Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/spijkervet/torchaudio-augmentations
Audio transformations library for PyTorch
https://github.com/spijkervet/torchaudio-augmentations
audio machine-learning pytorch
Last synced: 17 days ago
JSON representation
Audio transformations library for PyTorch
- Host: GitHub
- URL: https://github.com/spijkervet/torchaudio-augmentations
- Owner: Spijkervet
- License: mit
- Created: 2021-01-24T00:01:58.000Z (almost 4 years ago)
- Default Branch: master
- Last Pushed: 2022-04-19T12:45:17.000Z (over 2 years ago)
- Last Synced: 2024-10-12T05:46:26.265Z (about 1 month ago)
- Topics: audio, machine-learning, pytorch
- Language: Python
- Homepage:
- Size: 1.19 MB
- Stars: 223
- Watchers: 2
- Forks: 26
- Open Issues: 6
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# PyTorch Audio Augmentations
[](https://codecov.io/gh/Spijkervet/torchaudio-augmentations)
[](https://pepy.tech/project/torchaudio-augmentations)
[](https://zenodo.org/record/4748582#)
Audio data augmentations library for PyTorch for audio in the time-domain. The focus of this repository is to:
- Provide many audio transformations in an easy Python interface.
- Have a high test coverage.
- Easily control stochastic (sequential) audio transformations.
- Make every audio transformation differentiable with PyTorch's `nn.Module`.
- Optimise audio transformations for CPU and GPU.
It supports stochastic transformations as used often in self-supervised, semi-supervised learning methods. One can apply a single stochastic augmentation or create as many stochastically transformed audio examples from a single interface.
This package follows the conventions set out by `torchvision` and `torchaudio`, with audio defined as a tensor of `[channel, time]`, or a batched representation `[batch, channel, time]`. Each individual augmentation can be initialized on its own, or be wrapped around a `RandomApply` interface which will apply the augmentation with probability `p`.
## Usage
We can define a single or several audio augmentations, which are applied sequentially to an audio waveform.
```python
from audio_augmentations import *
audio, sr = torchaudio.load("tests/classical.00002.wav")
num_samples = sr * 5
transforms = [
RandomResizedCrop(n_samples=num_samples),
RandomApply([PolarityInversion()], p=0.8),
RandomApply([Noise(min_snr=0.001, max_snr=0.005)], p=0.3),
RandomApply([Gain()], p=0.2),
HighLowPass(sample_rate=sr), # this augmentation will always be applied in this aumgentation chain!
RandomApply([Delay(sample_rate=sr)], p=0.5),
RandomApply([PitchShift(
n_samples=num_samples,
sample_rate=sr
)], p=0.4),
RandomApply([Reverb(sample_rate=sr)], p=0.3)
]
```
We can also define a stochastic augmentation on multiple transformations. The following will apply both polarity inversion and white noise with a probability of 80%, a gain of 20%, and delay and reverb with a probability of 50%:
```python
transforms = [
RandomResizedCrop(n_samples=num_samples),
RandomApply([PolarityInversion(), Noise(min_snr=0.001, max_snr=0.005)], p=0.8),
RandomApply([Gain()], p=0.2),
RandomApply([Delay(sample_rate=sr), Reverb(sample_rate=sr)], p=0.5)
]
```
We can return either one or many versions of the same audio example:
```python
transform = Compose(transforms=transforms)
transformed_audio = transform(audio)
>> transformed_audio.shape = [num_channels, num_samples]
```
```
audio = torchaudio.load("testing/classical.00002.wav")
transform = ComposeMany(transforms=transforms, num_augmented_samples=4)
transformed_audio = transform(audio)
>> transformed_audio.shape = [4, num_channels, num_samples]
```
Similar to the `torchvision.datasets` interface, an instance of the `Compose` or `ComposeMany` class can be supplied to `torchaudio` dataloaders that accept `transform=`.
## Optional
Install WavAugment for reverberation / pitch shifting:
```
pip install git+https://github.com/facebookresearch/WavAugment
```
# Cite
You can cite this work with the following BibTeX:
```
@misc{spijkervet_torchaudio_augmentations,
doi = {10.5281/ZENODO.4748582},
url = {https://zenodo.org/record/4748582},
author = {Spijkervet, Janne},
title = {Spijkervet/torchaudio-augmentations},
publisher = {Zenodo},
year = {2021},
copyright = {MIT License}
}
```