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
update
游雁3e99d0e8269c
diff --git a/examples/industrial_data_pretraining/fun_asr_nano/model.py b/examples/industrial_data_pretraining/fun_asr_nano/model.pyindex 308740083..c2830e815 100644--- a/examples/industrial_data_pretraining/fun_asr_nano/model.py+++ b/examples/industrial_data_pretraining/fun_asr_nano/model.py@@ -15,7 +15,7 @@ from funasr.register import tablesfrom funasr.train_utils.device_funcs import force_gatherable, to_devicefrom funasr.utils.datadir_writer import DatadirWriterfrom funasr.utils.load_utils import extract_fbank, load_audio_text_image_video-from transformers import AutoConfig, AutoModelForCausalLM+from ctc import CTCfrom tools.utils import forced_align@@ -72,6 +72,7 @@ class FunASRNano(nn.Module):self.llm = Noneinit_param_path = llm_conf.get("init_param_path", None)llm_dim = None+ from transformers import AutoConfig, AutoModelForCausalLMllm_load_kwargs = llm_conf.get("load_kwargs", {})config = AutoConfig.from_pretrained(init_param_path)