All Tools
S
OtherFreeOpen Source
SKORCH
scikit-learn compatible neural network library for PyTorch
BSD-3-Clause
ABOUT
PyTorch provides powerful deep learning primitives but lacks the high-level training abstractions that scikit-learn users expect, such as fit/predict interfaces, grid search, and cross-validation. This creates a steep learning curve for data scientists who want to adopt neural networks without rewriting their entire ML pipeline. Skorch bridges this gap by wrapping PyTorch modules in a scikit-learn compatible estimator interface, enabling seamless integration with existing scikit-learn workflows, pipelines, and model selection.
INSTALL
pip install skorchINTEGRATION GUIDE
1. Train PyTorch neural networks using familiar scikit-learn fit and predict methods
2. Perform hyperparameter tuning with scikit-learn GridSearchCV on PyTorch models
3. Build ensemble models combining PyTorch neural networks with traditional ML estimators
4. Deploy PyTorch models in production pipelines designed for scikit-learn interfaces
5. Migrate existing scikit-learn workflows to deep learning without pipeline rewrites
TAGS
pytorchscikit-learnneural-networksmachine-learningpythonopen-source