Installation

Python ≥ 3.10, < 3.14 is required. Python 3.14 is not yet supported.

CPU

pip install wsi-patching

GPU

pip install wsi-patching[gpu]

The [gpu] extra adds torch, torchvision, and cupy-cuda12x (Linux only). The following components require it:

Component	Reason
PenArtifactFilter	Batched pen-colour detection runs on CUDA
CellVitTissueClassifierFilter	MobileNetV3 classifier runs on GPU
MacenkoNormalizer	Stain normalisation uses CuPy for speed
TorchStreamWriter	Output type is torch.Tensor

All GPU components fail clearly at import time with an ImportError if the extras are missing.

Development

git clone https://github.com/amspath/wsi-patching-pipeline.git
cd wsi-patching-pipeline
uv sync --extra dev,gpu

uv is recommended. Plain pip also works:

pip install -e ".[dev,gpu]"

---

Pinned fastslide version

The slide-reading backend fastslide==0.5.2 is a hard-pinned dependency. If you see version conflicts from another package requiring a different fastslide version, isolate wsi-patching in its own virtual environment.