speech_recognition/FunASR
1
0
Fork 0
mirror of https://github.com/modelscope/FunASR synced 2025-09-15 14:48:36 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

Merge branch 'cpp-python-websocket-compatible' of github.com:zhaomingwork/FunASR into cpp-python-websocket-compatible

Browse Source
This commit is contained in:
zhaomingwork 2023-05-12 07:00:06 +00:00
commit ec1f0f087f
7 changed files with 95 additions and 45 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
funasr/export/export_model.py
View File

@ -27,15 +27,13 @@ class ModelExport:
):
assert check_argument_types()
self.set_all_random_seed(0)
if cache_dir is None:
cache_dir = Path.home() / ".cache" / "export"
self.cache_dir = Path(cache_dir)
self.cache_dir = cache_dir
self.export_config = dict(
feats_dim=560,
onnx=False,
)
print("output dir: {}".format(self.cache_dir))
self.onnx = onnx
self.device = device
self.quant = quant
@ -52,7 +50,7 @@ class ModelExport:
verbose: bool = False,
):
export_dir = self.cache_dir / tag_name.replace(' ', '-')
export_dir = self.cache_dir
os.makedirs(export_dir, exist_ok=True)
# export encoder1
@ -174,6 +172,7 @@ class ModelExport:
if model_dir.startswith('damo'):
from modelscope.hub.snapshot_download import snapshot_download
model_dir = snapshot_download(model_dir, cache_dir=self.cache_dir)
self.cache_dir = model_dir
if mode is None:
import json

