biondizzle/vllm-glm
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

well need cuda

Browse Source
This commit is contained in:
biondizzle
2026-04-15 07:41:18 +00:00
parent fec79d93e5
commit 9cbc1e2777
1 changed files with 5 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
Dockerfile
View File

@@ -10,9 +10,12 @@ COPY utils.py /usr/local/lib/python3.12/dist-packages/vllm/tool_parsers/utils.py
COPY vllm_patches/hf.py /usr/local/lib/python3.12/dist-packages/vllm/renderers/hf.py
# Install LMCache for KV cache offloading / sharing across nodes
# NO_CUDA_EXT=1 skips CUDA extension build (not needed for Redis backend)
# Reinstall torch with CUDA 13 support, then build LMCache with CUDA extensions
RUN apt-get update && apt-get install -y git && \
# Reinstall torch with CUDA 13.x support to match base image CUDA version
pip install --no-cache-dir --upgrade torch --index-url https://download.pytorch.org/whl/cu131 && \
# Clone and build LMCache from source with CUDA extensions
git clone --depth 1 https://github.com/LMCache/LMCache.git /tmp/lmcache && \
cd /tmp/lmcache && \
NO_CUDA_EXT=1 pip install --no-cache-dir . && \
pip install --no-cache-dir . && \
rm -rf /tmp/lmcache