Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.
Python asraudiochineseemotion-recognitionfunasrmcp-servermultilingual-asropenai-compatible-apiparaformerpunctuationpytorchreal-time-asrspeaker-diarizationspeech-recognitionspeech-to-textstreaming-asrtranscriptionvllmvoice-activity-detectionwhisper-alternative
add
游雁22025a68d9e3
diff --git a/examples/industrial_data_pretraining/sense_voice/demo.py b/examples/industrial_data_pretraining/sense_voice/demo.pyindex ed583f0f8..b3af796ba 100644--- a/examples/industrial_data_pretraining/sense_voice/demo.py+++ b/examples/industrial_data_pretraining/sense_voice/demo.py@@ -21,6 +21,7 @@ DecodingOptions = {"language": "auto","fp16": True,"gain_event": True,+ "beam_size": 5,}res = model.generate(input=input_wav, batch_size_s=0, DecodingOptions=DecodingOptions)diff --git a/examples/industrial_data_pretraining/sense_voice/demo_fsmn.py b/examples/industrial_data_pretraining/sense_voice/demo_fsmn.pyindex e063e1fde..ce4bdf8a2 100644--- a/examples/industrial_data_pretraining/sense_voice/demo_fsmn.py+++ b/examples/industrial_data_pretraining/sense_voice/demo_fsmn.py@@ -21,6 +21,7 @@ DecodingOptions = {"language": "auto","fp16": True,"gain_event": True,+ "beam_size": 5,}res = model.generate(input=input_wav, batch_size_s=0, DecodingOptions=DecodingOptions, beam_size=5)