EngineX-Iluvatar/enginex-mr_series-sherpa-onnx
Archived
16
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Encode hotwords in C++ side (#828)

Browse Source 
* Encode hotwords in C++ side
This commit is contained in:
Wei Kang
2024-05-20 19:41:36 +08:00
committed by GitHub
parent 8af2af8466
commit b012b78ceb
43 changed files with 714 additions and 102 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
sherpa-onnx/csrc/online-model-config.cc
View File

@@ -32,6 +32,19 @@ void OnlineModelConfig::Register(ParseOptions *po) {
po->Register("provider", &provider,
"Specify a provider to use: cpu, cuda, coreml");
po->Register("modeling-unit", &modeling_unit,
"The modeling unit of the model, commonly used units are bpe, "
"cjkchar, cjkchar+bpe, etc. Currently, it is needed only when "
"hotwords are provided, we need it to encode the hotwords into "
"token sequence.");
po->Register("bpe-vocab", &bpe_vocab,
"The vocabulary generated by google's sentencepiece program. "
"It is a file has two columns, one is the token, the other is "
"the log probability, you can get it from the directory where "
"your bpe model is generated. Only used when hotwords provided "
"and the modeling unit is bpe or cjkchar+bpe");
po->Register("model-type", &model_type,
"Specify it to reduce model initialization time. "
"Valid values are: conformer, lstm, zipformer, zipformer2, "
@@ -50,6 +63,14 @@ bool OnlineModelConfig::Validate() const {
return false;
}
if (!modeling_unit.empty() &&
(modeling_unit == "bpe" || modeling_unit == "cjkchar+bpe")) {
if (!FileExists(bpe_vocab)) {
SHERPA_ONNX_LOGE("bpe_vocab: %s does not exist", bpe_vocab.c_str());
return false;
}
}
if (!paraformer.encoder.empty()) {
return paraformer.Validate();
}
@@ -83,7 +104,9 @@ std::string OnlineModelConfig::ToString() const {
os << "warm_up=" << warm_up << ", ";
os << "debug=" << (debug ? "True" : "False") << ", ";
os << "provider=\"" << provider << "\", ";
os << "model_type=\"" << model_type << "\")";
os << "model_type=\"" << model_type << "\", ";
os << "modeling_unit=\"" << modeling_unit << "\", ";
os << "bpe_vocab=\"" << bpe_vocab << "\")";
return os.str();
}