modify yaml, test=doc

pull/1539/head
lym0302 3 years ago
parent 10ab7aabfe
commit 3b304544f6

@ -11,21 +11,14 @@ This demo is an implementation of starting the voice service and accessing the s
see [installation](https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/docs/source/install.md).
It is recommended to use **paddlepaddle 2.2.1** or above.
You can choose one way from easy, meduim and hard to install paddlespeech.
You can choose one way from meduim and hard to install paddlespeech.
### 2. Prepare config File
The configuration file contains the service-related configuration files and the model configuration related to the voice tasks contained in the service. They are all under the `conf` folder.
The configuration file can be found in `conf/application.yaml` .
Among them, `engine_list` indicates the speech engine that will be included in the service to be started, in the format of <speech task>_<engine type>.
At present, the speech tasks integrated by the service include: asr (speech recognition) and tts (speech synthesis).
Currently the engine type supports two forms: python and inference (Paddle Inference)
**Note: The configuration of `engine_backend` in `application.yaml` represents all speech tasks included in the started service.**
If the service you want to start contains only a certain speech task, then you need to comment out the speech tasks that do not need to be included. For example, if you only want to use the speech recognition (ASR) service, then you can comment out the speech synthesis (TTS) service, as in the following example:
```bash
engine_backend:
asr: 'conf/asr/asr.yaml'
#tts: 'conf/tts/tts.yaml'
```
**Note: The configuration file of `engine_backend` in `application.yaml` needs to match the configuration type of `engine_type`.**
When the configuration file of `engine_backend` is `XXX.yaml`, the configuration type of `engine_type` needs to be set to `python`; when the configuration file of `engine_backend` is `XXX_pd.yaml`, the configuration of `engine_type` needs to be set type is `inference`;
The input of ASR client demo should be a WAV file(`.wav`), and the sample rate must be the same as the model.

@ -11,20 +11,15 @@
请看 [安装文档](https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/docs/source/install.md).
推荐使用 **paddlepaddle 2.2.1** 或以上版本。
你可以从 easymediumhard 三中方式中选择一种方式安装 PaddleSpeech。
你可以从 mediumhard 三中方式中选择一种方式安装 PaddleSpeech。
### 2. 准备配置文件
配置文件包含服务相关的配置文件和服务中包含的语音任务相关的模型配置。 它们都在 `conf` 文件夹下。
**注意:`application.yaml` 中 `engine_backend` 的配置表示启动的服务中包含的所有语音任务。**
如果你想启动的服务中只包含某项语音任务那么你需要注释掉不需要包含的语音任务。例如你只想使用语音识别ASR服务那么你可以将语音合成TTS服务注释掉如下示例
```bash
engine_backend:
asr: 'conf/asr/asr.yaml'
#tts: 'conf/tts/tts.yaml'
```
**注意:`application.yaml` 中 `engine_backend` 的配置文件需要和 `engine_type` 的配置类型匹配。**
当`engine_backend` 的配置文件为`XXX.yaml`时,需要设置`engine_type`的配置类型为`python`;当`engine_backend` 的配置文件为`XXX_pd.yaml`时,需要设置`engine_type`的配置类型为`inference`;
配置文件可参见 `conf/application.yaml`
其中,`engine_list`表示即将启动的服务将会包含的语音引擎,格式为 <语音任务>_<引擎类型>。
目前服务集成的语音任务有: asr(语音识别)、tts(语音合成)。
目前引擎类型支持两种形式python 及 inference (Paddle Inference)
这个 ASR client 的输入应该是一个 WAV 文件(`.wav`),并且采样率必须与模型的采样率相同。

