Skip to main content

麻瓜语音

Project description

MUGGLE-SPEECH - 麻瓜中文语音识别

注:MUGGLE-OCR当初是配套 Captcha-Trainer 项目的SDK调用,经过几轮的重构和迭代,新的框架已经准备好了,将在未来不久的某一天,重新归来。

MUGGLE-SPEECH 是基于 Transformer 的端到端语音识别模型,采用ONNX部署方案。目前在语音识别领域只是试试水,毕竟样本受限,目前免费白嫖到的公开样本来自:AISHELL-1, AISHELL-3, MAGICDATA,尝试申请AISHELL-2,很难过被告知不对个人提供开源。十分欢迎愿意贡献数据集的朋友们,希望能够给社区尽一份绵薄之力。

训练相关代码将整合到 MUGGLE-DL 框架中再开源,初次接触语音识别,还有很多不足的地方,将会慢慢改进。

以下是Python-SDK调用方法

import time

from muggle_speech import MuggleSpeech

sdk = MuggleSpeech(mode='wave')
# 文件格式必须是 wav 格式

if __name__ == '__main__':
    for i in range(1000):
        st = time.time()
        # 从 bytes 打开
        # wav_bytes = open(r"test.wav", "rb").read()
        # inputs = sdk.from_bytes(wav_bytes)
        # 从 文件 打开
        inputs = sdk.from_file(r"test.wav")
        predict_text = sdk.predict_file(inputs)
        print(predict_text, time.time() - st)

测试结果:

CPU 识别12个字大概 80-100毫秒 左右。

安装命令

pip install muggle-speech

交流群

857149419 (1群,已满)

934889548(2群)

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

muggle_speech-0.1.1.tar.gz (65.4 MB view details)

Uploaded Source

File details

Details for the file muggle_speech-0.1.1.tar.gz.

File metadata

  • Download URL: muggle_speech-0.1.1.tar.gz
  • Upload date:
  • Size: 65.4 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.0 CPython/3.9.10

File hashes

Hashes for muggle_speech-0.1.1.tar.gz
Algorithm Hash digest
SHA256 587ea4e57d81a3f280c86becd4373766076e3181c55df97e60e293885835c656
MD5 efbf19935a90be16d41d2d7766229b3c
BLAKE2b-256 e9f37d5d3862afff11e1e1b4f58cb8cfebc47c2d30faf484ced6a686d32a5f43

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page