Loss decrease for a few and sudden explodes to inf

[Angus Xu]

Hi Nils, i encounted a problem while training the DTLN-aec, when i train only on the real data it stable at the loss 0.03.
However when i trained on both real and synthetic data, the loss derease for some batch and explode to inf at a sudden.
i thought it is because the log in the snr. Whether or not it is normal? Or will the loss to be stable at a finite number after several epochs.

[MuhammadurRehman19058]

@anguschowchowxu, I have been encountering the same issue. Have you come up with its solution?

[Angus Xu]

@MuhammadurRehman19058 In synthetic near_end speech data, some of them with no noise will be an exact zeros-array, it raise inf when computing SNR