Files
llama-4-bi150/Dockerfile

40 lines
1.7 KiB
Docker
Raw Normal View History

# Llama 4 Scout / Maverick on Iluvatar BI-V150
#
# 一个镜像同时支持:
# - Llama-4-Scout-17B-16E-Instruct(BF16,TP8,8 卡)
# - Llama-4-Maverick-17B-128E-Instruct-FP8(FP8,TP8+PP2,16 卡)
#
# 镜像只包含运行时依赖和 Llama 4 补丁,不设置启动命令、ENV、WORKDIR 或 EXPOSE;
# 模型路径、端口、TP/PP、GPU、挂载等全部由 run.sh / docker-compose.yml 在启动时传入。
#
# 构建:
# docker build -t llama4-bi150:2026-10-26 .
#
# 运行(由 run.sh 传入 entrypoint 和全部参数):
# ./run.sh /mnt/disk0/models/LLM-Research/Llama-4-Scout-17B-16E-Instruct
# ./run.sh /mnt/disk0/models/LLM-Research/Llama-4-Maverick-17B-128E-Instruct-FP8
#
# 补丁是针对下面这个精确基础镜像生成的;换基础镜像需要重新生成 diff。
# registry.iluvatar.com.cn:10443/customer/sz/vllm0.17.0-4.4.0-x86:v7
# (vLLM 0.17.0+corex.20260612070235)
FROM registry.iluvatar.com.cn:10443/customer/sz/vllm0.17.0-4.4.0-x86:v7
ARG IMAGE_VERSION=2026-10-26
LABEL org.opencontainers.image.title="llama4-bi150" \
org.opencontainers.image.description="Llama 4 Scout / Maverick on Iluvatar BI-V150 with vLLM 0.17.0+corex" \
org.opencontainers.image.version="${IMAGE_VERSION}"
COPY patches /opt/llama4/patches
COPY apply_patch.py /opt/llama4/apply_patch.py
COPY detect_model.py /opt/llama4/detect_model.py
COPY entrypoint.sh /usr/local/bin/llama4-entrypoint.sh
RUN chmod +x /usr/local/bin/llama4-entrypoint.sh \
&& chmod +x /opt/llama4/detect_model.py \
&& python3 /opt/llama4/apply_patch.py
# 故意不设置 ENV / WORKDIR / EXPOSE / ENTRYPOINT / CMD:
# 镜像保持通用,启动方式由 run.sh 或 docker-compose.yml 决定。