ExternalVendorCode/LocalAI
1
0
Fork 0
mirror of https://github.com/mudler/LocalAI.git synced 2024-12-18 12:26:26 +00:00
Code Issues Actions 12 Packages Projects Releases Wiki Activity

fix(llama.cpp): embed metal file into result binary for darwin (#4279)

Browse Source 
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
This commit is contained in:
Ettore Di Giacinto 2024-11-28 05:17:00 +01:00 committed by GitHub
parent 0597f3b9e9
commit cbedf2f428
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 1 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
Makefile
View File

@ -745,10 +745,6 @@ backend-assets/grpc/llama-cpp-fallback: backend-assets/grpc backend/cpp/llama/ll
$(info ${GREEN}I llama-cpp build info:fallback${RESET})
CMAKE_ARGS="$(CMAKE_ARGS) -DGGML_AVX=off -DGGML_AVX2=off -DGGML_AVX512=off -DGGML_FMA=off -DGGML_F16C=off" $(MAKE) VARIANT="llama-fallback" build-llama-cpp-grpc-server
cp -rfv backend/cpp/llama-fallback/grpc-server backend-assets/grpc/llama-cpp-fallback
# TODO: every binary should have its own folder instead, so can have different metal implementations
ifeq ($(BUILD_TYPE),metal)
cp backend/cpp/llama-fallback/llama.cpp/build/bin/ggml-metal.metal backend-assets/grpc/
endif
backend-assets/grpc/llama-cpp-cuda: backend-assets/grpc backend/cpp/llama/llama.cpp
cp -rf backend/cpp/llama backend/cpp/llama-cuda

4
backend/cpp/llama/Makefile
View File

@ -30,9 +30,7 @@ else ifeq ($(OS),Darwin)
CMAKE_ARGS+=-DGGML_METAL=OFF
else
CMAKE_ARGS+=-DGGML_METAL=ON
# Until this is tested properly, we disable embedded metal file
# as we already embed it as part of the LocalAI assets
CMAKE_ARGS+=-DGGML_METAL_EMBED_LIBRARY=OFF
CMAKE_ARGS+=-DGGML_METAL_EMBED_LIBRARY=ON
TARGET+=--target ggml-metal
endif
endif