wehub-resource-sync
22 lines
792 B
Docker
22 lines
792 B
Docker
# This is a lightweight image of the latest release of LMCache (without NIXL) bundled with the latest release of vLLM.
|
|
# Note: You will not be able to run Prefill-Decode Disaggregation (PD) with this image because NIXL is not installed.
|
|
FROM vllm/vllm-openai:latest
|
|
|
|
RUN uv pip install --system --no-cache lmcache
|
|
|
|
|
|
# Build:
|
|
# docker build -f Dockerfile.lightweight -t 'lmcache/vllm-openai:lightweight' .
|
|
|
|
# Run:
|
|
# export HF_TOKEN=<your_huggingface_token>
|
|
# docker run --runtime nvidia --gpus all \
|
|
# -v ~/.cache/huggingface:/root/.cache/huggingface \
|
|
# --env "HF_TOKEN=$HF_TOKEN" \
|
|
# -p 8000:8000 \
|
|
# --ipc=host \
|
|
# lmcache/vllm-openai:lightweight \
|
|
# --model Qwen/Qwen3-0.6B \
|
|
# --kv-transfer-config \
|
|
# '{"kv_connector":"LMCacheConnectorV1","kv_role":"kv_both"}'
|