From a987adfcc6ce1bc199d2640dd6aacdd999c97a95 Mon Sep 17 00:00:00 2001 From: Florent BENOIT Date: Fri, 3 May 2024 15:51:03 +0200 Subject: [PATCH] chore: handle some AVX flags properly on certain Intel CPUs (#459) Signed-off-by: Florent Benoit --- model_servers/llamacpp_python/cuda/Containerfile | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/model_servers/llamacpp_python/cuda/Containerfile b/model_servers/llamacpp_python/cuda/Containerfile index 8c116dae9..92875a967 100644 --- a/model_servers/llamacpp_python/cuda/Containerfile +++ b/model_servers/llamacpp_python/cuda/Containerfile @@ -1,7 +1,10 @@ FROM quay.io/opendatahub/workbench-images:cuda-ubi9-python-3.9-20231206 +USER root +RUN dnf install -y gcc-toolset-13-gcc gcc-toolset-13-gcc-c++ +USER 1001 WORKDIR /locallm COPY src . ENV CMAKE_ARGS="-DLLAMA_CUBLAS=on -DLLAMA_AVX2=OFF -DLLAMA_FMA=OFF -DLLAMA_F16C=OFF" ENV FORCE_CMAKE=1 -RUN pip install --no-cache-dir -r ./requirements.txt +RUN CC="/opt/rh/gcc-toolset-13/root/usr/bin/gcc" CXX="/opt/rh/gcc-toolset-13/root/usr/bin/g++" pip install --no-cache-dir -r ./requirements.txt ENTRYPOINT [ "sh", "run.sh" ]