@ -1,27 +1,107 @@
# This is the parameter configuration file for PaddleSpeech Serving.
##################################################################
#################################################################################
# SERVER SETTING #
##################################################################
#################################################################################
host: 127.0.0.1
port: 8090
##################################################################
# CONFIG FILE #
##################################################################
# add engine backend type (Options: asr, tts) and config file here.
# Adding a speech task to engine_backend means starting the service.
engine_backend:
asr: 'conf/asr/asr.yaml'
tts: 'conf/tts/tts.yaml'
# The engine_type of speech task needs to keep the same type as the config file of speech task.
# E.g: The engine_type of asr is 'python', the engine_backend of asr is 'XX/asr.yaml'
# E.g: The engine_type of asr is 'inference', the engine_backend of asr is 'XX/asr_pd.yaml'
#
# add engine type (Options: python, inference)
engine_type:
asr: 'python'
tts: 'python'
# The task format in the engin_list is: <speech task>_<engine type>
# task choices = ['asr_python', 'asr_inference', 'tts_python', 'tts_inference']
engine_list: ['asr_python', 'tts_python']
#################################################################################
# ENGINE CONFIG #
#################################################################################
################### speech task: asr; engine_type: python #######################
asr_python:
model: 'conformer_wenetspeech'
lang: 'zh'
sample_rate: 16000
cfg_path: # [optional]
ckpt_path: # [optional]
decode_method: 'attention_rescoring'
force_yes: True
device: # set 'gpu:id' or 'cpu'
################### speech task: asr; engine_type: inference #######################
asr_inference:
# model_type choices=['deepspeech2offline_aishell']
model_type: 'deepspeech2offline_aishell'
am_model: # the pdmodel file of am static model [optional]
am_params: # the pdiparams file of am static model [optional]
lang: 'zh'
sample_rate: 16000
cfg_path:
decode_method:
force_yes: True
am_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
################### speech task: tts; engine_type: python #######################
tts_python:
# am (acoustic model) choices=['speedyspeech_csmsc', 'fastspeech2_csmsc',
# 'fastspeech2_ljspeech', 'fastspeech2_aishell3',
# 'fastspeech2_vctk']
am: 'fastspeech2_csmsc'
am_config:
am_ckpt:
am_stat:
phones_dict:
tones_dict:
speaker_dict:
spk_id: 0
# voc (vocoder) choices=['pwgan_csmsc', 'pwgan_ljspeech', 'pwgan_aishell3',
# 'pwgan_vctk', 'mb_melgan_csmsc']
voc: 'pwgan_csmsc'
voc_config:
voc_ckpt:
voc_stat:
# others
lang: 'zh'
device: # set 'gpu:id' or 'cpu'
################### speech task: tts; engine_type: inference #######################
tts_inference:
# am (acoustic model) choices=['speedyspeech_csmsc', 'fastspeech2_csmsc']
am: 'fastspeech2_csmsc'
am_model: # the pdmodel file of your am static model (XX.pdmodel)
am_params: # the pdiparams file of your am static model (XX.pdipparams)
am_sample_rate: 24000
phones_dict:
tones_dict:
speaker_dict:
spk_id: 0
am_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
# voc (vocoder) choices=['pwgan_csmsc', 'mb_melgan_csmsc','hifigan_csmsc']
voc: 'pwgan_csmsc'
voc_model: # the pdmodel file of your vocoder static model (XX.pdmodel)
voc_params: # the pdiparams file of your vocoder static model (XX.pdipparams)
voc_sample_rate: 24000
voc_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
# others
lang: 'zh'

@ -1,8 +0,0 @@
model: 'conformer_wenetspeech'
lang: 'zh'
sample_rate: 16000
cfg_path: # [optional]
ckpt_path: # [optional]
decode_method: 'attention_rescoring'
force_yes: True
device: # set 'gpu:id' or 'cpu'

@ -1,26 +0,0 @@
# This is the parameter configuration file for ASR server.
# These are the static models that support paddle inference.
##################################################################
# ACOUSTIC MODEL SETTING #
# am choices=['deepspeech2offline_aishell'] TODO
##################################################################
model_type: 'deepspeech2offline_aishell'
am_model: # the pdmodel file of am static model [optional]
am_params: # the pdiparams file of am static model [optional]
lang: 'zh'
sample_rate: 16000
cfg_path:
decode_method:
force_yes: True
am_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
##################################################################
# OTHERS #
##################################################################

@ -1,32 +0,0 @@
# This is the parameter configuration file for TTS server.
##################################################################
# ACOUSTIC MODEL SETTING #
# am choices=['speedyspeech_csmsc', 'fastspeech2_csmsc',
# 'fastspeech2_ljspeech', 'fastspeech2_aishell3',
# 'fastspeech2_vctk']
##################################################################
am: 'fastspeech2_csmsc'
am_config:
am_ckpt:
am_stat:
phones_dict:
tones_dict:
speaker_dict:
spk_id: 0
##################################################################
# VOCODER SETTING #
# voc choices=['pwgan_csmsc', 'pwgan_ljspeech', 'pwgan_aishell3',
# 'pwgan_vctk', 'mb_melgan_csmsc']
##################################################################
voc: 'pwgan_csmsc'
voc_config:
voc_ckpt:
voc_stat:
##################################################################
# OTHERS #
##################################################################
lang: 'zh'
device: # set 'gpu:id' or 'cpu'

