Patch deepseek_v4 sparse attn cutedsl op for cutlass.min
This commit is contained in:
@@ -1,3 +1,4 @@
|
||||
FROM vllm/vllm-openai:nightly
|
||||
|
||||
RUN pip3 install --force-reinstall --no-deps nvidia-cutlass-dsl-libs-cu13==4.5.2
|
||||
RUN pip3 install --force-reinstall --no-deps nvidia-cutlass-dsl-libs-cu13==4.5.2 && \
|
||||
sed -i 's/cute.arch.fmin/cutlass.min/g' /usr/local/lib/python3.12/dist-packages/vllm/models/deepseek_v4/nvidia/ops/sparse_attn_compress_cutedsl.py
|
||||
|
||||
Reference in New Issue
Block a user