huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-11 04:14:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix Dockerfile

Browse Source 
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
This commit is contained in:
Adrien Gallouët 2025-02-07 12:26:28 +01:00
parent 0d27ee74de
commit 4841f71a0e
No known key found for this signature in database
2 changed files with 9 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
Dockerfile_llamacpp
View File

@ -3,7 +3,6 @@ FROM nvidia/cuda:12.8.0-cudnn-devel-ubuntu24.04 AS deps
ARG llamacpp_version=b4651
ARG llamacpp_cuda=OFF
ARG cuda_arch=75-real;80-real;86-real;89-real;90-real
ENV TGI_LLAMA_PKG_CUDA=cuda-${CUDA_VERSION%.*}
WORKDIR /opt/src

15
backends/llamacpp/build.rs
View File

@ -31,10 +31,18 @@ impl ParseCallbacks for PrefixStripper {
}
fn main() {
let pkg_cuda = option_env!("TGI_LLAMA_PKG_CUDA");
let lib_search_path = option_env!("TGI_LLAMA_LD_LIBRARY_PATH");
let lib_target_hardware = option_env!("TGI_LLAMA_HARDWARE_TARGET").unwrap_or("cpu");
if let Some(cuda_version) = option_env!("CUDA_VERSION") {
let mut version: Vec<&str> = cuda_version.split('.').collect();
if version.len() > 2 {
version.pop();
}
pkg_config::Config::new().probe(&version.join(".")).unwrap();
}
pkg_config::Config::new().probe("llama").unwrap();
let bindings = bindgen::Builder::default()
.header("src/wrapper.h")
.prepend_enum_name(false)
@ -48,10 +56,5 @@ fn main() {
.write_to_file(out_path.join("llamacpp.rs"))
.expect("Couldn't write bindings!");
if let Some(pkg_cuda) = pkg_cuda {
pkg_config::Config::new().probe(pkg_cuda).unwrap();
}
pkg_config::Config::new().probe("llama").unwrap();
inject_transient_dependencies(lib_search_path, lib_target_hardware);
}