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
python runtime
游雁f3d3362fd33a
diff --git a/runtime/python/onnxruntime/funasr_onnx/sensevoice_bin.py b/runtime/python/onnxruntime/funasr_onnx/sensevoice_bin.pyindex 7d6f341e5..433acd192 100644--- a/runtime/python/onnxruntime/funasr_onnx/sensevoice_bin.py+++ b/runtime/python/onnxruntime/funasr_onnx/sensevoice_bin.py@@ -107,8 +107,8 @@ class SenseVoiceSmall:ctc_logits, encoder_out_lens = self.infer(feats,feats_len,- np.array(language, dtype=np.int32),- np.array(textnorm, dtype=np.int32),+ np.array([language], dtype=np.int32),+ np.array([textnorm], dtype=np.int32),)# back to torch.Tensorctc_logits = torch.from_numpy(ctc_logits).float()@@ -120,7 +120,7 @@ class SenseVoiceSmall:mask = yseq != self.blank_idtoken_int = yseq[mask].tolist()- asr_res.append(self.tokenizer.encode(token_int))+ asr_res.append(self.tokenizer.decode(token_int))return asr_res