software-engineering/spark-vllm-docker
2
0
Fork 0
Code Issues Actions Packages Projects Releases Wiki Activity

Removed temp fastsafetensors patch

Browse Source
This commit is contained in:
Eugene Rakhmatulin
2026-02-09 10:21:14 -08:00
parent 06e8817f18
commit 2923fe6ea5
3 changed files with 10 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
Dockerfile
View File

@@ -216,8 +216,8 @@ RUN --mount=type=cache,id=uv-cache,target=/root/.cache/uv \
# Apply Patches # Apply Patches
# TEMPORARY PATCH for fastsafetensors loading in cluster setup - tracking https://github.com/foundation-model-stack/fastsafetensors/issues/36 # TEMPORARY PATCH for fastsafetensors loading in cluster setup - tracking https://github.com/foundation-model-stack/fastsafetensors/issues/36
COPY fastsafetensors.patch . # COPY fastsafetensors.patch .
RUN patch -p1 < fastsafetensors.patch # RUN patch -p1 < fastsafetensors.patch
# Final Compilation # Final Compilation
# We mount the ccache directory here. Ideally, map this to a host volume for persistence # We mount the ccache directory here. Ideally, map this to a host volume for persistence

4
Dockerfile.wheels
View File

@@ -41,7 +41,7 @@ RUN mkdir -p tiktoken_encodings && \
wget -O tiktoken_encodings/cl100k_base.tiktoken "https://openaipublic.blob.core.windows.net/encodings/cl100k_base.tiktoken" wget -O tiktoken_encodings/cl100k_base.tiktoken "https://openaipublic.blob.core.windows.net/encodings/cl100k_base.tiktoken"
# Cache TEMPORARY PATCH for fastsafetensors loading in cluster setup - tracking https://github.com/foundation-model-stack/fastsafetensors/issues/36 # Cache TEMPORARY PATCH for fastsafetensors loading in cluster setup - tracking https://github.com/foundation-model-stack/fastsafetensors/issues/36
COPY fastsafetensors.patch . # COPY fastsafetensors.patch .
# Install fastsafetensors # Install fastsafetensors
RUN --mount=type=cache,id=uv-cache,target=/root/.cache/uv \ RUN --mount=type=cache,id=uv-cache,target=/root/.cache/uv \
@@ -68,7 +68,7 @@ RUN --mount=type=cache,id=uv-cache,target=/root/.cache/uv \
# Apply TEMPORARY PATCH for fastsafetensors loading in cluster setup - tracking https://github.com/foundation-model-stack/fastsafetensors/issues/36 # Apply TEMPORARY PATCH for fastsafetensors loading in cluster setup - tracking https://github.com/foundation-model-stack/fastsafetensors/issues/36
# Apply in site-packages # Apply in site-packages
RUN patch -p1 -d /usr/local/lib/python3.12/dist-packages < ${VLLM_BASE_DIR}/fastsafetensors.patch # RUN patch -p1 -d /usr/local/lib/python3.12/dist-packages < ${VLLM_BASE_DIR}/fastsafetensors.patch
ARG FLASHINFER_PRE="" ARG FLASHINFER_PRE=""

6
README.md
View File

@@ -159,6 +159,12 @@ Don't do it every time you rebuild, because it will slow down compilation times.
For periodic maintenance, I recommend using a filter: `docker builder prune --filter until=72h` For periodic maintenance, I recommend using a filter: `docker builder prune --filter until=72h`
### 2026-02-09
- Migrated to a new base image with PyTorch 2.10 compiled with Spark support.
- Triton 3.6.0 is now default.
- Removed temporary fastsafetensors patch, as proper fix is now merged into vLLM main branch.
### 2026-02-04 ### 2026-02-04
#### Recipes support #### Recipes support