A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 05:30 UTC
Upgrade FlashInfer to v0.2.11 (#22613)
Signed-off-by: Po-Han Huang <pohanh@nvidia.com> Co-authored-by: mgoin <mgoin64@gmail.com>
Po-Han Huang (NVIDIA) committed
Aug 12, 2025 at 10:58 UTC
dc5e4a653c859573dfcca99f1b0141c2db9f94cc
2 files changed
+2
-2
docker/Dockerfile
+1
-1
@@ -387,7 +387,7 @@ RUN --mount=type=bind,from=build,src=/workspace/dist,target=/vllm-workspace/dist
387
ARG FLASHINFER_GIT_REPO="https://github.com/flashinfer-ai/flashinfer.git"
388
# Keep this in sync with https://github.com/vllm-project/vllm/blob/main/requirements/cuda.txt
389
# We use `--force-reinstall --no-deps` to avoid issues with the existing FlashInfer wheel.
390
-ARG FLASHINFER_GIT_REF="v0.2.10"
390
+ARG FLASHINFER_GIT_REF="v0.2.11"
391
RUN --mount=type=cache,target=/root/.cache/uv bash - <<'BASH'
392
. /etc/environment
393
git clone --depth 1 --recursive --shallow-submodules \
setup.py
+1
-1
@@ -684,7 +684,7 @@ setup(
684
"mistral_common[audio]"], # Required for audio processing
685
"video": [], # Kept for backwards compatibility
686
# FlashInfer should be updated together with the Dockerfile
687
- "flashinfer": ["flashinfer-python==0.2.10"],
687
+ "flashinfer": ["flashinfer-python==0.2.11"],
688
},
689
cmdclass=cmdclass,
690
package_data=package_data,