diff --git a/Dockerfile b/Dockerfile index 3300cfd..b8dae57 100644 --- a/Dockerfile +++ b/Dockerfile @@ -80,6 +80,9 @@ RUN python3 use_existing_torch.py && \ COPY 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 # Uses --no-build-isolation to respect the pre-installed Torch/FlashInfer RUN pip install --no-build-isolation . -v diff --git a/README.md b/README.md index e9fdc76..120b059 100644 --- a/README.md +++ b/README.md @@ -12,6 +12,11 @@ The Dockerfile builds from the main branch of VLLM, so depending on when you run ## 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 Added `build-and-copy.sh` for convenience.