size(), 0);
}
TEST(TEXT2TOKEN, TEST_bpe) {
... ... @@ -68,17 +72,22 @@ TEST(TEXT2TOKEN, TEST_bpe) {
auto sym_table = SymbolTable(tokens);
auto bpe_processor = std::make_unique<ssentencepiece::Ssentencepiece>(bpe);
std::string text = "HELLO WORLD\nI LOVE YOU";
std::string text = "HELLO WORLD\nI LOVE YOU :2.0";
std::istringstream iss(text);
std::vector<std::vector<int32_t>> ids;
std::vector<float> scores;
auto r = EncodeHotwords(iss, "bpe", sym_table, bpe_processor.get(), &ids);
auto r =
EncodeHotwords(iss, "bpe", sym_table, bpe_processor.get(), &ids, &scores);
std::vector<std::vector<int32_t>> expected_ids(
{{22, 58, 24, 425}, {19, 370, 47}});
EXPECT_EQ(ids, expected_ids);
std::vector<float> expected_scores({0, 2.0});
EXPECT_EQ(scores, expected_scores);
}
TEST(TEXT2TOKEN, TEST_cjkchar_bpe) {
... ... @@ -101,19 +110,23 @@ TEST(TEXT2TOKEN, TEST_cjkchar_bpe) {
auto sym_table = SymbolTable(tokens);
auto bpe_processor = std::make_unique<ssentencepiece::Ssentencepiece>(bpe);
std::string text = "世界人民 GOES TOGETHER\n中国 GOES WITH 美国";
std::string text = "世界人民 GOES TOGETHER :1.5\n中国 GOES WITH 美国 :0.5";
std::istringstream iss(text);
std::vector<std::vector<int32_t>> ids;
std::vector<float> scores;
auto r =
EncodeHotwords(iss, "cjkchar+bpe", sym_table, bpe_processor.get(), &ids);
auto r = EncodeHotwords(iss, "cjkchar+bpe", sym_table, bpe_processor.get(),
&ids, &scores);
std::vector<std::vector<int32_t>> expected_ids(
{{1368, 1392, 557, 680, 275, 178, 475},
{685, 736, 275, 178, 179, 921, 736}});
EXPECT_EQ(ids, expected_ids);
std::vector<float> expected_scores({1.5, 0.5});
EXPECT_EQ(scores, expected_scores);
}
TEST(TEXT2TOKEN, TEST_bbpe) {
... ... @@ -136,17 +149,22 @@ TEST(TEXT2TOKEN, TEST_bbpe) {
auto sym_table = SymbolTable(tokens);
auto bpe_processor = std::make_unique<ssentencepiece::Ssentencepiece>(bpe);
std::string text = "频繁\n李鞑靼";
std::string text = "频繁 :1.0\n李鞑靼";
std::istringstream iss(text);
std::vector<std::vector<int32_t>> ids;
std::vector<float> scores;
auto r = EncodeHotwords(iss, "bpe", sym_table, bpe_processor.get(), &ids);
auto r =
EncodeHotwords(iss, "bpe", sym_table, bpe_processor.get(), &ids, &scores);
std::vector<std::vector<int32_t>> expected_ids(
{{259, 1118, 234, 188, 132}, {259, 1585, 236, 161, 148, 236, 160, 191}});
EXPECT_EQ(ids, expected_ids);
std::vector<float> expected_scores({1.0, 0});
EXPECT_EQ(scores, expected_scores);
}
} // namespace sherpa_onnx
... ...
... ... @@ -103,7 +103,8 @@ static bool EncodeBase(const std::vector<std::string> &lines,
bool EncodeHotwords(std::istream &is, const std::string &modeling_unit,
const SymbolTable &symbol_table,
const ssentencepiece::Ssentencepiece *bpe_encoder,
std::vector<std::vector<int32_t>> *hotwords) {
std::vector<std::vector<int32_t>> *hotwords,
std::vector<float> *boost_scores) {
std::vector<std::string> lines;
std::string line;
std::string word;
... ... @@ -131,7 +132,12 @@ bool EncodeHotwords(std::istream &is, const std::string &modeling_unit,
break;
}
}
phrase = oss.str().substr(1);
phrase = oss.str();
if (phrase.empty()) {
continue;
} else {
phrase = phrase.substr(1);
}
std::istringstream piss(phrase);
oss.clear();
oss.str("");
... ... @@ -177,7 +183,8 @@ bool EncodeHotwords(std::istream &is, const std::string &modeling_unit,
}
lines.push_back(oss.str());
}
return EncodeBase(lines, symbol_table, hotwords, nullptr, nullptr, nullptr);
return EncodeBase(lines, symbol_table, hotwords, nullptr, boost_scores,
nullptr);
}
bool EncodeKeywords(std::istream &is, const SymbolTable &symbol_table,
... ...
... ... @@ -29,7 +29,8 @@ namespace sherpa_onnx {
bool EncodeHotwords(std::istream &is, const std::string &modeling_unit,
const SymbolTable &symbol_table,
const ssentencepiece::Ssentencepiece *bpe_encoder_,
std::vector<std::vector<int32_t>> *hotwords_id);
std::vector<std::vector<int32_t>> *hotwords_id,
std::vector<float> *boost_scores);
/* Encode the keywords in an input stream to be tokens ids.
*
... ...