msb-public
/
PaddleSpeech
mirror of https://github.com/PaddlePaddle/PaddleSpeech
1
0
Code Issues Projects Releases Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from '662eb3fbb8'
${ noResults }
PaddleSpeech/demos/speech_server/README_cn.md

20 KiB
Raw Blame History Escape

(简体中文|English)

语音服务

介绍

这个demo是一个启动离线语音服务和访问服务的实现。它可以通过使用paddlespeech_server 和 paddlespeech_client的单个命令或 python 的几行代码来实现。

使用方法

1. 安装

请看 安装文档.

推荐使用 paddlepaddle 2.2.2 或以上版本。 你可以从 medium，hard 两种方式中选择一种方式安装 PaddleSpeech。

2. 准备配置文件

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)以及cls(音频分类)。 目前引擎类型支持两种形式：python 及 inference (Paddle Inference) 注意： 如果在容器里可正常启动服务，但客户端访问 ip 不可达，可尝试将配置文件中 host 地址换成本地 ip 地址。

ASR client 的输入是一个 WAV 文件（.wav），并且采样率必须与模型的采样率相同。

可以下载此 ASR client的示例音频：

wget -c https://paddlespeech.bj.bcebos.com/PaddleAudio/zh.wav https://paddlespeech.bj.bcebos.com/PaddleAudio/en.wav

3. 服务端使用方法

  • 命令行 (推荐使用)

    # 启动服务
paddlespeech_server start --config_file ./conf/application.yaml

    使用方法：

    paddlespeech_server start --help

    参数:

    • config_file: 服务的配置文件，默认： ./conf/application.yaml
    • log_file: log 文件. 默认：./log/paddlespeech.log

    输出:

    [2022-02-23 11:17:32] [INFO] [server.py:64] Started server process [6384]
INFO:     Waiting for application startup.
[2022-02-23 11:17:32] [INFO] [on.py:26] Waiting for application startup.
INFO:     Application startup complete.
[2022-02-23 11:17:32] [INFO] [on.py:38] Application startup complete.
INFO:     Uvicorn running on http://127.0.0.1:8090 (Press CTRL+C to quit)
[2022-02-23 11:17:32] [INFO] [server.py:204] Uvicorn running on http://127.0.0.1:8090 (Press CTRL+C to quit)

  • Python API

    from paddlespeech.server.bin.paddlespeech_server import ServerExecutor

server_executor = ServerExecutor()
server_executor(
    config_file="./conf/application.yaml", 
    log_file="./log/paddlespeech.log")

    输出：

    INFO:     Started server process [529]
[2022-02-23 14:57:56] [INFO] [server.py:64] Started server process [529]
INFO:     Waiting for application startup.
[2022-02-23 14:57:56] [INFO] [on.py:26] Waiting for application startup.
INFO:     Application startup complete.
[2022-02-23 14:57:56] [INFO] [on.py:38] Application startup complete.
INFO:     Uvicorn running on http://127.0.0.1:8090 (Press CTRL+C to quit)
[2022-02-23 14:57:56] [INFO] [server.py:204] Uvicorn running on http://127.0.0.1:8090 (Press CTRL+C to quit)

4. ASR 客户端使用方法

注意： 初次使用客户端时响应时间会略长

  • 命令行 (推荐使用)

    paddlespeech_client asr --server_ip 127.0.0.1 --port 8090 --input ./zh.wav

    使用帮助:

    paddlespeech_client asr --help

    参数:

    • server_ip: 服务端ip地址，默认: 127.0.0.1。
    • port: 服务端口，默认: 8090。
    • input(必须输入): 用于识别的音频文件。
    • sample_rate: 音频采样率，默认值：16000。
    • lang: 模型语言，默认值：zh_cn。
    • audio_format: 音频格式，默认值：wav。

    输出:

    [2022-02-23 18:11:22,819] [    INFO] - {'success': True, 'code': 200, 'message': {'description': 'success'}, 'result': {'transcription': '我认为跑步最重要的就是给我带来了身体健康'}}
[2022-02-23 18:11:22,820] [    INFO] - time cost 0.689145 s.

