Did you normalize signals when you calculate loss?

echocatzh / MTFAA-Net

Multi-Scale Temporal Frequency Convolutional Network With Axial Attention for Speech Enhancement

MIT License

187 stars 55 forks source link

Did you normalize signals when you calculate loss? #7

Closed YangangCao closed 2 years ago

YangangCao commented 2 years ago

Hi, thanks for you great work. I find that the loss decrease hardly when I train your MTFAA, I dont normalize signals when I calculate loss. Maybe I should normalize signals like 《Data augmentation and loss normalization for deep noise suppression》，I want to know your way to calculate loss.

echocatzh commented 2 years ago

Hi Yangang, thanks for your attention,

I'm not sure if normalized signals work, in the paper, the loss seems to be an extension of MSE.