@ -1,42 +0,0 @@
# This is the parameter configuration file for TTS server.
# These are the static models that support paddle inference.
##################################################################
# ACOUSTIC MODEL SETTING #
# am choices=['speedyspeech_csmsc', 'fastspeech2_csmsc']
##################################################################
am: 'fastspeech2_csmsc'
am_model: # the pdmodel file of your am static model (XX.pdmodel)
am_params: # the pdiparams file of your am static model (XX.pdipparams)
am_sample_rate: 24000
phones_dict:
tones_dict:
speaker_dict:
spk_id: 0
am_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
##################################################################
# VOCODER SETTING #
# voc choices=['pwgan_csmsc', 'mb_melgan_csmsc','hifigan_csmsc']
##################################################################
voc: 'pwgan_csmsc'
voc_model: # the pdmodel file of your vocoder static model (XX.pdmodel)
voc_params: # the pdiparams file of your vocoder static model (XX.pdipparams)
voc_sample_rate: 24000
voc_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
##################################################################
# OTHERS #
##################################################################
lang: 'zh'

@ -34,7 +34,7 @@ def init(config):
bool:
"""
# init api
api_list = list(config.engine_backend)
api_list = list(engine.split("_")[0] for engine in config.engine_list)
api_router = setup_router(api_list)
app.include_router(api_router)

@ -62,7 +62,7 @@ class ServerExecutor(BaseExecutor):
bool:
"""
# init api
api_list = list(config.engine_backend)
api_list = list(engine.split("_")[0] for engine in config.engine_list)
api_router = setup_router(api_list)
app.include_router(api_router)

@ -1,27 +1,107 @@
# This is the parameter configuration file for PaddleSpeech Serving.
##################################################################
#################################################################################
# SERVER SETTING #
##################################################################
#################################################################################
host: 127.0.0.1
port: 8090
##################################################################
# CONFIG FILE #
##################################################################
# add engine backend type (Options: asr, tts) and config file here.
# Adding a speech task to engine_backend means starting the service.
engine_backend:
asr: 'conf/asr/asr.yaml'
tts: 'conf/tts/tts.yaml'
# The engine_type of speech task needs to keep the same type as the config file of speech task.
# E.g: The engine_type of asr is 'python', the engine_backend of asr is 'XX/asr.yaml'
# E.g: The engine_type of asr is 'inference', the engine_backend of asr is 'XX/asr_pd.yaml'
#
# add engine type (Options: python, inference)
engine_type:
asr: 'python'
tts: 'python'
# The task format in the engin_list is: <speech task>_<engine type>
# task choices = ['asr_python', 'asr_inference', 'tts_python', 'tts_inference']
engine_list: ['asr_python', 'tts_python']
#################################################################################
# ENGINE CONFIG #
#################################################################################
################### speech task: asr; engine_type: python #######################
asr_python:
model: 'conformer_wenetspeech'
lang: 'zh'
sample_rate: 16000
cfg_path: # [optional]
ckpt_path: # [optional]
decode_method: 'attention_rescoring'
force_yes: True
device: # set 'gpu:id' or 'cpu'
################### speech task: asr; engine_type: inference #######################
asr_inference:
# model_type choices=['deepspeech2offline_aishell']
model_type: 'deepspeech2offline_aishell'
am_model: # the pdmodel file of am static model [optional]
am_params: # the pdiparams file of am static model [optional]
lang: 'zh'
sample_rate: 16000
cfg_path:
decode_method:
force_yes: True
am_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
################### speech task: tts; engine_type: python #######################
tts_python:
# am (acoustic model) choices=['speedyspeech_csmsc', 'fastspeech2_csmsc',
# 'fastspeech2_ljspeech', 'fastspeech2_aishell3',
# 'fastspeech2_vctk']
am: 'fastspeech2_csmsc'
am_config:
am_ckpt:
am_stat:
phones_dict:
tones_dict:
speaker_dict:
spk_id: 0
# voc (vocoder) choices=['pwgan_csmsc', 'pwgan_ljspeech', 'pwgan_aishell3',
# 'pwgan_vctk', 'mb_melgan_csmsc']
voc: 'pwgan_csmsc'
voc_config:
voc_ckpt:
voc_stat:
# others
lang: 'zh'
device: # set 'gpu:id' or 'cpu'
################### speech task: tts; engine_type: inference #######################
tts_inference:
# am (acoustic model) choices=['speedyspeech_csmsc', 'fastspeech2_csmsc']
am: 'fastspeech2_csmsc'
am_model: # the pdmodel file of your am static model (XX.pdmodel)
am_params: # the pdiparams file of your am static model (XX.pdipparams)
am_sample_rate: 24000
phones_dict:
tones_dict:
speaker_dict:
spk_id: 0
am_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
# voc (vocoder) choices=['pwgan_csmsc', 'mb_melgan_csmsc','hifigan_csmsc']
voc: 'pwgan_csmsc'
voc_model: # the pdmodel file of your vocoder static model (XX.pdmodel)
voc_params: # the pdiparams file of your vocoder static model (XX.pdipparams)
voc_sample_rate: 24000
voc_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
# others
lang: 'zh'