  • Python API

    from paddlespeech.server.bin.paddlespeech_client import ASRClientExecutor
import json

asrclient_executor = ASRClientExecutor()
res = asrclient_executor(
    input="./zh.wav",
    server_ip="127.0.0.1",
    port=8090,
    sample_rate=16000,
    lang="zh_cn",
    audio_format="wav")
print(res.json())

    输出:

    {'success': True, 'code': 200, 'message': {'description': 'success'}, 'result': {'transcription': '我认为跑步最重要的就是给我带来了身体健康'}}

5. TTS 客户端使用方法

注意： 初次使用客户端时响应时间会略长

  • 命令行 (推荐使用)

    paddlespeech_client tts --server_ip 127.0.0.1 --port 8090 --input "您好，欢迎使用百度飞桨语音合成服务。" --output output.wav

    使用帮助:

    paddlespeech_client tts --help

    参数:

    • server_ip: 服务端ip地址，默认: 127.0.0.1。
    • port: 服务端口，默认: 8090。
    • input(必须输入): 待合成的文本。
    • spk_id: 说话人 id，用于多说话人语音合成，默认值： 0。
    • speed: 音频速度，该值应设置在 0 到 3 之间。 默认值：1.0
    • volume: 音频音量，该值应设置在 0 到 3 之间。 默认值： 1.0
    • sample_rate: 采样率，可选 [0, 8000, 16000]，默认与模型相同。 默认值：0
    • output: 输出音频的路径， 默认值：None，表示不保存音频到本地。

    输出:

    [2022-02-23 15:20:37,875] [    INFO] - {'description': 'success.'}
[2022-02-23 15:20:37,875] [    INFO] - Save synthesized audio successfully on output.wav.
[2022-02-23 15:20:37,875] [    INFO] - Audio duration: 3.612500 s.
[2022-02-23 15:20:37,875] [    INFO] - Response time: 0.348050 s.

  • Python API

    from paddlespeech.server.bin.paddlespeech_client import TTSClientExecutor
import json

ttsclient_executor = TTSClientExecutor()
res = ttsclient_executor(
    input="您好，欢迎使用百度飞桨语音合成服务。",
    server_ip="127.0.0.1",
    port=8090,
    spk_id=0,
    speed=1.0,
    volume=1.0,
    sample_rate=0,
    output="./output.wav")

response_dict = res.json()
print(response_dict["message"])
print("Save synthesized audio successfully on %s." % (response_dict['result']['save_path']))
print("Audio duration: %f s." %(response_dict['result']['duration']))

    输出:

    {'description': 'success.'}
Save synthesized audio successfully on ./output.wav.
Audio duration: 3.612500 s.

    6. CLS 客户端使用方法

    注意： 初次使用客户端时响应时间会略长

    • 命令行 (推荐使用)
    paddlespeech_client cls --server_ip 127.0.0.1 --port 8090 --input ./zh.wav

    使用帮助:

    paddlespeech_client cls --help

    参数:

    • server_ip: 服务端ip地址，默认: 127.0.0.1。
    • port: 服务端口，默认: 8090。
    • input(必须输入): 用于分类的音频文件。
    • topk: 分类结果的topk。

    输出:

    [2022-03-09 20:44:39,974] [    INFO] - {'success': True, 'code': 200, 'message': {'description': 'success'}, 'result': {'topk': 1, 'results': [{'class_name': 'Speech', 'prob': 0.9027184844017029}]}}
[2022-03-09 20:44:39,975] [    INFO] - Response time 0.104360 s.

  • Python API

    from paddlespeech.server.bin.paddlespeech_client import CLSClientExecutor
import json

clsclient_executor = CLSClientExecutor()
res = clsclient_executor(
    input="./zh.wav",
    server_ip="127.0.0.1",
    port=8090,
    topk=1)
print(res.json())

    输出:

    {'success': True, 'code': 200, 'message': {'description': 'success'}, 'result': {'topk': 1, 'results': [{'class_name': 'Speech', 'prob': 0.9027184844017029}]}}

