Package Details: spleeter 1.4.0-1

Git Clone URL: https://aur.archlinux.org/spleeter.git (read-only, click to copy)
Package Base: spleeter
Description: Deezer source separation library including pretrained models
Upstream URL: https://github.com/deezer/spleeter
Licenses: MIT
Submitter: fordprefect
Maintainer: fordprefect
Last Packager: fordprefect
Votes: 0
Popularity: 0.000000
First Submitted: 2019-11-14 08:38
Last Updated: 2019-11-14 11:35

Latest Comments

Nover commented on 2019-12-03 13:46

I tried to install it on my work computer and it broked Xorg because of all the Nvidia drivers installed at the same time (I use an AMD GPU). I raged, did a small pacman -Rcns mhdw or something like that and it worked again. It really needs to be repackaged :/ I think there is a better way to do it as my coworkers installed it from the git without all the hassle. Good luck

fordprefect commented on 2019-11-18 09:22

I have taken some time to look into packaging this - and it is a pain. Not only does tensorflow114 not properly build, also building it from ALA does not reliably work. And then it needs more versioned dependencies. A last attempt will be made (when time is available) to unversion the dependencies and test if that works.

pha-qu commented on 2019-11-16 19:09

A pre-built binary release would be great, otherwise installing Cuda as a build dependency is a deal breaker @ ~2 Gigs with a 4 Gig footprint... Including the unnecessary Nvidia Eclipse based IDE for GPU programming, and the huge amount of Nvidia documentation that includes

fordprefect commented on 2019-11-14 13:13

Since I'm currenlty away from my arch machine I couldn't test - will be done once I'm back on it. one should also brush one's teeth twice a day… but you are absolutely right!

sekret commented on 2019-11-14 13:11

Thanks! I cannot build it right now, because tensorflow114 requires more storage space than I can spare right now lol

The question is: Does it work for you? Did it work for you before? You should always test your packages before uploading them here ;-)

fordprefect commented on 2019-11-14 12:23

I packed and uploaded tensorflow-1.14 (as tensorflow114), as it has been in the official repos a while ago - but couldn't test it. Also added it as a dependency. Does it work for you now?

sekret commented on 2019-11-14 10:25

https://pastebin.com/C52tgBJs

Your package is still missing tensorflow, of which version 1.14.0 is required. Unfortunately we have version 2.* in the repos.