build error for version 0.2.81 and 0.2.80

[charliboy]

Prerequisites

When I install via pip install llama-cpp-python, there will be an error. It will occur on versions 0.2.81 and 0.2.80, The version 0.2.79 can be successfully installed.

python 3.11.9 ubuntu 22.04

# Failure Information (for bugs) [26/26] : && /usr/bin/g++ -pthread -B /home/szh/vs/gpt/text-generation-webui/installer_files/conda/envs/inference/compiler_compat -O3 -DNDEBUG vendor/llama.cpp/examples/llava/CMakeFiles/llava.dir/llava.cpp.o vendor/llama.cpp/examples/llava/CMakeFiles/llava.dir/clip.cpp.o vendor/llama.cpp/examples/llava/CMakeFiles/llama-llava-cli.dir/llava-cli.cpp.o -o vendor/llama.cpp/examples/llava/llama-llava-cli -Wl,-rpath,/tmp/tmp2secpzim/build/vendor/llama.cpp/src:/tmp/tmp2secpzim/build/vendor/llama.cpp/ggml/src: vendor/llama.cpp/common/libcommon.a vendor/llama.cpp/src/libllama.so vendor/llama.cpp/ggml/src/libggml.so && : FAILED: vendor/llama.cpp/examples/llava/llama-llava-cli : && /usr/bin/g++ -pthread -B /home/szh/vs/gpt/text-generation-webui/installer_files/conda/envs/inference/compiler_compat -O3 -DNDEBUG vendor/llama.cpp/examples/llava/CMakeFiles/llava.dir/llava.cpp.o vendor/llama.cpp/examples/llava/CMakeFiles/llava.dir/clip.cpp.o vendor/llama.cpp/examples/llava/CMakeFiles/llama-llava-cli.dir/llava-cli.cpp.o -o vendor/llama.cpp/examples/llava/llama-llava-cli -Wl,-rpath,/tmp/tmp2secpzim/build/vendor/llama.cpp/src:/tmp/tmp2secpzim/build/vendor/llama.cpp/ggml/src: vendor/llama.cpp/common/libcommon.a vendor/llama.cpp/src/libllama.so vendor/llama.cpp/ggml/src/libggml.so && : /home/szh/vs/gpt/text-generation-webui/installer_files/conda/envs/inference/compiler_compat/ld: warning: libgomp.so.1, needed by vendor/llama.cpp/ggml/src/libggml.so, not found (try using -rpath or -rpath-link) /home/szh/vs/gpt/text-generation-webui/installer_files/conda/envs/inference/compiler_compat/ld: vendor/llama.cpp/ggml/src/libggml.so: undefined reference to GOMP_barrier@GOMP_1.0' /home/szh/vs/gpt/text-generation-webui/installer_files/conda/envs/inference/compiler_compat/ld: vendor/llama.cpp/ggml/src/libggml.so: undefined reference to GOMP_parallel@GOMP_4.0' /home/szh/vs/gpt/text-generation-webui/installer_files/conda/envs/inference/compiler_compat/ld: vendor/llama.cpp/ggml/src/libggml.so: undefined reference to omp_get_thread_num@OMP_1.0' /home/szh/vs/gpt/text-generation-webui/installer_files/conda/envs/inference/compiler_compat/ld: vendor/llama.cpp/ggml/src/libggml.so: undefined reference to GOMP_single_start@GOMP_1.0' /home/szh/vs/gpt/text-generation-webui/installer_files/conda/envs/inference/compiler_compat/ld: vendor/llama.cpp/ggml/src/libggml.so: undefined reference to omp_get_num_threads@OMP_1.0' collect2: error: ld returned 1 exit status ninja: build stopped: subcommand failed.

  *** CMake build failed

[Rybens92]

Disabling llava module fixes this on my end:

CMAKE_ARGS="-DLLAVA_BUILD=OFF" pip install -U llama-cpp-python

I am using Ubuntu through WSL2 on Windows 11.

[justme1135]

Disabling llava module fixes this on my end:

CMAKE_ARGS="-DLLAVA_BUILD=OFF" pip install -U llama-cpp-python

I am using Ubuntu through WSL2 on Windows 11.

This command did not work for me on Arch Linux and produced the same error as above

set CMAKE_ARGS="-DLLAMA_CUDA=on -DLLAVA_BUILD=off" && set FORCE_CMAKE=1 && pip install llama-cpp-python --no-cache-dir

[nmstoker]

Thanks @Rybens92 - that worked for me when I combined it with my existing settings, like so, on Arch Linux under a conda env with Python 3.11:

CMAKE_ARGS="-DGGML_CUDA=on -DLLAVA_BUILD=off" pip install llama-cpp-python --upgrade --force-reinstall

---

I get that this particular issue will need a change somewhere to resolve it, but independently I think the README could do with an update to point people away from LLAMA_CUBLAS and toward GGML_CUDA.

[LEON-REIN]

Solved for me:

CMAKE_ARGS="-DCMAKE_CXX_FLAGS=-fopenmp" pip install llama-cpp-python

[charliboy]

Disabling llava module fixes this on my end:

CMAKE_ARGS="-DLLAVA_BUILD=OFF" pip install -U llama-cpp-python

I am using Ubuntu through WSL2 on Windows 11.

thanks ,it's worked for me

[javi22020]

Disabling llava module fixes this on my end:

CMAKE_ARGS="-DLLAVA_BUILD=OFF" pip install -U llama-cpp-python

I am using Ubuntu through WSL2 on Windows 11.

Worked for me too

[yimuu]

Disabling llava module fixes this on my end:

CMAKE_ARGS="-DLLAVA_BUILD=OFF" pip install -U llama-cpp-python

I am using Ubuntu through WSL2 on Windows 11.

Is there any other solution? Because I am going to use llava

[ianozsvald]

Super useful, this worked for me CMAKE_ARGS="-DGGML_CUDA=on -DLLAVA_BUILD=off" pip install llama-cpp-python --upgrade --force-reinstall on Linux Mint Cinnamon, thanks

[victorolaiya]

Solved for me:

CMAKE_ARGS="-DCMAKE_CXX_FLAGS=-fopenmp" pip install llama-cpp-python

Solved it for me!

[kpm]

What does disabling LLAVA_BUILD do? Do we lose any functionality?

[javi22020]

What does disabling LLAVA_BUILD do? Do we lose any functionality?

Probably Llava compatibility, they're VLMs.