mirror/LocalAI
1
0
Fork 0
mirror of https://github.com/mudler/LocalAI.git synced 2026-06-25 00:59:28 -04:00
Code Issues Packages Projects Releases Wiki Activity

fix(vllm): fetch the vllm-metal wheel without the GitHub API

Browse Source 
The darwin build resolved the wheel URL via api.github.com, whose
unauthenticated rate limit (60/hr per IP) 403s on shared macOS runners
(observed after the 9-min vLLM source build). Construct the release-asset
download URL deterministically from the pinned tag and the cp312/arm64 wheel
name instead - no API call, no rate limit. Verified the URL resolves (200).

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Assisted-by: Claude:opus-4.8 [Claude Code]
This commit is contained in:
Ettore Di Giacinto
2026-06-24 20:54:30 +00:00
parent af7d0e8b40
commit bfb9a40d58
1 changed files with 8 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
backend/python/vllm/install.sh
View File

@@ -135,19 +135,14 @@ if [ "$(uname -s)" = "Darwin" ]; then
popd
popd
# 2) Install the prebuilt vllm-metal wheel from the PINNED release
# (${VLLM_METAL_VERSION}). It pulls mlx / mlx-metal as deps and registers
# the `metal` platform plugin that backend.py resolves to at engine-init
# time. Pinning the tag (vs releases/latest) keeps the wheel and the vLLM
# source build above reproducible and coupled; .github/bump_vllm_metal.sh
# advances both together.
_metal_wheel_url=$(curl -fsSL "https://api.github.com/repos/vllm-project/vllm-metal/releases/tags/${VLLM_METAL_VERSION}" \
| grep -oE '"browser_download_url"[[:space:]]*:[[:space:]]*"[^"]+\.whl"' \
| head -n1 | sed -E 's/.*"(https[^"]+)".*/\1/')
if [ -z "${_metal_wheel_url}" ]; then
echo "ERROR: could not resolve a vllm-metal wheel URL for release ${VLLM_METAL_VERSION}" >&2
exit 1
fi
# 2) Install the prebuilt vllm-metal wheel for the PINNED release. It pulls
# mlx / mlx-metal as deps and registers the `metal` platform plugin that
# backend.py resolves to at engine-init time. Build the release-asset URL
# deterministically (tag + the cp312/arm64 wheel name) rather than querying
# api.github.com, whose unauthenticated rate limit (60/hr per IP) 403s on
# shared CI runners. The wheel version is the tag without its leading 'v'.
_metal_wheel="vllm_metal-${VLLM_METAL_VERSION#v}-cp312-cp312-macosx_11_0_arm64.whl"
_metal_wheel_url="https://github.com/vllm-project/vllm-metal/releases/download/${VLLM_METAL_VERSION}/${_metal_wheel}"
echo "Installing vllm-metal wheel: ${_metal_wheel_url}"
uv pip install "${_metal_wheel_url}"