Translate Zero-Latency Whitening to Torch

[wbenoit26]

The zlw package has a nice implementation of a zero-latency whitening filter (as well as the standard linear-phase whitener). We should translate the functionality to ml4gw to improve our online latency.

[EthanMarx]

Cool. It looks like you can just implement a torch based backend: https://git.ligo.org/james.kennington/zlw/-/blob/main/src/zlw/fourier.py?ref_type=heads

[wbenoit26]

There are other numpy operations involved in calculating the whitening kernel, so I think it'll take more than that.