project: switch submodule dependency from ggml to llama.cpp (#16)

* currently not using anything from llama.cpp, but all development of ggml happens there, and it's easier to pull/upstream ggml changes to llama.cpp directly
* may switch back to ggml-org/ggml once it's maintained more independently rather than just synced occasionally

Author: Acly

.gitmodules

@@ -1,3 +1,3 @@
-[submodule "depend/ggml"]
-	path = depend/ggml
-	url = https://github.com/Acly/ggml.git
+[submodule "depend/llama"]
+	path = depend/llama
+	url = https://github.com/Acly/llama.cpp.git

CMakeLists.txt

@@ -109,7 +109,7 @@ if(VISP_CI)
     endforeach()
   endif()
 endif()
-add_subdirectory(depend/ggml)
+add_subdirectory(depend/llama/ggml)
 
 # Vision.cpp libraries and executables