Installation

TorchKM can be installed from PyPI or from source.

Basic installation

pip install torchkm

The default installation includes the high-level scikit-learn-style estimator API used throughout the examples.

Development installation

For development, clone the repository and install the package in editable mode:

git clone https://github.com/YikaiZhang95/torchkm.git
cd torchkm
pip install -e ".[dev,examples,viz]"

Then run the test suite:

pytest -q

GPU support

TorchKM uses PyTorch for tensor computation. To run on a GPU, install a PyTorch build that is compatible with your CUDA version. A typical device-selection pattern is:

import torch

device = "cuda" if torch.cuda.is_available() else "cpu"

Then pass the device to the estimator:

clf = TorchKMSVC(kernel="rbf", Cs=Cs, nC=len(Cs), cv=5, device=device)

CPU fallback

TorchKM is designed to run on CPU when CUDA is not available. This is useful for testing, examples, and smaller data sets. For large kernel matrices, GPU execution is usually preferred when available.

Verifying the installation

import torch
from torchkm.estimators import TorchKMSVC

print("PyTorch:", torch.__version__)
print("CUDA available:", torch.cuda.is_available())
print("TorchKMSVC:", TorchKMSVC)