@ -1,8 +0,0 @@
model: 'conformer_wenetspeech'
lang: 'zh'
sample_rate: 16000
cfg_path: # [optional]
ckpt_path: # [optional]
decode_method: 'attention_rescoring'
force_yes: True
device: # set 'gpu:id' or 'cpu'

@ -1,26 +0,0 @@
# This is the parameter configuration file for ASR server.
# These are the static models that support paddle inference.
##################################################################
# ACOUSTIC MODEL SETTING #
# am choices=['deepspeech2offline_aishell'] TODO
##################################################################
model_type: 'deepspeech2offline_aishell'
am_model: # the pdmodel file of am static model [optional]
am_params: # the pdiparams file of am static model [optional]
lang: 'zh'
sample_rate: 16000
cfg_path:
decode_method:
force_yes: True
am_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
##################################################################
# OTHERS #
##################################################################

@ -1,32 +0,0 @@
# This is the parameter configuration file for TTS server.
##################################################################
# ACOUSTIC MODEL SETTING #
# am choices=['speedyspeech_csmsc', 'fastspeech2_csmsc',
# 'fastspeech2_ljspeech', 'fastspeech2_aishell3',
# 'fastspeech2_vctk']
##################################################################
am: 'fastspeech2_csmsc'
am_config:
am_ckpt:
am_stat:
phones_dict:
tones_dict:
speaker_dict:
spk_id: 0
##################################################################
# VOCODER SETTING #
# voc choices=['pwgan_csmsc', 'pwgan_ljspeech', 'pwgan_aishell3',
# 'pwgan_vctk', 'mb_melgan_csmsc']
##################################################################
voc: 'pwgan_csmsc'
voc_config:
voc_ckpt:
voc_stat:
##################################################################
# OTHERS #
##################################################################
lang: 'zh'
device: # set 'gpu:id' or 'cpu'

@ -1,42 +0,0 @@
# This is the parameter configuration file for TTS server.
# These are the static models that support paddle inference.
##################################################################
# ACOUSTIC MODEL SETTING #
# am choices=['speedyspeech_csmsc', 'fastspeech2_csmsc']
##################################################################
am: 'fastspeech2_csmsc'
am_model: # the pdmodel file of your am static model (XX.pdmodel)
am_params: # the pdiparams file of your am static model (XX.pdipparams)
am_sample_rate: 24000
phones_dict:
tones_dict:
speaker_dict:
spk_id: 0
am_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
##################################################################
# VOCODER SETTING #
# voc choices=['pwgan_csmsc', 'mb_melgan_csmsc','hifigan_csmsc']
##################################################################
voc: 'pwgan_csmsc'
voc_model: # the pdmodel file of your vocoder static model (XX.pdmodel)
voc_params: # the pdiparams file of your vocoder static model (XX.pdipparams)
voc_sample_rate: 24000
voc_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
##################################################################
# OTHERS #
##################################################################
lang: 'zh'

@ -26,7 +26,6 @@ from paddlespeech.s2t.frontend.featurizer.text_featurizer import TextFeaturizer
from paddlespeech.s2t.modules.ctc import CTCDecoder
from paddlespeech.s2t.utils.utility import UpdateConfig
from paddlespeech.server.engine.base_engine import BaseEngine
from paddlespeech.server.utils.config import get_config
from paddlespeech.server.utils.paddle_predictor import init_predictor
from paddlespeech.server.utils.paddle_predictor import run_model
@ -184,7 +183,7 @@ class ASREngine(BaseEngine):
def __init__(self):
super(ASREngine, self).__init__()
def init(self, config_file: str) -> bool:
def init(self, config: dict) -> bool:
"""init engine resource
Args:
@ -196,7 +195,7 @@ class ASREngine(BaseEngine):
self.input = None
self.output = None
self.executor = ASRServerExecutor()
self.config = get_config(config_file)
self.config = config
self.executor._init_from_path(
model_type=self.config.model_type,

