Brian McFee
Brian McFee
> What part of mir_eval is too slow? The separation module is particularly slow. Yes, it could benefit from a thorough efficiency audit -- I've noticed some pretty obvious examples...
Sure. Side note: is it worth making up a separate issue to put mir_eval in conda-forge?
Btw, numba now has wheels on pypi, so we probably could make it a hard dependency.
Update: llvmlite >=0.20 (now up) ships with windows wheels.
> Could you please add jit to those eval functions I'm not sure that jitting would help for bss eval. Last I checked, the bottleneck here was in computing a...
I see what you mean, but I think the example that Harte gives is inconsistent with the general semantics of added / suppressed notes. My understanding, which may diverge from...
I think that makes sense -- I'd like to get @ejhumphrey and @mattmcvicar 's take on things as well, since they've been at chords longer than I. Having thought about...
It's not common (it's not even possible right now :grin:) but I think there's no harm in exposing this parameter and preserving the default behavior. Down the road, I could...
Almost done implementing this, and ran into a couple of interesting points. I expected, since the PCPs are equivalent, that they should get perfect scores on all metrics. This turns...
> well then, i'm pleasantly surprised this works. lgtm so far > > grumble chords grumble The strange thing is that the test case hits on thirds, triads, and tetrads,...