Wrong output shape in enhancement.py

sp-uhh / sgmse

Score-based Generative Models (Diffusion Models) for Speech Enhancement and Dereverberation

MIT License

520 stars 76 forks source link

Closed PavelPanjaya closed 7 months ago

PavelPanjaya commented 7 months ago

The shape of the output tensor x_hat is torch.Size([2, 758949]) which can't be written as audio file because of the first dimension.

cobalamin commented 7 months ago

The model is designed for single-channel speech enhancement; you may be providing a stereo input file?

PavelPanjaya commented 7 months ago

Thanks for your response, I've converted my audio to mono and it works now. Thanks.

cobalamin commented 7 months ago

Great! Closing this issue.