7. 声纹客户端使用方法

7.1 提取声纹特征

注意： 初次使用客户端时响应时间会略长

  • 命令行 (推荐使用)
paddlespeech_client vector --task spk  --server_ip 127.0.0.1 --port 8590 --input 85236145389.wav
  • 使用帮助:
paddlespeech_client vector --help
  • 参数:
    • server_ip: 服务端ip地址，默认: 127.0.0.1。
    • port: 服务端口，默认: 8090。
    • input(必须输入): 用于识别的音频文件。
    • task: vector 的任务，可选spk或者score。默认是 spk。
    • enroll: 注册音频；。
    • test: 测试音频。
  • 输出:
  [2022-05-08 00:18:44,249] [    INFO] - vector http client start
  [2022-05-08 00:18:44,250] [    INFO] - the input audio: 85236145389.wav
  [2022-05-08 00:18:44,250] [    INFO] - endpoint: http://127.0.0.1:8590/paddlespeech/vector
  [2022-05-08 00:18:44,250] [    INFO] - http://127.0.0.1:8590/paddlespeech/vector
  [2022-05-08 00:18:44,406] [    INFO] - The vector: {'success': True, 'code': 200, 'message': {'description': 'success'}, 'result': {'vec': [1.421751856803894, 5.626245498657227, -5.342077255249023, 1.1773887872695923, 3.3080549240112305, 1.7565933465957642, 5.167886257171631, 10.806358337402344, -3.8226819038391113, -5.614140033721924, 2.6238479614257812, -0.8072972893714905, 1.9635076522827148, -7.312870025634766, 0.011035939678549767, -9.723129272460938, 0.6619706153869629, -6.976806163787842, 10.213476181030273, 7.494769096374512, 2.9105682373046875, 3.8949244022369385, 3.799983501434326, 7.106168746948242, 16.90532875061035, -7.149388313293457, 8.733108520507812, 3.423006296157837, -4.831653594970703, -11.403363227844238, 11.232224464416504, 7.127461910247803, -4.282842636108398, 2.452359437942505, -5.130749702453613, -18.17766761779785, -2.6116831302642822, -11.000344276428223, -6.731433391571045, 1.6564682722091675, 0.7618281245231628, 1.125300407409668, -2.0838370323181152, 4.725743293762207, -8.782588005065918, -3.5398752689361572, 3.8142364025115967, 5.142068862915039, 2.1620609760284424, 4.09643030166626, -6.416214942932129, 12.747446060180664, 1.9429892301559448, -15.15294361114502, 6.417416095733643, 16.09701156616211, -9.716667175292969, -1.9920575618743896, -3.36494779586792, -1.8719440698623657, 11.567351341247559, 3.6978814601898193, 11.258262634277344, 7.442368507385254, 9.183408737182617, 4.528149127960205, -1.2417854070663452, 4.395912170410156, 6.6727728843688965, 5.88988733291626, 7.627128601074219, -0.6691966652870178, -11.889698028564453, -9.20886516571045, -7.42740535736084, -3.777663230895996, 6.917238712310791, -9.848755836486816, -2.0944676399230957, -5.1351165771484375, 0.4956451654434204, 9.317537307739258, -5.914181232452393, -1.809860348701477, -0.11738915741443634, -7.1692705154418945, -1.057827353477478, -5.721670627593994, -5.117385387420654, 16.13765525817871, -4.473617076873779, 7.6624321937561035, -0.55381840467453, 9.631585121154785, -6.470459461212158, -8.548508644104004, 4.371616840362549, -0.7970245480537415, 4.4789886474609375, -2.975860834121704, 3.2721822261810303, 2.838287830352783, 5.134591102600098, -9.19079875946045, -0.5657302737236023, -4.8745832443237305, 2.3165574073791504, -5.984319686889648, -2.1798853874206543, 0.3554139733314514, -0.3178512752056122, 9.493552207946777, 2.1144471168518066, 4.358094692230225, -12.089824676513672, 8.451693534851074, -7.925466537475586, 4.624246597290039, 4.428936958312988, 18.69200897216797, -2.6204581260681152, -5.14918851852417, -0.3582090139389038, 8.488558769226074, 4.98148775100708, -9.326835632324219, -2.2544219493865967, 6.641760349273682, 1.2119598388671875, 10.977124214172363, 16.555034637451172, 3.3238420486450195, 9.551861763000488, -1.6676981449127197, -0.7953944206237793, -8.605667114257812, -0.4735655188560486, 2.674196243286133, -5.359177112579346, -2.66738224029541, 0.6660683155059814, 15.44322681427002, 4.740593433380127, -3.472534418106079, 11.592567443847656, -2.0544962882995605, 1.736127495765686, -8.265326499938965, -9.30447769165039, 5.406829833984375, -1.518022894859314, -7.746612548828125, -6.089611053466797, 0.07112743705511093, -0.3490503430366516, -8.64989185333252, -9.998957633972168, -2.564845085144043, -0.5399947762489319, 2.6018123626708984, -0.3192799389362335, -1.8815255165100098, -2.0721492767333984, -3.410574436187744, -8.29980754852295, 1.483638048171997, -15.365986824035645, -8.288211822509766, 3.884779930114746, -3.4876468181610107, 7.362999439239502, 0.4657334089279175, 3.1326050758361816, 12.438895225524902, -1.8337041139602661, 4.532927989959717, 2.7264339923858643, 10.14534854888916, -6.521963596343994, 2.897155523300171, -3.392582654953003, 5.079153060913086, 7.7597246170043945, 4.677570819854736, 5.845779895782471, 2.402411460876465, 7.7071051597595215, 3.9711380004882812, -6.39003849029541, 6.12687873840332, -3.776029348373413, -11.118121147155762]}}
  [2022-05-08 00:18:44,406] [    INFO] - Response time 0.156481 s.
  • Python API
from paddlespeech.server.bin.paddlespeech_client import VectorClientExecutor

