daviestechlabs/kuberay-images
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

more fixes.
Some checks failed
Build and Push Images / determine-version (push) Successful in 6s
Details
Build and Push Images / build (Dockerfile.ray-worker-nvidia, nvidia) (push) Failing after 40s
Details
Build and Push Images / build (Dockerfile.ray-worker-intel, intel) (push) Failing after 43s
Details
Build and Push Images / Release (push) Has been cancelled
Details
Build and Push Images / Notify (push) Has been cancelled
Details
Build and Push Images / build (Dockerfile.ray-worker-strixhalo, strixhalo) (push) Has been cancelled
Details
Build and Push Images / build (Dockerfile.ray-worker-rdna2, rdna2) (push) Has been cancelled
Details

Browse Source
This commit is contained in:
Billy Davies
2026-02-09 10:56:51 -05:00
parent 2c38cce20c
commit c9cf143821
2 changed files with 6 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
dockerfiles/Dockerfile.ray-worker-strixhalo
View File

@@ -86,6 +86,11 @@ RUN --mount=type=cache,target=/home/ray/.cache/uv,uid=1000,gid=1000 \
--index-url https://rocm.nightlies.amd.com/v2/gfx110X-all/ \
torch torchaudio torchvision
# FIX: Uninstall flash_attn — it was compiled against the vLLM ROCm wheel's
# PyTorch, but the TheRock nightly above has a different c10::hip ABI.
# vLLM ROCm uses its own Triton/CK attention backends, so flash_attn is not needed.
RUN pip uninstall -y flash-attn 2>/dev/null || true
# FIX: Ray base image has pandas 1.5.3 which is incompatible with numpy 2.x
# The TheRock PyTorch wheels require numpy 2.x, so upgrade pandas to match.
# Pin numpy <2.3 because numba (required by vLLM for speculative decoding)