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
[fix] fix empty asr result (#1794)
Xingchen Song(宋星辰)93a02feda184
diff --git a/funasr/auto/auto_model.py b/funasr/auto/auto_model.pyindex 7b5a02fff..9b4f6d92c 100644--- a/funasr/auto/auto_model.py+++ b/funasr/auto/auto_model.py@@ -388,6 +388,7 @@ class AutoModel:):max_len_in_batch = max(max_len_in_batch, sample_length)end_idx += 1+ results_sorted.append({'key': 'bad_data', 'text': '', 'timestamp': []})continuespeech_j, speech_lengths_j = slice_padding_audio_samples(@@ -417,7 +418,7 @@ class AutoModel:end_idx += 1max_len_in_batch = sample_lengthif len(results) < 1:- continue+ results.append({'key': 'bad_data', 'text': '', 'timestamp': []})results_sorted.extend(results)# end_asr_total = time.time()