# SPDX-FileCopyrightText: Copyright (c) 2024-2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: Apache-2.0
#
# This file represents the default ARG values of Dockerfiles generated
# by render.py. These are the recommended default values for users and
# is the source of truth for the values used in our delivered images.
#
# Some ARGs have multiple valid values and can be changed for local testing,
# you can do so locally in this file, or pass the --build-arg into docker build
# when building.

dynamo:
  cuda12.9:
    base_image: nvcr.io/nvidia/cuda-dl-base
    base_image_tag: 25.06-cuda12.9-devel-ubuntu24.04
  cuda13.0:
    base_image: nvcr.io/nvidia/cuda-dl-base
    base_image_tag: 25.11-cuda13.0-devel-ubuntu24.04
  epp_image: us-central1-docker.pkg.dev/k8s-staging-images/gateway-api-inference-extension/epp:v0.5.1
  frontend_image: nvcr.io/nvidia/base/ubuntu:noble-20250619
  planner_build_image: python
  planner_build_image_tag: 3.12-slim
  planner_runtime_image: nvcr.io/nvidia/distroless/python
  planner_runtime_image_tag: 3.12-v4.0.3
  python_version: "3.12"

  nats_version: v2.10.28
  etcd_version: v3.5.21

  nixl_ref: 0.10.1
  nixl_ucx_ref: v1.20.x
  nixl_gdrcopy_ref: v2.5.1
  nixl_ucx_efa_ref: 9d2b88a1f67faf9876f267658bd077b379b8bb76
  nixl_libfabric_ref: v2.3.0
  enable_kvbm: "true"
  enable_media_ffmpeg: "false"
  enable_gpu_memory_service: "true"
  ffmpeg_version: "8.1"
  efa_version: 1.47.0

vllm:
  cuda12.9:
    base_image: nvcr.io/nvidia/cuda-dl-base
    runtime_image: nvcr.io/nvidia/cuda
    base_image_tag: 25.06-cuda12.9-devel-ubuntu24.04
    runtime_image_tag: 12.9.1-runtime-ubuntu24.04
    vllm_ref: v0.19.1
  cuda13.0:
    base_image: nvcr.io/nvidia/cuda-dl-base
    runtime_image: nvcr.io/nvidia/cuda
    base_image_tag: 25.11-cuda13.0-devel-ubuntu24.04
    runtime_image_tag: 13.0.2-runtime-ubuntu24.04
    vllm_ref: v0.19.1
  xpu:
    base_image: intel/deep-learning-essentials
    runtime_image: intel/deep-learning-essentials
    base_image_tag: 2025.3.2-0-devel-ubuntu24.04
    runtime_image_tag: 2025.3.2-0-devel-ubuntu24.04
    vllm_ref: v0.17.1
  cpu:
    base_image: ubuntu
    runtime_image: ubuntu
    base_image_tag: 24.04
    runtime_image_tag: 24.04
    vllm_ref: v0.16.0
  flashinf_ref: v0.6.6
  lmcache_ref: 0.4.2
  vllm_omni_ref: "release/v0.19.0rc1"
  nixl_ref: 0.10.1
  max_jobs: "10"
  enable_media_ffmpeg: "false"
  enable_gpu_memory_service: "true"
  enable_kvbm: "true"
  enable_modelexpress_p2p: "false"
  modelexpress_ref: "76fc5d7f06c37121ee8789a29fac6f9b08c4743a"  # v0.3.0

sglang:
  cuda12.9:
    base_image: nvcr.io/nvidia/cuda-dl-base
    runtime_image: lmsysorg/sglang
    base_image_tag: 25.06-cuda12.9-devel-ubuntu24.04
    runtime_image_tag: v0.5.10.post1-runtime
  cuda13.0:
    base_image: nvcr.io/nvidia/cuda-dl-base
    runtime_image: lmsysorg/sglang
    base_image_tag: 25.11-cuda13.0-devel-ubuntu24.04
    runtime_image_tag: v0.5.10.post1-cu130-runtime
  # SGLang uses the NIXL stack from the upstream lmsysorg/sglang runtime image.
  # Do not add nixl_ref here: Dynamo does not build or install its NIXL wheel
  # for SGLang, and SGLang does not use Dynamo KVBM/block-manager at runtime.
  enable_media_ffmpeg: "true"
  enable_gpu_memory_service: "true"
  enable_kvbm: "false"

trtllm:
  cuda13.1:
    base_image: nvcr.io/nvidia/pytorch
    runtime_image: nvcr.io/nvidia/cuda-dl-base
    base_image_tag: 26.02-py3
    runtime_image_tag: 26.02-cuda13.1-runtime-ubuntu24.04
  nixl_ref: 0.10.1
  enable_media_ffmpeg: "false"
  enable_gpu_memory_service: "true"
  enable_kvbm: "true"
  python_version: "3.12"
  index_url: https://pypi.nvidia.com/
  pip_wheel_dir: /tmp/trtllm_wheel/
  pip_wheel: tensorrt-llm==1.3.0rc11
  trtllm_wheel_image: nvcr.io/nvidia/tensorrt-llm/release:${TENSORRTLLM_PIP_WHEEL#*==}

  github_trtllm_commit: v1.3.0rc11
  torch_version: 2.11.0a0+eb65b36914.nv26.2
  torch_tensorrt_version: 2.11.0a0
  torchvision_version: 0.25.0a0+1e53952f.nv26.2.44259020
  torchao_ver: 0.16.0+gita89eaab2
  torchdata_ver: 0.11.0
  torchtitan_ver: 0.2.1+git9f211ec1
  jinja2_version: 3.1.6
  sympy_version: 1.14.0
  pytorch_triton_ver: 3.6.0+git9844da95.nv26.2
  flash_attn_version: 2.7.4.post1+nv26.2.44259020
  flashinfer_python_ver: 0.6.6
  has_trtllm_context: "0"