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

Implemented a temporary patch for recently broken MiniMax-M2 (in builds after 12/10) for some quants.

Browse Source
This commit is contained in:
eugr
2025-12-11 11:13:05 -08:00
parent 9d351cd6d5
commit 5fba205db4
2 changed files with 8 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
Dockerfile
View File

@@ -80,6 +80,9 @@ RUN python3 use_existing_torch.py && \
COPY fastsafetensors.patch . COPY fastsafetensors.patch .
RUN patch -p1 < fastsafetensors.patch RUN patch -p1 < fastsafetensors.patch
# TEMPORARY PATCH for broken MiniMax M2 - tracking https://github.com/vllm-project/vllm/issues/30445 and https://github.com/vllm-project/vllm/pull/30389
RUN curl -L https://patch-diff.githubusercontent.com/raw/vllm-project/vllm/pull/30389.diff | git apply
# Final Build # Final Build
# Uses --no-build-isolation to respect the pre-installed Torch/FlashInfer # Uses --no-build-isolation to respect the pre-installed Torch/FlashInfer
RUN pip install --no-build-isolation . -v RUN pip install --no-build-isolation . -v

5
README.md
View File

@@ -12,6 +12,11 @@ The Dockerfile builds from the main branch of VLLM, so depending on when you run
## CHANGELOG ## CHANGELOG
### 2025-12-11
Applied a patch to fix broken MiniMax-M2 in some quants after [this commit](https://github.com/vllm-project/vllm/commit/d017bceb08eaac7bae2c499124ece737fb4fb22b) until [this PR](https://github.com/vllm-project/vllm/pull/30389) is approved.
See [this issue](https://github.com/vllm-project/vllm/issues/30445) for details.
### 2025-12-05 ### 2025-12-05
Added `build-and-copy.sh` for convenience. Added `build-and-copy.sh` for convenience.