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
atsr
语帆21b49bd56f44
diff --git a/funasr/auto/auto_model.py b/funasr/auto/auto_model.pyindex ef261dfcc..56dd5b5e4 100644--- a/funasr/auto/auto_model.py+++ b/funasr/auto/auto_model.py@@ -213,7 +213,7 @@ class AutoModel:batch_size = kwargs.get("batch_size", 1)# if kwargs.get("device", "cpu") == "cpu":# batch_size = 1- pdb.set_trace()+key_list, data_list = prepare_data_iterator(input, input_len=input_len, data_type=kwargs.get("data_type", None), key=key)speed_stats = {}diff --git a/funasr/bin/inference.py b/funasr/bin/inference.pyindex 3943d7ed4..d2f0c149d 100644--- a/funasr/bin/inference.py+++ b/funasr/bin/inference.py@@ -1,7 +1,7 @@import hydraimport loggingfrom omegaconf import DictConfig, OmegaConf, ListConfig-import pdb+from funasr.auto.auto_model import AutoModel@@ -23,7 +23,6 @@ def main_hydra(cfg: DictConfig):if kwargs.get("debug", False):import pdb; pdb.set_trace()model = AutoModel(**kwargs)- pdb.set_trace()res = model.generate(input=kwargs["input"])print(res)