Search Criteria
Package Details: whisper.cpp-vulkan 1.7.1-4
Package Actions
Git Clone URL: | https://aur.archlinux.org/whisper.cpp.git (read-only, click to copy) |
---|---|
Package Base: | whisper.cpp |
Description: | Port of OpenAI's Whisper model in C/C++ (with Vulkan optimizations) |
Upstream URL: | https://github.com/ggerganov/whisper.cpp |
Licenses: | MIT |
Conflicts: | whisper.cpp |
Provides: | whisper.cpp |
Submitter: | robertfoster |
Maintainer: | robertfoster |
Last Packager: | robertfoster |
Votes: | 11 |
Popularity: | 0.90 |
First Submitted: | 2023-03-10 17:32 (UTC) |
Last Updated: | 2024-11-04 13:56 (UTC) |
Dependencies (10)
- gcc-libs (gcc-libs-gitAUR, gccrs-libs-gitAUR, gcc11-libsAUR, gcc-libs-snapshotAUR)
- glibc (glibc-gitAUR, glibc-linux4AUR, glibc-eacAUR, glibc-eac-binAUR, glibc-eac-rocoAUR)
- vulkan-driver (nvidia-410xx-utilsAUR, nvidia-440xx-utilsAUR, nvidia-430xx-utilsAUR, swiftshader-gitAUR, amdvlk-debugAUR, nvidia-vulkan-utilsAUR, amdvlk-2023q3.3AUR, amdvlk-2021q2.5AUR, vulkan-amdgpu-proAUR, nvidia-390xx-utilsAUR, nvidia-470xx-utilsAUR, amdvlk-gitAUR, nvidia-utils-betaAUR, vulkan-nouveau-gitAUR, mesa-minimal-gitAUR, mesa-gitAUR, amdonly-gaming-vulkan-radeon-gitAUR, amdonly-gaming-vulkan-swrast-gitAUR, vulkan-amdgpu-pro-legacyAUR, vulkan-radeon-amd-bc250AUR, nvidia-utils-teslaAUR, amdvlk-binAUR, mesa-wsl2-gitAUR, nvidia-535xx-utilsAUR, nvidia-525xx-utilsAUR, nvidia-510xx-utilsAUR, nvidia-550xx-utilsAUR, amdvlk, nvidia-utils, vulkan-intel, vulkan-nouveau, vulkan-radeon, vulkan-swrast, vulkan-virtio)
- vulkan-icd-loader (vulkan-icd-loader-gitAUR)
- blas-openblas (make)
- cmake (cmake-gitAUR) (make)
- cuda (cuda11.1AUR, cuda-12.2AUR, cuda12.0AUR, cuda11.4AUR, cuda11.4-versionedAUR, cuda12.0-versionedAUR) (make)
- git (git-gitAUR, git-glAUR) (make)
- openvinoAUR (openvino-gitAUR) (make)
- vulkan-icd-loader (vulkan-icd-loader-gitAUR) (make)
Required by (1)
- shisper-git (requires whisper.cpp)
Latest Comments
1 2 3 Next › Last »
Marzal commented on 2024-11-04 13:37 (UTC) (edited on 2024-11-04 13:44 (UTC) by Marzal)
Hi, there is copy/paste errata on package_whisper.cpp-vulkan()
cd "${pkgbase}-openvino"
Also did you check
-DCMAKE_SKIP_RPATH=YES
? fixes a lot of warnings on my compilations.My
/etc/makepkg.conf
looks like upstream https://gitlab.archlinux.org/archlinux/packaging/packages/pacman/-/blob/main/makepkg.conf?ref_type=heads#L98 I think Arch enableddebug
a while ago (1 year) https://gitlab.archlinux.org/archlinux/packaging/packages/pacman/-/commit/90bf367e61b4f77f8351d0412be3d0c4ddadb85arobertfoster commented on 2024-11-04 11:58 (UTC) (edited on 2024-11-04 12:12 (UTC) by robertfoster)
@Marzal clblas removed and vulkan added. (!debug) option is already the default (check /etc/makepkg.conf).
@sixtyfive, here there are no models, it's up to you.
@BrainDamage. I don't agree with you. You can always manually edit the PKGBUILD, skip the makedepends or use the chaotic aur repo to fetch binaries. AUR is not a personal repo, is public
BrainDamage commented on 2024-10-23 09:02 (UTC) (edited on 2024-10-23 09:02 (UTC) by BrainDamage)
I'd strongly suggest to split the package into separate files, there's nothing in their compilation steps in common apart from downloading <1MB sources and in order to do that, you're making impossible to programmatically select the packages and forcing everyone to either:
install the 5GB/10GB of cuda/openvino make dependencies irregardless if they need the relevant packages
edit the pkbuild every update cutting out the unneeded packages, operation further complicated by the unnecessary pkgbuild splitting functionality plastering the sections all around the code
Official arch packages always build with all variants because they end up in a public repository, so it doesn't make sense for them to even consider selecting a single one so they often adopt a split system for maintainer's ease. Here we're talking about personal repos and I'd argue that being able to save downloading 5/10/15 GB of useless makedeps would be a worthy feature
You'd have to edit the version number/hash in 3 files instead of one, but in exchange the pkgbuild themselves would be much cleaner
sixtyfive commented on 2024-10-21 16:09 (UTC) (edited on 2024-10-21 16:10 (UTC) by sixtyfive)
With this package, (1) how does one access "command"? (as in
./command -m ./models/ggml-small.en.bin -t 8
) and (2) where in the filesystem are the models stored? Thank you kindly!homocomputeris commented on 2024-10-11 16:59 (UTC) (edited on 2024-10-11 17:05 (UTC) by homocomputeris)
The flag for openblas has changed:
Marzal commented on 2024-10-11 16:38 (UTC) (edited on 2024-10-19 22:17 (UTC) by Marzal)
CLBLAST it's no longer supported on 1.7 https://github.com/ggerganov/whisper.cpp/issues/2292
But Vulkan it's faster on my RX580
I would also add:
OPTIONS+=(!debug)
after makdepends or sourceand
-DCMAKE_SKIP_RPATH=YES
in build()PD: +1 to NOT have this packages as a split package, compiling cuda + openvino is overkill for my computer. But thanks for your work in any case. https://bbs.archlinux.org/viewtopic.php?id=223171
homocomputeris commented on 2024-10-10 23:12 (UTC)
Where are models supposed to be placed?
dreieck commented on 2024-10-10 11:12 (UTC) (edited on 2024-10-13 10:12 (UTC) by dreieck)
Your packages
llama.cpp-vulkan-git
andwhisper.cpp-clblas
do conflict which each other, which is not reflected in theconflicts
array.Please add the corresponding
conflicts
entry, or, to ensure compatibility, think of stripping out thelibggml
stuff and depend onlibggml
:Regards!
dreieck commented on 2024-10-10 11:06 (UTC)
You need to install the license file:
Regards and thanks for maintaining!
homocomputeris commented on 2024-10-07 22:17 (UTC)
Can the CUDA dependency be dropped somehow for Intel?
1 2 3 Next › Last »