Xiaohui Zhang

@BriansIDP as we discussed offline, our team welcomes this feature (as a prototype feature first). So feel free to proceed with PRs at your convenience.

@BriansIDP can you share you progress here? Thanks!

the "log-prob as input" is being addressed by @carolineechen. @carolineechen if possible, when you start working on it, please give a more accurate ETA here. Thanks!

> > We may consider deprecating AxisMasking, TimeMasking and FrequencyMasking in future. > > I am okay with that, but in that case, instead of extending them, I'd start a...

sure! done.

@pzelasko here is the reason why we need #0 in G http://kaldi-asr.org/doc/graph.html#graph_disambig. @qindazhu thanks for your update, looking forward to your decoding example!

cc @hwangjeff who was interested in this as well

it'll be in the next release (around 3 months or so). it's available in nightly build right now.

thanks @rgt-yncrea . We wonder whether does it make sense to pass waveform composed of complex values to the spectrum centroid function? What's the use case? Thanks!

thanks @rgt-yncrea . We discussed among our team and thought it makes sense. Can you go ahead making an PR? Thanks!