diff --git a/Dockerfile b/Dockerfile
index 5484be5bc578..620f549cf395 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -82,6 +82,7 @@ ENV BUILDKITE_COMMIT=${buildkite_commit}
 ARG USE_SCCACHE
 ARG SCCACHE_BUCKET_NAME=vllm-build-sccache
 ARG SCCACHE_REGION_NAME=us-west-2
+ARG SCCACHE_S3_NO_CREDENTIALS=0
 # if USE_SCCACHE is set, use sccache to speed up compilation
 RUN --mount=type=cache,target=/root/.cache/pip \
     if [ "$USE_SCCACHE" = "1" ]; then \
@@ -92,6 +93,7 @@ RUN --mount=type=cache,target=/root/.cache/pip \
         && rm -rf sccache.tar.gz sccache-v0.8.1-x86_64-unknown-linux-musl \
         && export SCCACHE_BUCKET=${SCCACHE_BUCKET_NAME} \
         && export SCCACHE_REGION=${SCCACHE_REGION_NAME} \
+        && export SCCACHE_S3_NO_CREDENTIALS=${SCCACHE_S3_NO_CREDENTIALS} \
         && export SCCACHE_IDLE_TIMEOUT=0 \
         && export CMAKE_BUILD_TYPE=Release \
         && sccache --show-stats \
diff --git a/setup.py b/setup.py
index 8930ea7239dc..7da911544043 100644
--- a/setup.py
+++ b/setup.py
@@ -371,7 +371,9 @@ def get_vllm_version() -> str:
         cuda_version = str(get_nvcc_cuda_version())
         if cuda_version != MAIN_CUDA_VERSION:
             cuda_version_str = cuda_version.replace(".", "")[:3]
-            version += f"+cu{cuda_version_str}"
+            # skip this for source tarball, required for pypi
+            if "sdist" not in sys.argv:
+                version += f"+cu{cuda_version_str}"
     elif _is_hip():
         # Get the HIP version
         hipcc_version = get_hipcc_rocm_version()