huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-06-19 15:52:08 +00:00
Code Issues Packages Projects Releases Wiki Activity

Set TGI_LLAMA_PKG_CUDA from CUDA_VERSION

Browse Source 
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
This commit is contained in:
Adrien Gallouët 2025-02-05 10:57:50 +00:00
parent dbee804129
commit c52f08351f
No known key found for this signature in database
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
Dockerfile_llamacpp
View File

@ -2,6 +2,7 @@ FROM nvidia/cuda:12.6.3-cudnn-devel-ubuntu24.04 AS deps
ARG llama_version=b4628 ARG llama_version=b4628
ARG llama_cuda_arch=75-real;80-real;86-real;89-real;90-real ARG llama_cuda_arch=75-real;80-real;86-real;89-real;90-real
ENV TGI_LLAMA_PKG_CUDA=cuda-${CUDA_VERSION%.*}
WORKDIR /opt/src WORKDIR /opt/src
@ -50,7 +51,6 @@ RUN cargo chef cook \
--profile release-opt \ --profile release-opt \
--package text-generation-router-llamacpp --package text-generation-router-llamacpp
COPY . . COPY . .
ENV TGI_LLAMA_PKG_CUDA=cuda-12.6
RUN cargo build \ RUN cargo build \
--profile release-opt \ --profile release-opt \
--package text-generation-router-llamacpp --frozen --package text-generation-router-llamacpp --frozen