Entrypoint /bin/bash /opt/nim/start_server.sh; WorkingDir /opt/nim; ExposedPorts 6006/tcp, 8000/tcp, 8888/tcp

--chmod=755 ./entrypoint.sh /opt/nim/start_server.sh

SERVER_START_SCRIPT_PATH=/opt/nim/start_server.sh

model_manifest.yaml /opt/nim/etc/default/model_manifest.yaml

nvs:1000

TELEMETRY_MODE=0 TELEMETRY_ENDPOINT=0 PIP_INDEX_URL=https://urm.nvidia.com/artifactory/api/pypi/nv-shared-pypi/simple [ -d /etc/ld.so.conf.d ] ||
  mkdir -p /etc/ld.so.conf.d &&
  [ -f /etc/ld.so.conf.d/nvidia.conf ] ||
  touch /etc/ld.so.conf.d/nvidia.conf &&
  echo "/usr/local/nvidia/lib" >> /etc/ld.so.conf.d/nvidia.conf &&
  echo "/usr/local/nvidia/lib64" >> /etc/ld.so.conf.d/nvidia.conf &&
  echo "/usr/local/cuda/lib64" >> /etc/ld.so.conf.d/nvidia.conf &&
  if command -v ldconfig >/dev/null 2>&1; then ldconfig; fi

TELEMETRY_MODE=0 TELEMETRY_ENDPOINT=0 PIP_INDEX_URL=https://urm.nvidia.com/artifactory/api/pypi/nv-shared-pypi/simple groupadd --gid 1000 --non-unique nvs &&
  useradd --create-home --shell /usr/sbin/nologin --uid 1000 --non-unique --gid 1000 nvs &&
  chown 1000.1000 $NIM_DIR_PATH

TELEMETRY_MODE=0 TELEMETRY_ENDPOINT=0 PIP_INDEX_URL=https://urm.nvidia.com/artifactory/api/pypi/nv-shared-pypi/simple patch /usr/local/lib/python3.12/dist-packages/nimlib/profiles.py < /tmp/patches/nimlib/profiles.patch &&
  patch /usr/local/lib/python3.12/dist-packages/nimlib/llm/profile_selection.py < /tmp/patches/nimlib/profile_selection.patch &&
  patch /usr/local/lib/python3.12/dist-packages/nimlib/download.py < /tmp/patches/nimlib/download.patch &&
  patch -p1 -d /usr/local/lib/python3.12/dist-packages/ < /tmp/patches/nimlib/custom_error_models.patch &&
  patch -p1 -d /usr/local/lib/python3.12/dist-packages/ < /tmp/patches/nimlib/start_server_banner.patch &&
  patch -p1 -d /usr/local/lib/python3.12/dist-packages/ < /tmp/patches/nimlib/api_validation_error.patch &&
  patch /workspace/vllm-source/vllm/entrypoints/openai/serving_chat.py < /tmp/patches/vllm/serving_chat_v7.patch &&
  patch /workspace/vllm-source/vllm/sampling_params.py < /tmp/patches/vllm/sampling_params.patch &&
  patch -p0 -d /workspace/vllm-source/vllm/entrypoints/openai < /tmp/patches/vllm/protocol_validation.patch &&
  patch -p1 -d /workspace/vllm-source < /tmp/patches/vllm/config_offline.patch &&
  patch -p1 -d /workspace/vllm-source < /tmp/patches/vllm/metrics_rename.patch &&
  patch -p0 -d /usr/local/lib/python3.12/dist-packages/nim_inference_sdk < /tmp/patches/nim_inference_sdk/inference_engine.patch &&
  patch -p0 -d /usr/local/lib/python3.12/dist-packages/nim_inference_sdk < /tmp/patches/nim_inference_sdk/backends/vllm_backend.patch &&
  rm -rf /tmp/patches

./patches /tmp/patches

./banner.txt /opt/nim