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
test
语帆044199f80279
diff --git a/funasr/models/lcbnet/model.py b/funasr/models/lcbnet/model.pyindex 9646e1e0d..563ff26e9 100644--- a/funasr/models/lcbnet/model.py+++ b/funasr/models/lcbnet/model.py@@ -21,7 +21,7 @@ from funasr.utils.load_utils import load_audio_text_image_video, extract_fbankfrom funasr.utils import postprocess_utilsfrom funasr.utils.datadir_writer import DatadirWriterfrom funasr.register import tables-+import pdb@tables.register("model_classes", "LCBNet")class LCBNet(nn.Module):"""@@ -90,7 +90,7 @@ class LCBNet(nn.Module):fusion_encoder_class = tables.encoder_classes.get(fusion_encoder)fusion_encoder = fusion_encoder_class(**fusion_encoder_conf)bias_predictor_class = tables.encoder_classes.get_class(bias_predictor)- bias_predictor = bias_predictor_class(args.bias_predictor_conf)+ bias_predictor = bias_predictor_class(bias_predictor_conf)if decoder is not None:decoder_class = tables.decoder_classes.get(decoder)@@ -117,9 +117,13 @@ class LCBNet(nn.Module):self.specaug = specaugself.normalize = normalizeself.encoder = encoder+ # lcbnetself.text_encoder = text_encoderself.fusion_encoder = fusion_encoderself.bias_predictor = bias_predictor+ self.select_num = select_num+ self.select_length = select_length+ self.insert_blank = insert_blankif not hasattr(self.encoder, "interctc_use_conditioning"):self.encoder.interctc_use_conditioning = False@@ -409,7 +413,8 @@ class LCBNet(nn.Module):logging.info("enable beam_search")self.init_beam_search(**kwargs)self.nbest = kwargs.get("nbest", 1)-+ pdb.set_trace()+meta_data = {}if isinstance(data_in, torch.Tensor) and kwargs.get("data_type", "sound") == "fbank": # fbankspeech, speech_lengths = data_in, data_lengths