@ -19,7 +19,6 @@ import paddle
from paddlespeech.cli.asr.infer import ASRExecutor
from paddlespeech.cli.log import logger
from paddlespeech.server.engine.base_engine import BaseEngine
from paddlespeech.server.utils.config import get_config
__all__ = ['ASREngine']
@ -40,7 +39,7 @@ class ASREngine(BaseEngine):
def __init__(self):
super(ASREngine, self).__init__()
def init(self, config_file: str) -> bool:
def init(self, config: dict) -> bool:
"""init engine resource
Args:
@ -52,8 +51,7 @@ class ASREngine(BaseEngine):
self.input = None
self.output = None
self.executor = ASRServerExecutor()
self.config = get_config(config_file)
self.config = config
try:
if self.config.device:
self.device = self.config.device

@ -28,11 +28,13 @@ def init_engine_pool(config) -> bool:
""" Init engine pool
"""
global ENGINE_POOL
for engine in config.engine_backend:
for engine_and_type in config.engine_list:
engine = engine_and_type.split("_")[0]
engine_type = engine_and_type.split("_")[1]
ENGINE_POOL[engine] = EngineFactory.get_engine(
engine_name=engine, engine_type=config.engine_type[engine])
if not ENGINE_POOL[engine].init(
config_file=config.engine_backend[engine]):
engine_name=engine, engine_type=engine_type)
if not ENGINE_POOL[engine].init(config=config[engine_and_type]):
return False
return True

@ -29,7 +29,6 @@ from paddlespeech.cli.utils import download_and_decompress
from paddlespeech.cli.utils import MODEL_HOME
from paddlespeech.server.engine.base_engine import BaseEngine
from paddlespeech.server.utils.audio_process import change_speed
from paddlespeech.server.utils.config import get_config
from paddlespeech.server.utils.errors import ErrorCode
from paddlespeech.server.utils.exception import ServerBaseException
from paddlespeech.server.utils.paddle_predictor import init_predictor
@ -357,11 +356,11 @@ class TTSEngine(BaseEngine):
"""
super(TTSEngine, self).__init__()
def init(self, config_file: str) -> bool:
def init(self, config: dict) -> bool:
self.executor = TTSServerExecutor()
try:
self.config = get_config(config_file)
self.config = config
self.executor._init_from_path(
am=self.config.am,
am_model=self.config.am_model,

@ -25,7 +25,6 @@ from paddlespeech.cli.log import logger
from paddlespeech.cli.tts.infer import TTSExecutor
from paddlespeech.server.engine.base_engine import BaseEngine
from paddlespeech.server.utils.audio_process import change_speed
from paddlespeech.server.utils.config import get_config
from paddlespeech.server.utils.errors import ErrorCode
from paddlespeech.server.utils.exception import ServerBaseException
@ -50,11 +49,11 @@ class TTSEngine(BaseEngine):
"""
super(TTSEngine, self).__init__()
def init(self, config_file: str) -> bool:
def init(self, config: dict) -> bool:
self.executor = TTSServerExecutor()
try:
self.config = get_config(config_file)
self.config = config
if self.config.device:
self.device = self.config.device
else:

