Probably incorrect normalization?

[xwhjy]

Hi,
Thanks for open-sourcing wav2clip!
I noticed that you normalize the data (line 189) within the ResNet forward function. I am wondering if you want to normalize each sample or each batch. Because I find if the batch size is larger than 1, the encoded embedding would be different from that when extracted with batch size as 1.

I change line 189 into line 190 and the obtained embedding would be the same no matter what the batch size is.

I might miss something. Please let me know if my fix is correct or not. Thanks a lot!