Ryan Hileman

> Better than that however would be to store an initial hidden state that's updated via gradients during model training. I might try this, so to make sure I'm understanding...

ideally these are also tested on apple silicon, to see how it behaves with the unified gpu memory

what's the easiest way to repro your test?

I'm very interested in logging and labeling realtime audio when tracing [Talon](https://talonvoice.com) with Rerun! I'll note that Talon's audio is realtime/continuous/infinite, but it might make more sense efficiency wise to...