@ -5,7 +5,7 @@ import os
import yaml
def change_speech_yaml(yaml_name: str, device: str):
def change_device(yamlfile: str, engine: str, device: str):
"""Change the settings of the device under the voice task configuration file
Args:
@ -13,68 +13,54 @@ def change_speech_yaml(yaml_name: str, device: str):
cpu (bool): True means set device to "cpu"
model_type (dict): change model type
"""
if "asr" in yaml_name:
dirpath = "./conf/asr/"
elif 'tts' in yaml_name:
dirpath = "./conf/tts/"
yamlfile = dirpath + yaml_name + ".yaml"
tmp_yamlfile = dirpath + yaml_name + "_tmp.yaml"
tmp_yamlfile = yamlfile.split(".yaml")[0] + "_tmp.yaml"
os.system("cp %s %s" % (yamlfile, tmp_yamlfile))
with open(tmp_yamlfile) as f, open(yamlfile, "w+", encoding="utf-8") as fw:
y = yaml.safe_load(f)
if device == 'cpu':
print("Set device: cpu")
if yaml_name == 'asr':
y['device'] = 'cpu'
elif yaml_name == 'asr_pd':
y['am_predictor_conf']['device'] = 'cpu'
elif yaml_name == 'tts':
y['device'] = 'cpu'
elif yaml_name == 'tts_pd':
y['am_predictor_conf']['device'] = 'cpu'
y['voc_predictor_conf']['device'] = 'cpu'
set_device = 'cpu'
elif device == 'gpu':
print("Set device: gpu")
if yaml_name == 'asr':
y['device'] = 'gpu:0'
elif yaml_name == 'asr_pd':
y['am_predictor_conf']['device'] = 'gpu:0'
elif yaml_name == 'tts':
y['device'] = 'gpu:0'
elif yaml_name == 'tts_pd':
y['am_predictor_conf']['device'] = 'gpu:0'
y['voc_predictor_conf']['device'] = 'gpu:0'
set_device = 'gpu:0'
else:
print("Please set correct device: cpu or gpu.")
print("The content of '%s': " % (yamlfile))
with open(tmp_yamlfile) as f, open(yamlfile, "w+", encoding="utf-8") as fw:
y = yaml.safe_load(f)
if engine == 'asr_python' or engine == 'tts_python':
y[engine]['device'] = set_device
elif engine == 'asr_inference':
y[engine]['am_predictor_conf']['device'] = set_device
elif engine == 'tts_inference':
y[engine]['am_predictor_conf']['device'] = set_device
y[engine]['voc_predictor_conf']['device'] = set_device
else:
print(
"Please set correct engine: asr_python, tts_python, asr_inference, tts_inference."
)
print(yaml.dump(y, default_flow_style=False, sort_keys=False))
yaml.dump(y, fw, allow_unicode=True)
os.system("rm %s" % (tmp_yamlfile))
print("Change %s successfully." % (yamlfile))
def change_app_yaml(task: str, engine_type: str):
def change_engine_type(yamlfile: str, engine_type):
"""Change the engine type and corresponding configuration file of the speech task in application.yaml
Args:
task (str): asr or tts
"""
yamlfile = "./conf/application.yaml"
tmp_yamlfile = "./conf/application_tmp.yaml"
tmp_yamlfile = yamlfile.split(".yaml")[0] + "_tmp.yaml"
os.system("cp %s %s" % (yamlfile, tmp_yamlfile))
speech_task = engine_type.split("_")[0]
with open(tmp_yamlfile) as f, open(yamlfile, "w+", encoding="utf-8") as fw:
y = yaml.safe_load(f)
y['engine_type'][task] = engine_type
path_list = ["./conf/", task, "/", task]
if engine_type == 'python':
path_list.append(".yaml")
elif engine_type == 'inference':
path_list.append("_pd.yaml")
y['engine_backend'][task] = ''.join(path_list)
print("The content of './conf/application.yaml': ")
engine_list = y['engine_list']
for engine in engine_list:
if speech_task in engine:
engine_list.remove(engine)
engine_list.append(engine_type)
y['engine_list'] = engine_list
print(yaml.dump(y, default_flow_style=False, sort_keys=False))
yaml.dump(y, fw, allow_unicode=True)
os.system("rm %s" % (tmp_yamlfile))
@ -83,32 +69,37 @@ def change_app_yaml(task: str, engine_type: str):
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument(
'--config_file',
type=str,
default='./conf/application.yaml',
help='server yaml file.')
parser.add_argument(
'--change_task',
type=str,
default=None,
help='Change task',
choices=[
'app-asr-python',
'app-asr-inference',
'app-tts-python',
'app-tts-inference',
'speech-asr-cpu',
'speech-asr-gpu',
'speech-asr_pd-cpu',
'speech-asr_pd-gpu',
'speech-tts-cpu',
'speech-tts-gpu',
'speech-tts_pd-cpu',
'speech-tts_pd-gpu',
'enginetype-asr_python',
'enginetype-asr_inference',
'enginetype-tts_python',
'enginetype-tts_inference',
'device-asr_python-cpu',
'device-asr_python-gpu',
'device-asr_inference-cpu',
'device-asr_inference-gpu',
'device-tts_python-cpu',
'device-tts_python-gpu',
'device-tts_inference-cpu',
'device-tts_inference-gpu',
],
required=True)
args = parser.parse_args()
types = args.change_task.split("-")
if types[0] == "app":
change_app_yaml(types[1], types[2])
elif types[0] == "speech":
change_speech_yaml(types[1], types[2])
if types[0] == "enginetype":
change_engine_type(args.config_file, types[1])
elif types[0] == "device":
change_device(args.config_file, types[1], types[2])
else:
print("Error change task, please check change_task.")

