It seems that gguf-py is not an independent code repository, but a part of llama.cpp.
In that sense you should try to incorporate the python binary package to src:llama.cpp and collaborate with Christian (@ckk). https://salsa.debian.org/deeplearning-team/llama.cpp/-/tree/master/gguf-py?ref_type=heads It might be a little bit complicated to produce both C++ binary packages and python packages from the same source, but you can take a look at OpenCV's example: https://salsa.debian.org/science-team/opencv/-/blob/master/debian/rules?ref_type=heads On Fri, 2025-07-04 at 13:50 +0000, Kohei Sendai wrote: > Package: wnpp > Severity: wishlist > Owner: Kohei Sendai <kouhei.sen...@gmail.com> > X-Debbugs-Cc: debian-de...@lists.debian.org, kouhei.sen...@gmail.com > > * Package name : python3-gguf > Version : 0.17.1 > Upstream Contact: GGML <g...@ggml.ai> > * URL : > * https://github.com/ggml-org/llama.cpp/tree/master/gguf-py > * License : Expat > Programming Lang: python > Description : Read and write ML models in GGUF format > > - This is important package to treat llm models.
