Using manual deps management¶

This page explains how to turn off automatic deps management and handle dependencies yourself.

1. Install the hardware backend¶

Install only the backend matching your setup.

uv add "raitap[torch-cuda]"

pip

pip install "raitap[torch-cuda]"

	CPU	CUDA	Intel GPU
Torch	`torch-cpu`	`torch-cuda`	`torch-intel`
ONNX	`onnx-cpu`	`onnx-cuda`	`onnx-intel`

Note

CUDA = NVIDIA GPUs.
torch-intel uses the Intel XPU API; onnx-intel uses the OpenVINO ONNX Runtime.
Apple MPS support is coming soon.

CUDA / Intel wheels do not live on PyPI. Re-declare the index routing per hardware:

Important

Replace BACKEND in the snippets below with torch or onnx depending on your setup.

CPU

On Linux, the plain torch package from PyPI is the CUDA build. Route it to the CPU index by adding the following to your pyproject.toml:

[project]
dependencies = [
  "raitap[BACKEND-cpu]>=0.9",
  "torch;       sys_platform == 'linux'", # redeclare to make the index below apply
  "torchvision; sys_platform == 'linux'", # redeclare to make the index below apply
]

[[tool.uv.index]]
name = "pytorch-cpu"
url = "https://download.pytorch.org/whl/cpu"
explicit = true

[tool.uv.sources]
torch = { index = "pytorch-cpu" }
torchvision = { index = "pytorch-cpu" }

CUDA

torch does not publish separate CUDA builds on PyPI. You control the version via the index by adding the following to your pyproject.toml:

[project]
dependencies = [
  "raitap[BACKEND-cuda]>=0.9; sys_platform != 'darwin'",
  "torch;       sys_platform != 'darwin'", # redeclare to make the index below apply
  "torchvision; sys_platform != 'darwin'", # redeclare to make the index below apply
]

[[tool.uv.index]]
name = "pytorch-cuda"
url = "https://download.pytorch.org/whl/cu126" # CUDA 12.6, modern default; for older GPUs see below
explicit = true

[tool.uv.sources]
torch = { index = "pytorch-cuda" }
torchvision = { index = "pytorch-cuda" }

Intel GPU

triton-xpu's PyPI releases are either yanked or pre-release. Route it to the Intel index by adding the following to your pyproject.toml:

[project]
dependencies = [
  "raitap[BACKEND-intel]>=0.9; sys_platform != 'darwin'",
  "torch;       sys_platform != 'darwin'", # redeclare to make the index below apply
  "torchvision; sys_platform != 'darwin'", # redeclare to make the index below apply
  "triton-xpu>=3.0.0rc0; sys_platform != 'darwin' and python_full_version < '3.14'",
]

[[tool.uv.index]]
name = "pytorch-intel"
url = "https://download.pytorch.org/whl/xpu"
explicit = true

[tool.uv.sources]
torch = { index = "pytorch-intel" }
torchvision = { index = "pytorch-intel" }
triton-xpu = { index = "pytorch-intel" }

Wheel availability¶

PyTorch indexes don't publish for every Python minor:

Extra	Index	Python	Platforms
`torch-cpu` / `onnx-cpu`	`/whl/cpu`	3.11–3.13	Linux, macOS, Windows
`torch-cuda` / `onnx-cuda`	`/whl/cu126`	3.11–3.13	Linux, Windows
`torch-intel` / `onnx-intel`	`/whl/xpu`	3.11–3.13	Linux, Windows

Faster locking on one OS¶

If you only target one OS, scope the resolver so it doesn't solve for all platforms at once:

[tool.uv]
environments = ["sys_platform == 'linux' and python_version >= '3.12'"]

2. Pick assessment extras¶

You can either:

Install a whole module (every library it offers):

uv add "raitap[transparency]"

pip

pip install "raitap[transparency]"

Install a single library / report format. The extra is the library name (captum, mlflow) or the format (html, pdf). You can combine it all in one line:
uv
uv add "raitap[onnx-cpu,transparency,metrics]"
pip
pip install "raitap[onnx-cpu,transparency,metrics]"

3. Run the config with automatic deps management off¶

Pass --custom-deps; install everything the config needs first.

uv run raitap --config-dir my-configs --config-name assessment --custom-deps

pip

raitap --config-dir my-configs --config-name assessment --custom-deps