vectorclient_executor = VectorClientExecutor()
res = vectorclient_executor(
    input="85236145389.wav",
    server_ip="127.0.0.1",
    port=8590,
    task="spk")
print(res)
  • 输出:
  {'success': True, 'code': 200, 'message': {'description': 'success'}, 'result': {'vec': [1.421751856803894, 5.626245498657227, -5.342077255249023, 1.1773887872695923, 3.3080549240112305, 1.7565933465957642, 5.167886257171631, 10.806358337402344, -3.8226819038391113, -5.614140033721924, 2.6238479614257812, -0.8072972893714905, 1.9635076522827148, -7.312870025634766, 0.011035939678549767, -9.723129272460938, 0.6619706153869629, -6.976806163787842, 10.213476181030273, 7.494769096374512, 2.9105682373046875, 3.8949244022369385, 3.799983501434326, 7.106168746948242, 16.90532875061035, -7.149388313293457, 8.733108520507812, 3.423006296157837, -4.831653594970703, -11.403363227844238, 11.232224464416504, 7.127461910247803, -4.282842636108398, 2.452359437942505, -5.130749702453613, -18.17766761779785, -2.6116831302642822, -11.000344276428223, -6.731433391571045, 1.6564682722091675, 0.7618281245231628, 1.125300407409668, -2.0838370323181152, 4.725743293762207, -8.782588005065918, -3.5398752689361572, 3.8142364025115967, 5.142068862915039, 2.1620609760284424, 4.09643030166626, -6.416214942932129, 12.747446060180664, 1.9429892301559448, -15.15294361114502, 6.417416095733643, 16.09701156616211, -9.716667175292969, -1.9920575618743896, -3.36494779586792, -1.8719440698623657, 11.567351341247559, 3.6978814601898193, 11.258262634277344, 7.442368507385254, 9.183408737182617, 4.528149127960205, -1.2417854070663452, 4.395912170410156, 6.6727728843688965, 5.88988733291626, 7.627128601074219, -0.6691966652870178, -11.889698028564453, -9.20886516571045, -7.42740535736084, -3.777663230895996, 6.917238712310791, -9.848755836486816, -2.0944676399230957, -5.1351165771484375, 0.4956451654434204, 9.317537307739258, -5.914181232452393, -1.809860348701477, -0.11738915741443634, -7.1692705154418945, -1.057827353477478, -5.721670627593994, -5.117385387420654, 16.13765525817871, -4.473617076873779, 7.6624321937561035, -0.55381840467453, 9.631585121154785, -6.470459461212158, -8.548508644104004, 4.371616840362549, -0.7970245480537415, 4.4789886474609375, -2.975860834121704, 3.2721822261810303, 2.838287830352783, 5.134591102600098, -9.19079875946045, -0.5657302737236023, -4.8745832443237305, 2.3165574073791504, -5.984319686889648, -2.1798853874206543, 0.3554139733314514, -0.3178512752056122, 9.493552207946777, 2.1144471168518066, 4.358094692230225, -12.089824676513672, 8.451693534851074, -7.925466537475586, 4.624246597290039, 4.428936958312988, 18.69200897216797, -2.6204581260681152, -5.14918851852417, -0.3582090139389038, 8.488558769226074, 4.98148775100708, -9.326835632324219, -2.2544219493865967, 6.641760349273682, 1.2119598388671875, 10.977124214172363, 16.555034637451172, 3.3238420486450195, 9.551861763000488, -1.6676981449127197, -0.7953944206237793, -8.605667114257812, -0.4735655188560486, 2.674196243286133, -5.359177112579346, -2.66738224029541, 0.6660683155059814, 15.44322681427002, 4.740593433380127, -3.472534418106079, 11.592567443847656, -2.0544962882995605, 1.736127495765686, -8.265326499938965, -9.30447769165039, 5.406829833984375, -1.518022894859314, -7.746612548828125, -6.089611053466797, 0.07112743705511093, -0.3490503430366516, -8.64989185333252, -9.998957633972168, -2.564845085144043, -0.5399947762489319, 2.6018123626708984, -0.3192799389362335, -1.8815255165100098, -2.0721492767333984, -3.410574436187744, -8.29980754852295, 1.483638048171997, -15.365986824035645, -8.288211822509766, 3.884779930114746, -3.4876468181610107, 7.362999439239502, 0.4657334089279175, 3.1326050758361816, 12.438895225524902, -1.8337041139602661, 4.532927989959717, 2.7264339923858643, 10.14534854888916, -6.521963596343994, 2.897155523300171, -3.392582654953003, 5.079153060913086, 7.7597246170043945, 4.677570819854736, 5.845779895782471, 2.402411460876465, 7.7071051597595215, 3.9711380004882812, -6.39003849029541, 6.12687873840332, -3.776029348373413, -11.118121147155762]}}

