thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 4

spec: Split model converter with other python scripts into a separate package

Browse source 
Signed-off-by: Vitaly Chikunov <vt@altlinux.org>
This commit is contained in:
Vitaly Chikunov 2025-09-06 08:03:38 +03:00
parent 98cd666eee
commit 7bd1b78853
1 changed files with 30 additions and 14 deletions
Download patch file Download diff file Expand all files Collapse all files

44
.gear/llama.cpp.spec
View file

@ -20,6 +20,7 @@ Group: Sciences/Computer science
Url: https://github.com/ggerganov/llama.cpp
ExcludeArch: %ix86
Requires: %name-cpu = %EVR
Requires: %name-convert = %EVR
%if_with cuda
Requires: %name-cuda = %EVR
%filter_from_requires /(libcudart\.so\.12)/d
@ -77,11 +78,7 @@ Multimodal models:
LLaVA 1.5 models, BakLLaVA, Obsidian, ShareGPT4V, MobileVLM 1.7B/3B
models, Yi-VL, Mini CPM, Moondream, Bunny, GLM-EDGE, Qwen2-VL
NOTE 1: For data format conversion script to work you will need to:
pip3 install -r /usr/share/llama.cpp/requirements.txt
NOTE 2:
NOTE:
MODELS ARE NOT PROVIDED. You'll need to download them from the original
sites (or Hugging Face Hub).
@ -106,13 +103,6 @@ Requires: libllama = %EVR
Summary: %name tools including backend for CPU
Group: Sciences/Computer science
Requires: libllama = %EVR
AutoReqProv: nopython3
Requires: python3
Requires: python3(argparse)
Requires: python3(glob)
Requires: python3(os)
Requires: python3(pip)
Requires: python3(struct)
%add_findreq_skiplist %_datadir/%name/examples/*
%description cpu
@ -135,6 +125,22 @@ Requires: %name-cpu = %EVR
%description vulkan
%summary.
%package convert
Summary: %name model converters to GGUF
Group: Sciences/Computer science
AutoReqProv: nopython3
Requires: python3(pip)
%description convert
The main model format conversion script is unsupported and provided AS IS.
Other scripts are legacy and unmaintained.
For the scripts to work you will need to:
pip3 install -r /usr/share/llama.cpp/requirements.txt
or install individually from the separate requirements files.
%prep
%setup
%autopatch -p1
@ -238,8 +244,11 @@ llama-cli -m %_datadir/tinyllamas/stories260K.gguf -p "Once upon a time" -s 55 -
%define _customdocdir %_docdir/%name
%doc LICENSE README.md docs build-options.txt
%_bindir/llama-*
%_bindir/convert*.py
%_datadir/%name
%dir %_datadir/%name
%dir %_datadir/%name/examples
%_datadir/%name/examples/*.sh
%_datadir/%name/grammars
%_datadir/%name/prompts
%dir %_libexecdir/llama
%_libexecdir/llama/libggml-cpu*.so
%_libexecdir/llama/libggml-rpc.so
@ -257,6 +266,13 @@ llama-cli -m %_datadir/tinyllamas/stories260K.gguf -p "Once upon a time" -s 55 -
%_libexecdir/llama/libggml-vulkan.so
%endif
%files convert
%_bindir/convert*.py
%dir %_datadir/%name
%dir %_datadir/%name/examples
%_datadir/%name/examples/*.py
%_datadir/%name/requirements*
%changelog
* Sat Aug 09 2025 Vitaly Chikunov <vt@altlinux.org> 1:6121-alt1
- Update to b6121 (2025-08-08).