From e3ab23c1a688fee22f4fb341c14dfa2b31d64fb9 Mon Sep 17 00:00:00 2001 From: fzyzcjy <5236035+fzyzcjy@users.noreply.github.com> Date: Thu, 4 Dec 2025 16:41:24 +0800 Subject: [PATCH] Try to fix B200 DeepEP error (#14399) --- docker/Dockerfile | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docker/Dockerfile b/docker/Dockerfile index c6348364c..c3d7a7d21 100644 --- a/docker/Dockerfile +++ b/docker/Dockerfile @@ -138,7 +138,7 @@ RUN --mount=type=cache,target=/root/.cache/pip python3 -m pip install --upgrade # We use Tom's DeepEP fork for GB200 for now; the 1fd57b0276311d035d16176bb0076426166e52f3 commit is https://github.com/fzyzcjy/DeepEP/tree/gb200_blog_part_2 RUN set -eux; \ if [ "${CUDA_VERSION%%.*}" = "12" ]; then \ - pip install nvidia-nvshmem-cu12==${NVSHMEM_VERSION} ; \ + pip install nvidia-nvshmem-cu12==3.3.20 ; \ elif [ "${CUDA_VERSION%%.*}" = "13" ]; then \ pip install nvidia-nvshmem-cu13==${NVSHMEM_VERSION} ; \ else \ @@ -203,7 +203,7 @@ RUN --mount=type=cache,target=/root/.cache/pip python3 -m pip install \ RUN --mount=type=cache,target=/root/.cache/pip if [ "${CUDA_VERSION%%.*}" = "12" ]; then \ python3 -m pip install nvidia-nccl-cu12==2.28.3 --force-reinstall --no-deps ; \ python3 -m pip install nvidia-cudnn-cu12==9.16.0.29 --force-reinstall --no-deps; \ - python3 -m pip install nvidia-nvshmem-cu12==${NVSHMEM_VERSION} --force-reinstall --no-deps; \ + python3 -m pip install nvidia-nvshmem-cu12==3.3.20 --force-reinstall --no-deps; \ python3 -m pip install nvidia-cutlass-dsl==4.3.0 --force-reinstall --no-deps; \ elif [ "${CUDA_VERSION%%.*}" = "13" ]; then \ python3 -m pip install nvidia-nccl-cu13==2.28.3 --force-reinstall --no-deps ; \