How to compose torchaudio transforms

Hi
Is there an integrated solution to compose torchaudio transforms in the same style as torchvision do?

torchaudio doesn’t provide a dedicated compose transformation since 0.3.0 (see release notes). Instead, one can simply apply them one after the other x = transform1(x); x = transform2(x), or use nn.Sequential(transform1, transform2).

2 Likes

Is there any specific reason to why there is no compose transformation for torchaudio anymore?