paddle-nnAudio 0.0.2

A fast GPU audio processing toolbox with 1D convolutional neural network

paddle-nnAudio

paddle-nnAudio 是基于 PaddlePaddle 卷积神经网络作为后端的音频处理工具箱。通过这种方式， 梅尔频谱可以在神经网络训练过程中实时从音频生成，并且傅里叶核（如 CQT 核）可以被训练。paddle-nnAudio 移植自 nnAudio，旨在为 PaddlePaddle 生态提供类似的音频处理能力。

安装

pip install git+https://github.com/PlumBlossomMaid/paddle-nnAudio.git

或

pip install paddle-nnAudio

快速开始

import paddle
import librosa
import numpy as np
from ppAudio.features import MelSpectrogram

def main() -> None:
    paddle.device.set_device("gpu:0") # 使用gpu:0进行运算
    example_y, example_sr = librosa.load(librosa.example('vibeace', hq=False)) # 加载音频

    n_fft, win_length = (512, 400)
    melspec = MelSpectrogram(n_fft=n_fft, win_length=win_length, hop_length=512)
    X = melspec(paddle.to_tensor(example_y).unsqueeze(0)).squeeze() # 将波形前向传播以获取 spectrogram
    X_librosa = librosa.feature.melspectrogram(example_y, n_fft=n_fft, win_length=win_length, hop_length=512) # 设置对照
    assert np.allclose(X.cpu(), X_librosa, rtol=1e-3, atol=1e-3) # 精度对齐
    print("done")


if __name__ == "__main__":
    main()

依赖项

• Numpy >= 1.14.5
• Scipy >= 1.2.0
• PaddlePaddle >= 2.0.0 (或PaddlePaddle-gpu >= 2.0.0)
• Python >= 3.6
• librosa = 0.7.0

引用

如果您使用了 paddle-nnAudio，请引用原 nnAudio 的论文：

K. W. Cheuk, H. Anderson, K. Agres and D. Herremans, "nnAudio: An on-the-Fly GPU Audio to Spectrogram Conversion Toolbox Using 1D Convolutional Neural Networks," in IEEE Access, vol. 8, pp. 161981-162003, 2020, doi: 10.1109/ACCESS.2020.3019084.

BibTex

@ARTICLE{9174990,
  author={K. W. {Cheuk} and H. {Anderson} and K. {Agres} and D. {Herremans}},
  journal={IEEE Access}, 
  title={nnAudio: An on-the-Fly GPU Audio to Spectrogram Conversion Toolbox Using 1D Convolutional Neural Networks}, 
  year={2020},
  volume={8},
  number={},
  pages={161981-162003},
  doi={10.1109/ACCESS.2020.3019084}}

许可证

MIT License