#8585th of 126 in ML frameworks and libraries
skorch
scikit-learn compatible wrapper for training PyTorch models
- Stars
- 6.2k
- License
- BSD-3-Clause
- Last commit
- Oct 2026
- Last release
- May 2026
- Language
- Jupyter Notebook
6,131 on 3 Apr 20266,185 on 30 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
skorch wraps PyTorch modules in estimators such as NeuralNetClassifier, so they expose fit, predict_proba and set_params and work inside sklearn Pipeline and GridSearchCV. It adds callbacks for LR scheduling, early stopping, checkpointing, parameter freezing and sklearn scoring, plus integrations with GPyTorch and Hugging Face.
Who it is for: Python ML engineers who use scikit-learn tooling with PyTorch models
Strengths
- Drop-in sklearn API: Pipeline and GridSearchCV work with PyTorch modules
- Callbacks for early stopping, checkpointing, LR schedulers and sklearn scoring
- Documented integrations with GPyTorch and Hugging Face
- Officially supports the last four minor PyTorch versions
Weaknesses
- PyTorch is not a declared dependency; users must install a matching version
- Requires Python 3.9 or higher
- Conda package is maintained by conda-forge, not the skorch maintainers
- PyTorch only; no other tensor backends
What it needs
- GPU optional
- Docker
- Needs PyTorch, scikit-learn
- Models: PyTorch nn.Module, Hugging Face, GPyTorch
Quick answers
Is skorch open source?
Yes. Its code is on GitHub under the BSD-3-Clause license.
Can you run skorch with Docker?
Yes. Its repository has a Docker setup: a Dockerfile, a Compose file or a published image.
Does skorch need a GPU?
No. It runs on a CPU and can use a GPU when you have one.
Also in ML frameworks and libraries
See all 126| Rank | Project | Score |
|---|---|---|
| 1 | pytorchPython tensor library with GPU support and tape-based autograd for deep learning | 95 out of 100 |
| 2 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 94 out of 100 |
| 3 | ultralyticsPython library and CLI for training and running YOLO vision models | 93 out of 100 |
| #4 | pytorch-image-modelsPyTorch library of image models, pretrained weights and reference training scripts | 93 out of 100 |
| #5 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 91 out of 100 |