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
funasr1.0 update (#1278)
zhifu gao1159adbca076
diff --git a/README.md b/README.mdindex c9b9e8916..7500dd472 100644--- a/README.md+++ b/README.md@@ -93,7 +93,7 @@ from funasr import AutoModel# use vad, punc, spk or not as you needmodel = AutoModel(model="paraformer-zh", model_revision="v2.0.2",vad_model="fsmn-vad", vad_model_revision="v2.0.2",- punc_model="ct-punc-c", punc_model_revision="v2.0.2",+ punc_model="ct-punc-c", punc_model_revision="v2.0.3",# spk_model="cam++", spk_model_revision="v2.0.2",)res = model.generate(input=f"{model.model_path}/example/asr_example.wav",diff --git a/README_zh.md b/README_zh.mdindex 9cd18977d..b19e7c29a 100644--- a/README_zh.md+++ b/README_zh.md@@ -89,7 +89,7 @@ from funasr import AutoModel# use vad, punc, spk or not as you needmodel = AutoModel(model="paraformer-zh", model_revision="v2.0.2",vad_model="fsmn-vad", vad_model_revision="v2.0.2",- punc_model="ct-punc-c", punc_model_revision="v2.0.2",+ punc_model="ct-punc-c", punc_model_revision="v2.0.3",# spk_model="cam++", spk_model_revision="v2.0.2",)res = model.generate(input=f"{model.model_path}/example/asr_example.wav",diff --git a/examples/industrial_data_pretraining/bicif_paraformer/demo.py b/examples/industrial_data_pretraining/bicif_paraformer/demo.pyindex a06b308d1..f1b1496e5 100644--- a/examples/industrial_data_pretraining/bicif_paraformer/demo.py+++ b/examples/industrial_data_pretraining/bicif_paraformer/demo.py@@ -10,7 +10,7 @@ model = AutoModel(model="damo/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16kvad_model="damo/speech_fsmn_vad_zh-cn-16k-common-pytorch",vad_model_revision="v2.0.2",punc_model="damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch",- punc_model_revision="v2.0.2",+ punc_model_revision="v2.0.3",spk_model="damo/speech_campplus_sv_zh-cn_16k-common",spk_model_revision="v2.0.2",)diff --git a/examples/industrial_data_pretraining/bicif_paraformer/infer.sh b/examples/industrial_data_pretraining/bicif_paraformer/infer.shindex 09e1c8393..55efdf2ce 100644--- a/examples/industrial_data_pretraining/bicif_paraformer/infer.sh+++ b/examples/industrial_data_pretraining/bicif_paraformer/infer.sh@@ -4,7 +4,7 @@ model_revision="v2.0.2"vad_model="damo/speech_fsmn_vad_zh-cn-16k-common-pytorch"vad_model_revision="v2.0.2"punc_model="damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch"-punc_model_revision="v2.0.2"+punc_model_revision="v2.0.3"spk_model="damo/speech_campplus_sv_zh-cn_16k-common"spk_model_revision="v2.0.2"diff --git a/examples/industrial_data_pretraining/paraformer-zh-spk/demo.py b/examples/industrial_data_pretraining/paraformer-zh-spk/demo.pyindex b4453e927..e17a83168 100644--- a/examples/industrial_data_pretraining/paraformer-zh-spk/demo.py+++ b/examples/industrial_data_pretraining/paraformer-zh-spk/demo.py@@ -10,7 +10,7 @@ model = AutoModel(model="damo/speech_seaco_paraformer_large_asr_nat-zh-cn-16k-covad_model="damo/speech_fsmn_vad_zh-cn-16k-common-pytorch",vad_model_revision="v2.0.2",punc_model="damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch",- punc_model_revision="v2.0.2",+ punc_model_revision="v2.0.3",spk_model="damo/speech_campplus_sv_zh-cn_16k-common",spk_model_revision="v2.0.2")diff --git a/examples/industrial_data_pretraining/paraformer-zh-spk/infer.sh b/examples/industrial_data_pretraining/paraformer-zh-spk/infer.shindex 98a325dd4..b8610cb70 100644--- a/examples/industrial_data_pretraining/paraformer-zh-spk/infer.sh+++ b/examples/industrial_data_pretraining/paraformer-zh-spk/infer.sh@@ -4,7 +4,7 @@ model_revision="v2.0.2"vad_model="damo/speech_fsmn_vad_zh-cn-16k-common-pytorch"vad_model_revision="v2.0.2"punc_model="damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch"-punc_model_revision="v2.0.2"+punc_model_revision="v2.0.3"spk_model="damo/speech_campplus_sv_zh-cn_16k-common"spk_model_revision="v2.0.2"diff --git a/examples/industrial_data_pretraining/paraformer/demo.py b/examples/industrial_data_pretraining/paraformer/demo.pyindex 78af3aa1d..724191d9e 100644--- a/examples/industrial_data_pretraining/paraformer/demo.py+++ b/examples/industrial_data_pretraining/paraformer/demo.py@@ -5,7 +5,12 @@from funasr import AutoModel-model = AutoModel(model="damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch", model_revision="v2.0.2")+model = AutoModel(model="damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch", model_revision="v2.0.3",+ # vad_model="damo/speech_fsmn_vad_zh-cn-16k-common-pytorch",+ # vad_model_revision="v2.0.2",+ # punc_model="damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch",+ # punc_model_revision="v2.0.3",+ )res = model.generate(input="https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_zh.wav")print(res)diff --git a/examples/industrial_data_pretraining/seaco_paraformer/demo.py b/examples/industrial_data_pretraining/seaco_paraformer/demo.pyindex 19ad1c9c5..a2029564d 100644--- a/examples/industrial_data_pretraining/seaco_paraformer/demo.py+++ b/examples/industrial_data_pretraining/seaco_paraformer/demo.py@@ -10,7 +10,7 @@ model = AutoModel(model="damo/speech_seaco_paraformer_large_asr_nat-zh-cn-16k-covad_model="damo/speech_fsmn_vad_zh-cn-16k-common-pytorch",vad_model_revision="v2.0.2",punc_model="damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch",- punc_model_revision="v2.0.2",+ punc_model_revision="v2.0.3",spk_model="damo/speech_campplus_sv_zh-cn_16k-common",spk_model_revision="v2.0.2",)diff --git a/examples/industrial_data_pretraining/seaco_paraformer/infer.sh b/examples/industrial_data_pretraining/seaco_paraformer/infer.shindex 61029e114..f33568437 100644--- a/examples/industrial_data_pretraining/seaco_paraformer/infer.sh+++ b/examples/industrial_data_pretraining/seaco_paraformer/infer.sh@@ -4,7 +4,7 @@ model_revision="v2.0.2"vad_model="damo/speech_fsmn_vad_zh-cn-16k-common-pytorch"vad_model_revision="v2.0.2"punc_model="damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch"-punc_model_revision="v2.0.2"+punc_model_revision="v2.0.3"python funasr/bin/inference.py \+model=${model} \diff --git a/funasr/auto/auto_model.py b/funasr/auto/auto_model.pyindex 0538f6623..ca6189dd2 100644--- a/funasr/auto/auto_model.py+++ b/funasr/auto/auto_model.py@@ -391,7 +391,7 @@ class AutoModel:if self.punc_model is not None:self.punc_kwargs.update(cfg)punc_res = self.inference(result["text"], model=self.punc_model, kwargs=self.punc_kwargs, **cfg)- result["text_with_punc"] = punc_res[0]["text"]+ result["text"] = punc_res[0]["text"]# speaker embedding cluster after resortedif self.spk_model is not None:diff --git a/funasr/models/paraformer/model.py b/funasr/models/paraformer/model.pyindex 468d23f39..0c4f14aae 100644--- a/funasr/models/paraformer/model.py+++ b/funasr/models/paraformer/model.py@@ -451,7 +451,7 @@ class Paraformer(torch.nn.Module):self.nbest = kwargs.get("nbest", 1)meta_data = {}- if isinstance(data_in, torch.Tensor): # fbank+ if isinstance(data_in, torch.Tensor) and kwargs.get("data_type", "sound") == "fbank": # fbankspeech, speech_lengths = data_in, data_lengthsif len(speech.shape) < 3:speech = speech[None, :, :]