Skip to content

learnables/torchml

Repository files navigation


Test Status

torchml implements the scikit-learn API on top of PyTorch. This means we automatically get GPU support for scikit-learn and, when possible, differentiability.

Resources

Getting Started

pip install torchml

Minimal Linear Regression Example

import torchml as ml

(X_train, y_train), (X_test, y_test) = generate_data()

# API closely follows scikit-learn
linreg = ml.linear_model.LinearRegression()
linreg.fit(X_train, y_train)
linreg.predict(X_test)

Changelog

A human-readable changelog is available in the CHANGELOG.md file.

Citing

To cite torchml repository in your academic publications, please use the following reference.

Sébastien M. R. Arnold, Lucy Xiaoyang Shi, Xinran Gao, Zhiheng Zhang, and Bairen Chen. 2023. "torchml: a scikit-learn implementation on top of PyTorch".

You can also use the following Bibtex entry:

@misc{torchml,
  author={Arnold, S{\'e}bastien M R and Shi, Lucy Xiaoyang and Gao, Xinran and Zhang, Zhiheng and Chen, Bairen},
  title={torchml: A scikit-learn implementation on top of PyTorch},
  year={2023},
  url={https://github.com/learnables/torchml},
}