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
Dev gzf (#1474)
zhifu gao0d9384c8c016
diff --git a/README.md b/README.mdindex d15905025..38fd6868d 100644--- a/README.md+++ b/README.md@@ -227,6 +227,21 @@ model = AutoModel(model="paraformer", device="cpu")res = model.export(quantize=False)```+### Text ONNX+```python+# pip3 install -U funasr-onnx+from funasr_onnx import Paraformer+model_dir = "damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch"+model = Paraformer(model_dir, batch_size=1, quantize=True)++wav_path = ['~/.cache/modelscope/hub/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/example/asr_example.wav']++result = model(wav_path)+print(result)+```++More examples ref to [demo](runtime/python/onnxruntime)+## Deployment ServiceFunASR supports deploying pre-trained or further fine-tuned models for service. Currently, it supports the following types of service deployment:- File transcription service, Mandarin, CPU version, donediff --git a/README_zh.md b/README_zh.mdindex 4f4e15dbf..725b21056 100644--- a/README_zh.md+++ b/README_zh.md@@ -216,7 +216,7 @@ print(res)funasr-export ++model=paraformer ++quantize=false```-### 从python指令导出+### 从Python导出```pythonfrom funasr import AutoModel@@ -225,6 +225,20 @@ model = AutoModel(model="paraformer")res = model.export(quantize=False)```+### 测试ONNX+```python+# pip3 install -U funasr-onnx+from funasr_onnx import Paraformer+model_dir = "damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch"+model = Paraformer(model_dir, batch_size=1, quantize=True)++wav_path = ['~/.cache/modelscope/hub/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/example/asr_example.wav']++result = model(wav_path)+print(result)+```++更多例子请参考 [样例](runtime/python/onnxruntime)<a name="服务部署"></a>## 服务部署diff --git a/funasr/download/download_from_hub.py b/funasr/download/download_from_hub.pyindex 4d5a5346a..ef2832f64 100644--- a/funasr/download/download_from_hub.py+++ b/funasr/download/download_from_hub.py@@ -48,9 +48,10 @@ def download_from_ms(**kwargs):if "file_path_metas" in conf_json:add_file_root_path(model_or_path, conf_json["file_path_metas"], cfg)cfg.update(kwargs)- config = OmegaConf.load(cfg["config"])- kwargs = OmegaConf.merge(config, cfg)- kwargs["model"] = config["model"]+ if "config" in cfg:+ config = OmegaConf.load(cfg["config"])+ kwargs = OmegaConf.merge(config, cfg)+ kwargs["model"] = config["model"]elif os.path.exists(os.path.join(model_or_path, "config.yaml")) and os.path.exists(os.path.join(model_or_path, "model.pt")):config = OmegaConf.load(os.path.join(model_or_path, "config.yaml"))kwargs = OmegaConf.merge(config, kwargs)diff --git a/funasr/version.txt b/funasr/version.txtindex 5b09c67c2..a9707166b 100644--- a/funasr/version.txt+++ b/funasr/version.txt@@ -1 +1 @@-1.0.14+1.0.15diff --git a/runtime/python/onnxruntime/funasr_onnx/punc_bin.py b/runtime/python/onnxruntime/funasr_onnx/punc_bin.pyindex b1aca6ea2..d9bac14d0 100644--- a/runtime/python/onnxruntime/funasr_onnx/punc_bin.py+++ b/runtime/python/onnxruntime/funasr_onnx/punc_bin.py@@ -56,9 +56,9 @@ class CT_Transformer():"\npip3 install -U funasr -i https://mirror.sjtu.edu.cn/pypi/web/simple"model = AutoModel(model=model_dir)- model_dir = model.export(type="onnx", quantize=quantize)+ model_dir = model.export(quantize=quantize)- config_file = os.path.join(model_dir, 'punc.yaml')+ config_file = os.path.join(model_dir, 'confi.yaml')config = read_yaml(config_file)token_list = os.path.join(model_dir, 'tokens.json')with open(token_list, 'r', encoding='utf-8') as f:@@ -67,7 +67,7 @@ class CT_Transformer():self.converter = TokenIDConverter(token_list)self.ort_infer = OrtInferSession(model_file, device_id, intra_op_num_threads=intra_op_num_threads)self.batch_size = 1- self.punc_list = config['punc_list']+ self.punc_list = config["model_conf"]['punc_list']self.period = 0for i in range(len(self.punc_list)):if self.punc_list[i] == ",":@@ -76,9 +76,9 @@ class CT_Transformer():self.punc_list[i] = "?"elif self.punc_list[i] == "。":self.period = i- if "seg_jieba" in config:+ self.jieba_usr_dict_path = os.path.join(model_dir, 'jieba_usr_dict')+ if os.path.exists(self.jieba_usr_dict_path):self.seg_jieba = True- self.jieba_usr_dict_path = os.path.join(model_dir, 'jieba_usr_dict')self.code_mix_split_words_jieba = code_mix_split_words_jieba(self.jieba_usr_dict_path)else:self.seg_jieba = False@@ -175,7 +175,7 @@ class CT_Transformer_VadRealtime(CT_Transformer):intra_op_num_threads: int = 4,cache_dir: str = None):- super(CT_Transformer_VadRealtime, self).__init__(model_dir, batch_size, device_id, quantize, intra_op_num_threads, cache_dir=cache_dir)+ super().__init__(model_dir, batch_size, device_id, quantize, intra_op_num_threads, cache_dir=cache_dir)def __call__(self, text: str, param_dict: map, split_size=20):cache_key = "cache"diff --git a/runtime/python/onnxruntime/setup.py b/runtime/python/onnxruntime/setup.pyindex 3c128eefe..df20fd940 100644--- a/runtime/python/onnxruntime/setup.py+++ b/runtime/python/onnxruntime/setup.py@@ -13,7 +13,7 @@ def get_readme():MODULE_NAME = 'funasr_onnx'-VERSION_NUM = '0.2.5'+VERSION_NUM = '0.3.0'setuptools.setup(name=MODULE_NAME,