Skip to content

Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.

update cutsplit step

雾聪6dcd960fda8b
diff --git a/runtime/onnxruntime/src/audio.cpp b/runtime/onnxruntime/src/audio.cpp
index 9b93dc83b..0135ab4f1 100644
--- a/runtime/onnxruntime/src/audio.cpp
+++ b/runtime/onnxruntime/src/audio.cpp
@@ -1096,7 +1096,7 @@ void Audio::CutSplit(OfflineStream* offline_stream)
delete frame;
frame = nullptr;
- int step = dest_sample_rate*10;
+ int step = dest_sample_rate*1;
bool is_final=false;
vector<std::vector<int>> vad_segments;
for (int sample_offset = 0; sample_offset < speech_len; sample_offset += std::min(step, speech_len - sample_offset)) {