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

Offline decode support multi threads (#306)

Browse Source 
Co-authored-by: cuidongcai1035 <cuidongcai1035@wezhuiyi.com>
This commit is contained in:
keanu
2023-09-19 21:04:13 +08:00
committed by GitHub
parent b3e9986825
commit bd173b27cc
3 changed files with 319 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
sherpa-onnx/csrc/offline-whisper-model.cc
View File

@@ -78,9 +78,11 @@ class OfflineWhisperModel::Impl {
decoder_input.size(), decoder_output_names_ptr_.data(),
decoder_output_names_ptr_.size());
return {std::move(decoder_out[0]), std::move(decoder_out[1]),
std::move(decoder_out[2]), std::move(decoder_input[3]),
std::move(decoder_input[4]), std::move(decoder_input[5])};
return std::tuple<Ort::Value, Ort::Value, Ort::Value, Ort::Value,
Ort::Value, Ort::Value>{
std::move(decoder_out[0]), std::move(decoder_out[1]),
std::move(decoder_out[2]), std::move(decoder_input[3]),
std::move(decoder_input[4]), std::move(decoder_input[5])};
}
std::pair<Ort::Value, Ort::Value> GetInitialSelfKVCache() {