iCEM
mkolodziejczyk-piap opened this issue · comments
mkolodziejczyk-piap commented
Hi, I'd like to contribute to iCEM extensions of current CEM implementation. I'd like to ask:
- if it's not on your list?
- if is it ok, if I extend current cem() function with additional (optional) features and hyperparameters?
Regards,
mkolodziejczyk-piap commented
BTW In current implementation correlation in time axis is implemented through noise "sharing". In iCEM colored noise is suggested. Are you aware of any comparison of those? I also wonder why sampling_smoothing is zero in default hyperparameters set? Regards,