7.2 音频声纹打分

注意： 初次使用客户端时响应时间会略长

  • 命令行 (推荐使用)
paddlespeech_client vector --task score  --server_ip 127.0.0.1 --port 8590 --enroll 85236145389.wav --test 123456789.wav
  • 使用帮助:
paddlespeech_client vector --help
  • 参数:
    • server_ip: 服务端ip地址，默认: 127.0.0.1。
    • port: 服务端口，默认: 8090。
    • input(必须输入): 用于识别的音频文件。
    • task: vector 的任务，可选spk或者score。默认是 spk。
    • enroll: 注册音频；。
    • test: 测试音频。
  • 输出:
  [2022-05-09 10:28:40,556] [    INFO] - vector score http client start
  [2022-05-09 10:28:40,556] [    INFO] - enroll audio: 85236145389.wav, test audio: 123456789.wav
  [2022-05-09 10:28:40,556] [    INFO] - endpoint: http://127.0.0.1:8590/paddlespeech/vector/score
  [2022-05-09 10:28:40,731] [    INFO] - The vector score is: {'success': True, 'code': 200, 'message': {'description': 'success'}, 'result': {'score': 0.4292638897895813}}
  [2022-05-09 10:28:40,731] [    INFO] - The vector: None
  [2022-05-09 10:28:40,731] [    INFO] - Response time 0.175514 s.
  • Python API
from paddlespeech.server.bin.paddlespeech_client import VectorClientExecutor

vectorclient_executor = VectorClientExecutor()
res = vectorclient_executor(
    input=None,
    enroll_audio="85236145389.wav",
    test_audio="123456789.wav",
    server_ip="127.0.0.1",
    port=8590,
    task="score")
print(res)
  • 输出:
[2022-05-09 10:34:54,769] [    INFO] - vector score http client start
[2022-05-09 10:34:54,771] [    INFO] - enroll audio: 85236145389.wav, test audio: 123456789.wav
[2022-05-09 10:34:54,771] [    INFO] - endpoint: http://127.0.0.1:8590/paddlespeech/vector/score
[2022-05-09 10:34:55,026] [    INFO] - The vector score is: {'success': True, 'code': 200, 'message': {'description': 'success'}, 'result': {'score': 0.4292638897895813}}

8. 标点预测

注意： 初次使用客户端时响应时间会略长

  • 命令行 (推荐使用)
paddlespeech_client text --server_ip 127.0.0.1 --port 8090 --input "我认为跑步最重要的就是给我带来了身体健康"

使用帮助:

paddlespeech_client text --help

参数:

  • server_ip: 服务端ip地址，默认: 127.0.0.1。
  • port: 服务端口，默认: 8090。
  • input(必须输入): 用于标点预测的文本内容。

输出:

  [2022-05-09 18:19:04,397] [    INFO] - The punc text: 我认为跑步最重要的就是给我带来了身体健康。
  [2022-05-09 18:19:04,397] [    INFO] - Response time 0.092407 s.

  • Python API

    from paddlespeech.server.bin.paddlespeech_client import TextClientExecutor

textclient_executor = TextClientExecutor()
res = textclient_executor(
    input="我认为跑步最重要的就是给我带来了身体健康",
    server_ip="127.0.0.1",
    port=8390,)
print(res)

    输出:

    我认为跑步最重要的就是给我带来了身体健康。

服务支持的模型

ASR支持的模型

通过 paddlespeech_server stats --task asr 获取ASR服务支持的所有模型，其中静态模型可用于 paddle inference 推理。

TTS支持的模型

通过 paddlespeech_server stats --task tts 获取TTS服务支持的所有模型，其中静态模型可用于 paddle inference 推理。

CLS支持的模型

通过 paddlespeech_server stats --task cls 获取CLS服务支持的所有模型，其中静态模型可用于 paddle inference 推理。

Vector支持的模型

通过 paddlespeech_server stats --task vector 获取Vector服务支持的所有模型。

Text支持的模型

通过 paddlespeech_server stats --task text 获取Text服务支持的所有模型。