26
funasr/runtime/python/grpc/Readme.md
View File

@ -5,7 +5,6 @@ The audio data is in streaming, the asr inference process is in offline.
## For the Server
### Prepare server environment
#### Backend is modelscope pipeline (default)
Install the modelscope and funasr
```shell
@ -22,18 +21,6 @@ cd funasr/runtime/python/grpc
pip install -r requirements_server.txt
```
#### Backend is funasr_onnx (optional)
Install [`funasr_onnx`](https://github.com/alibaba-damo-academy/FunASR/tree/main/funasr/runtime/python/onnxruntime).
```
pip install funasr_onnx -i https://pypi.Python.org/simple
```
Export the model, more details ref to [export docs](https://github.com/alibaba-damo-academy/FunASR/tree/main/funasr/runtime/python/onnxruntime).
```shell
python -m funasr.export.export_model --model-name damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch --export-dir ./export --type onnx --quantize True
```
### Generate protobuf file
Run on server, the two generated pb files are both used for server and client
@ -51,11 +38,6 @@ python -m grpc_tools.protoc --proto_path=./proto -I ./proto --python_out=. -
python grpc_main_server.py --port 10095 --backend pipeline
```
If you want run server with onnxruntime, please set `backend` and `onnx_dir`.
```
# Start server.
python grpc_main_server.py --port 10095 --backend onnxruntime --onnx_dir /models/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch
```
## For the client
@ -87,9 +69,5 @@ python grpc_main_client_mic.py --host 127.0.0.1 --port 10095
<div align="left"><img src="proto/workflow.png" width="400"/>
## Reference
We borrow from or refer to some code as:
1)https://github.com/wenet-e2e/wenet/tree/main/runtime/core/grpc
2)https://github.com/Open-Speech-EkStep/inference_service/blob/main/realtime_inference_service.py
## Acknowledge
1. This project is maintained by [FunASR community](https://github.com/alibaba-damo-academy/FunASR).

20
funasr/runtime/python/onnxruntime/funasr_onnx/paraformer_bin.py
View File

@ -32,14 +32,30 @@ class Paraformer():
plot_timestamp_to: str = "",
quantize: bool = False,
intra_op_num_threads: int = 4,
cache_dir: str = None
):
if not Path(model_dir).exists():
raise FileNotFoundError(f'{model_dir} does not exist.')
from modelscope.hub.snapshot_download import snapshot_download
try:
model_dir = snapshot_download(model_dir, cache_dir=cache_dir)
except:
raise "model_dir must be model_name in modelscope or local path downloaded from modelscope, but is {}".format(model_dir)
model_file = os.path.join(model_dir, 'model.onnx')
if quantize:
model_file = os.path.join(model_dir, 'model_quant.onnx')
if not os.path.exists(model_file):
print(".onnx is not exist, begin to export onnx")
from funasr.export.export_model import ModelExport
export_model = ModelExport(
cache_dir=cache_dir,
onnx=True,
device="cpu",
quant=quantize,
)
export_model.export(model_dir)
config_file = os.path.join(model_dir, 'config.yaml')
cmvn_file = os.path.join(model_dir, 'am.mvn')
config = read_yaml(config_file)

30
funasr/runtime/python/onnxruntime/funasr_onnx/punc_bin.py
View File

@ -24,15 +24,32 @@ class CT_Transformer():
batch_size: int = 1,
device_id: Union[str, int] = "-1",
quantize: bool = False,
intra_op_num_threads: int = 4
intra_op_num_threads: int = 4,
cache_dir: str = None,
):
if not Path(model_dir).exists():
raise FileNotFoundError(f'{model_dir} does not exist.')
from modelscope.hub.snapshot_download import snapshot_download
try:
model_dir = snapshot_download(model_dir, cache_dir=cache_dir)
except:
raise "model_dir must be model_name in modelscope or local path downloaded from modelscope, but is {}".format(
model_dir)
model_file = os.path.join(model_dir, 'model.onnx')
if quantize:
model_file = os.path.join(model_dir, 'model_quant.onnx')
if not os.path.exists(model_file):
print(".onnx is not exist, begin to export onnx")
from funasr.export.export_model import ModelExport
export_model = ModelExport(
cache_dir=cache_dir,
onnx=True,
device="cpu",
quant=quantize,
)
export_model.export(model_dir)
config_file = os.path.join(model_dir, 'punc.yaml')
config = read_yaml(config_file)
@ -135,9 +152,10 @@ class CT_Transformer_VadRealtime(CT_Transformer):
batch_size: int = 1,
device_id: Union[str, int] = "-1",
quantize: bool = False,
intra_op_num_threads: int = 4
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)
super(CT_Transformer_VadRealtime, self).__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"

1
funasr/runtime/python/onnxruntime/funasr_onnx/utils/utils.py
View File

@ -271,4 +271,5 @@ def get_logger(name='funasr_onnx'):
logger.addHandler(sh)
logger_initialized[name] = True
logger.propagate = False
logging.basicConfig(level=logging.ERROR)
return logger

37
funasr/runtime/python/onnxruntime/funasr_onnx/vad_bin.py
View File

@ -31,14 +31,30 @@ class Fsmn_vad():
quantize: bool = False,
intra_op_num_threads: int = 4,
max_end_sil: int = None,
cache_dir: str = None
):
if not Path(model_dir).exists():
raise FileNotFoundError(f'{model_dir} does not exist.')
from modelscope.hub.snapshot_download import snapshot_download
try:
model_dir = snapshot_download(model_dir, cache_dir=cache_dir)
except:
raise "model_dir must be model_name in modelscope or local path downloaded from modelscope, but is {}".format(
model_dir)
model_file = os.path.join(model_dir, 'model.onnx')
if quantize:
model_file = os.path.join(model_dir, 'model_quant.onnx')
if not os.path.exists(model_file):
print(".onnx is not exist, begin to export onnx")
from funasr.export.export_model import ModelExport
export_model = ModelExport(
cache_dir=cache_dir,
onnx=True,
device="cpu",
quant=quantize,
)
export_model.export(model_dir)
config_file = os.path.join(model_dir, 'vad.yaml')
cmvn_file = os.path.join(model_dir, 'vad.mvn')
config = read_yaml(config_file)
@ -172,14 +188,29 @@ class Fsmn_vad_online():
quantize: bool = False,
intra_op_num_threads: int = 4,
max_end_sil: int = None,
cache_dir: str = None
):
if not Path(model_dir).exists():
raise FileNotFoundError(f'{model_dir} does not exist.')
from modelscope.hub.snapshot_download import snapshot_download
try:
model_dir = snapshot_download(model_dir, cache_dir=cache_dir)
except:
raise "model_dir must be model_name in modelscope or local path downloaded from modelscope, but is {}".format(
model_dir)
model_file = os.path.join(model_dir, 'model.onnx')
if quantize:
model_file = os.path.join(model_dir, 'model_quant.onnx')
if not os.path.exists(model_file):
print(".onnx is not exist, begin to export onnx")
from funasr.export.export_model import ModelExport
export_model = ModelExport(
cache_dir=cache_dir,
onnx=True,
device="cpu",
quant=quantize,
)
export_model.export(model_dir)
config_file = os.path.join(model_dir, 'vad.yaml')
cmvn_file = os.path.join(model_dir, 'vad.mvn')
config = read_yaml(config_file)

17
funasr/runtime/python/onnxruntime/setup.py
View File

@ -13,7 +13,7 @@ def get_readme():
MODULE_NAME = 'funasr_onnx'
VERSION_NUM = '0.0.8'
VERSION_NUM = '0.0.10'
setuptools.setup(
name=MODULE_NAME,
@ -27,10 +27,17 @@ setuptools.setup(
long_description=get_readme(),
long_description_content_type='text/markdown',
include_package_data=True,
install_requires=["librosa", "onnxruntime>=1.7.0",
"scipy", "numpy>=1.19.3",
"typeguard", "kaldi-native-fbank",
"PyYAML>=5.1.2"],
install_requires=["librosa",
"onnxruntime>=1.7.0",
"scipy",
"numpy>=1.19.3",
"typeguard",
"kaldi-native-fbank",
"PyYAML>=5.1.2",
"funasr",
"modelscope",
"onnx"
],
packages=[MODULE_NAME, f'{MODULE_NAME}.utils'],
keywords=[
'funasr,asr'