FROM intelanalytics/ipex-llm-xpu:2.1.0-SNAPSHOT

ARG http_proxy
ARG https_proxy

# Disable pip's cache behavior
ARG PIP_NO_CACHE_DIR=false


# Install Serving Dependencies
RUN cd /llm &&\
    # Install ipex-llm[serving] only will update ipex_llm source code without updating
    # bigdl-core-xe, which will lead to problems
    apt-get update && \
    apt-get install -y libfabric-dev wrk libaio-dev && \
    pip install --pre --upgrade ipex-llm[xpu,serving] && \
    pip install transformers==4.37.0 gradio==4.19.2 && \
    # Install vLLM-v2 dependencies
    cd /llm && \
    git clone -b sycl_xpu https://github.com/analytics-zoo/vllm.git && \
    cd vllm && \
    pip install -r requirements-xpu.txt && \
    pip install --no-deps xformers && \
    VLLM_BUILD_XPU_OPS=1 pip install --no-build-isolation -v -e . && \
    pip install outlines==0.0.34 --no-deps && \
    pip install interegular cloudpickle diskcache joblib lark nest-asyncio numba scipy && \
    # For Qwen series models support
    pip install transformers_stream_generator einops tiktoken

ADD ./vllm_offline_inference.py       /llm/
ADD ./payload-1024.lua                /llm/
ADD ./start-vllm-service.sh           /llm/
ADD ./benchmark_vllm_throughput.py   /llm/
ADD ./start-fastchat-service.sh       /llm/

WORKDIR /llm/