biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Install pre-built xformers-0.0.32.post2 built with pt-2.9.0 (#27598)

Browse Source 
Signed-off-by: Huy Do <huydhn@gmail.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
This commit is contained in:
Huy Do
2025-10-28 19:39:15 -07:00
committed by GitHub
parent 0b51c9bd8b
commit f257544709
2 changed files with 1 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
docker/Dockerfile
View File

@@ -361,13 +361,6 @@ RUN --mount=type=bind,from=build,src=/workspace/dist,target=/vllm-workspace/dist
&& uv pip install --system dist/*.whl --verbose \
--extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')
# TODO (huydhn): Remove this once xformers is released for 2.9.0
RUN --mount=type=cache,target=/root/.cache/uv bash - <<'BASH'
. /etc/environment
export TORCH_CUDA_ARCH_LIST='7.5 8.0+PTX 9.0a'
uv pip install --system --no-build-isolation "git+https://github.com/facebookresearch/xformers@v0.0.32.post2"
BASH
# Install FlashInfer pre-compiled kernel cache and binaries
# https://docs.flashinfer.ai/installation.html
RUN --mount=type=cache,target=/root/.cache/uv \