@ -1,27 +1,107 @@
# This is the parameter configuration file for PaddleSpeech Serving.
##################################################################
#################################################################################
# SERVER SETTING #
##################################################################
#################################################################################
host: 127.0.0.1
port: 8090
##################################################################
# CONFIG FILE #
##################################################################
# add engine backend type (Options: asr, tts) and config file here.
# Adding a speech task to engine_backend means starting the service.
engine_backend:
asr: 'conf/asr/asr.yaml'
tts: 'conf/tts/tts.yaml'
# The engine_type of speech task needs to keep the same type as the config file of speech task.
# E.g: The engine_type of asr is 'python', the engine_backend of asr is 'XX/asr.yaml'
# E.g: The engine_type of asr is 'inference', the engine_backend of asr is 'XX/asr_pd.yaml'
#
# add engine type (Options: python, inference)
engine_type:
asr: 'python'
tts: 'python'
# The task format in the engin_list is: <speech task>_<engine type>
# task choices = ['asr_python', 'asr_inference', 'tts_python', 'tts_inference']
engine_list: ['asr_python', 'tts_python']
#################################################################################
# ENGINE CONFIG #
#################################################################################
################### speech task: asr; engine_type: python #######################
asr_python:
model: 'conformer_wenetspeech'
lang: 'zh'
sample_rate: 16000
cfg_path: # [optional]
ckpt_path: # [optional]
decode_method: 'attention_rescoring'
force_yes: True
device: # set 'gpu:id' or 'cpu'
################### speech task: asr; engine_type: inference #######################
asr_inference:
# model_type choices=['deepspeech2offline_aishell']
model_type: 'deepspeech2offline_aishell'
am_model: # the pdmodel file of am static model [optional]
am_params: # the pdiparams file of am static model [optional]
lang: 'zh'
sample_rate: 16000
cfg_path:
decode_method:
force_yes: True
am_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
################### speech task: tts; engine_type: python #######################
tts_python:
# am (acoustic model) choices=['speedyspeech_csmsc', 'fastspeech2_csmsc',
# 'fastspeech2_ljspeech', 'fastspeech2_aishell3',
# 'fastspeech2_vctk']
am: 'fastspeech2_csmsc'
am_config:
am_ckpt:
am_stat:
phones_dict:
tones_dict:
speaker_dict:
spk_id: 0
# voc (vocoder) choices=['pwgan_csmsc', 'pwgan_ljspeech', 'pwgan_aishell3',
# 'pwgan_vctk', 'mb_melgan_csmsc']
voc: 'pwgan_csmsc'
voc_config:
voc_ckpt:
voc_stat:
# others
lang: 'zh'
device: # set 'gpu:id' or 'cpu'
################### speech task: tts; engine_type: inference #######################
tts_inference:
# am (acoustic model) choices=['speedyspeech_csmsc', 'fastspeech2_csmsc']
am: 'fastspeech2_csmsc'
am_model: # the pdmodel file of your am static model (XX.pdmodel)
am_params: # the pdiparams file of your am static model (XX.pdipparams)
am_sample_rate: 24000
phones_dict:
tones_dict:
speaker_dict:
spk_id: 0
am_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
# voc (vocoder) choices=['pwgan_csmsc', 'mb_melgan_csmsc','hifigan_csmsc']
voc: 'pwgan_csmsc'
voc_model: # the pdmodel file of your vocoder static model (XX.pdmodel)
voc_params: # the pdiparams file of your vocoder static model (XX.pdipparams)
voc_sample_rate: 24000
voc_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
# others
lang: 'zh'

@ -1,8 +0,0 @@
model: 'conformer_wenetspeech'
lang: 'zh'
sample_rate: 16000
cfg_path: # [optional]
ckpt_path: # [optional]
decode_method: 'attention_rescoring'
force_yes: True
device: # set 'gpu:id' or 'cpu'

@ -1,26 +0,0 @@
# This is the parameter configuration file for ASR server.
# These are the static models that support paddle inference.
##################################################################
# ACOUSTIC MODEL SETTING #
# am choices=['deepspeech2offline_aishell'] TODO
##################################################################
model_type: 'deepspeech2offline_aishell'
am_model: # the pdmodel file of am static model [optional]
am_params: # the pdiparams file of am static model [optional]
lang: 'zh'
sample_rate: 16000
cfg_path:
decode_method:
force_yes: True
am_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
##################################################################
# OTHERS #
##################################################################

