rebase: target the official vLLM 0.31.0 release

Use upstream release images at db9527a46873, preserve custom processor conditioning, and include weekly jobs in full Buildkite release validation.

Signed-off-by: tzhouam <tzhouam@connect.ust.hk>
This commit is contained in:
tzhouam committed 2026-10-03 21:11:54 +08:00
1 parent 008fa25263
commit ed876de879
15 files changed
+142 -45

No files matched your search

+4 -3
View File
@@ -1,7 +1,8 @@
# Keep the ROCm runtime aligned with the vLLM release targeted by
# docker/Dockerfile.ci. A stale base image can build successfully while failing
# at runtime when vLLM-Omni imports APIs added by the newer vLLM release.
ARG BASE_IMAGE=vllm/vllm-openai-rocm:v0.30.0
# Official vLLM v0.31.0 release image, before Docker Hub tag promotion.
ARG BASE_IMAGE=public.ecr.aws/q9t5s3a7/vllm-release-repo:db9527a46873454610df6dbedf79a36d6bf1a7f6-rocm
# Keep the dependency-heavy portion independent from the source tree. The
# registry cache is keyed on these inputs, while the final source copy can be
@@ -12,11 +13,11 @@ FROM ${BASE_IMAGE} AS deps
# BASE_IMAGE pins the default prebuilt runtime when USE_NIGHTLY_BUILD=0.
# VLLM_VERSION_OR_COMMIT_HASH is consumed only by the opt-in source rebuild when
# USE_NIGHTLY_BUILD=1. Keep both aligned with docker/Dockerfile.ci's
# VLLM_BASE_TAG. This is enforced by tests/buildkite/test_rocm_dockerfile.py. An
# release version. This is enforced by tests/buildkite/test_rocm_dockerfile.py. An
# intentional custom source pin must update that regression in the same change,
# and should return to the stable release as soon as one is available.
ARG USE_NIGHTLY_BUILD=0
ARG VLLM_VERSION_OR_COMMIT_HASH=v0.30.0
ARG VLLM_VERSION_OR_COMMIT_HASH=v0.31.0
ARG ARG_PYTORCH_ROCM_ARCH
ENV PYTORCH_ROCM_ARCH=${ARG_PYTORCH_ROCM_ARCH:-${PYTORCH_ROCM_ARCH}}