Add llama-quantize utility

diff --git a/debian/control b/debian/control
index ca3098a24ea092a275a4118ba8421485461d2dfe..530f8f4c471893b6d16e0ce96b766ee97022b18f 100644 (file)
--- a/debian/control
+++ b/debian/control
@@ -31,8 +31,16 @@ Architecture: any
 Priority: optional
 Depends: ${misc:Depends}, ${shlibs:Depends},
  libllama, ggml, curl, openssl
-Description: Inference of LLMs in pure C/C++ (CLI)
- Llama.cpp inference of LLMs in pure C/C++ (CLI).
+Description: Inference of LLMs in pure C/C++ (server)
+ Llama.cpp inference of LLMs in pure C/C++ (server).
+
+Package: llama-cpp-quantize
+Architecture: any
+Priority: optional
+Depends: ${misc:Depends}, ${shlibs:Depends},
+ libllama, ggml
+Description: Inference of LLMs in pure C/C++ (quantize)
+ Llama.cpp inference of LLMs in pure C/C++ (quantize).
 
 Package: libllama-dev
 Architecture: any

diff --git a/debian/llama-cpp-quantize.install b/debian/llama-cpp-quantize.install
new file mode 100644 (file)
index 0000000..d72d696
--- /dev/null
+++ b/debian/llama-cpp-quantize.install
@@ -0,0 +1,2 @@
+/usr/bin/llama-quantize
+/usr/libexec/*/ggml/llama-quantize