@ -1,32 +0,0 @@
# This is the parameter configuration file for TTS server.
##################################################################
# ACOUSTIC MODEL SETTING #
# am choices=['speedyspeech_csmsc', 'fastspeech2_csmsc',
# 'fastspeech2_ljspeech', 'fastspeech2_aishell3',
# 'fastspeech2_vctk']
##################################################################
am: 'fastspeech2_csmsc'
am_config:
am_ckpt:
am_stat:
phones_dict:
tones_dict:
speaker_dict:
spk_id: 0
##################################################################
# VOCODER SETTING #
# voc choices=['pwgan_csmsc', 'pwgan_ljspeech', 'pwgan_aishell3',
# 'pwgan_vctk', 'mb_melgan_csmsc']
##################################################################
voc: 'pwgan_csmsc'
voc_config:
voc_ckpt:
voc_stat:
##################################################################
# OTHERS #
##################################################################
lang: 'zh'
device: # set 'gpu:id' or 'cpu'

@ -1,42 +0,0 @@
# This is the parameter configuration file for TTS server.
# These are the static models that support paddle inference.
##################################################################
# ACOUSTIC MODEL SETTING #
# am choices=['speedyspeech_csmsc', 'fastspeech2_csmsc']
##################################################################
am: 'fastspeech2_csmsc'
am_model: # the pdmodel file of your am static model (XX.pdmodel)
am_params: # the pdiparams file of your am static model (XX.pdipparams)
am_sample_rate: 24000
phones_dict:
tones_dict:
speaker_dict:
spk_id: 0
am_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
##################################################################
# VOCODER SETTING #
# voc choices=['pwgan_csmsc', 'mb_melgan_csmsc','hifigan_csmsc']
##################################################################
voc: 'pwgan_csmsc'
voc_model: # the pdmodel file of your vocoder static model (XX.pdmodel)
voc_params: # the pdiparams file of your vocoder static model (XX.pdipparams)
voc_sample_rate: 24000
voc_predictor_conf:
device: # set 'gpu:id' or 'cpu'
switch_ir_optim: True
glog_info: False # True -> print glog
summary: True # False -> do not show predictor config
##################################################################
# OTHERS #
##################################################################
lang: 'zh'

@ -99,8 +99,8 @@ echo "**************************************************************************
# start server: asr engine type: python; tts engine type: python; device: cpu
python change_yaml.py --change_task speech-asr-cpu # change asr.yaml device: cpu
python change_yaml.py --change_task speech-tts-cpu # change tts.yaml device: cpu
python change_yaml.py --change_task device-asr_python-cpu # change asr.yaml device: cpu
python change_yaml.py --change_task device-tts_python-cpu # change tts.yaml device: cpu
echo "Start the service: asr engine type: python; tts engine type: python; device: cpu" | tee -a ./log/test_result.log
((target_start_num+=1))
@ -125,8 +125,8 @@ echo "**************************************************************************
# start server: asr engine type: inference; tts engine type: inference; device: gpu
python change_yaml.py --change_task app-asr-inference # change application.yaml, asr engine_type: inference; asr engine_backend: asr_pd.yaml
python change_yaml.py --change_task app-tts-inference # change application.yaml, tts engine_type: inference; tts engine_backend: tts_pd.yaml
python change_yaml.py --change_task enginetype-asr_inference # change application.yaml, asr engine_type: inference; asr engine_backend: asr_pd.yaml
python change_yaml.py --change_task enginetype-tts_inference # change application.yaml, tts engine_type: inference; tts engine_backend: tts_pd.yaml
echo "Start the service: asr engine type: inference; tts engine type: inference; device: gpu" | tee -a ./log/test_result.log
((target_start_num+=1))
@ -151,8 +151,8 @@ echo "**************************************************************************
# start server: asr engine type: inference; tts engine type: inference; device: cpu
python change_yaml.py --change_task speech-asr_pd-cpu # change asr_pd.yaml device: cpu
python change_yaml.py --change_task speech-tts_pd-cpu # change tts_pd.yaml device: cpu
python change_yaml.py --change_task device-asr_inference-cpu # change asr_pd.yaml device: cpu
python change_yaml.py --change_task device-tts_inference-cpu # change tts_pd.yaml device: cpu
echo "start the service: asr engine type: inference; tts engine type: inference; device: cpu" | tee -a ./log/test_result.log
((target_start_num+=1))
@ -182,4 +182,5 @@ echo "***************** Here are all the test results ********************"
cat ./log/test_result.log
# Restoring conf is the same as demos/speech_server
rm -rf ./conf
cp ../../../demos/speech_server/conf/ ./ -rf
Loading…
Cancel
Save