TabPFN

TabPFN is a foundation model for tabular data that outperforms traditional methods while being dramatically faster. This repository contains the core PyTorch implementation with CUDA optimization.

⚠️ Major Update: Version 2.0: Complete codebase overhaul with new architecture and features. Previous version available at v1.0.0 and pip install tabpfn<2.

📚 For detailed usage examples and best practices, check out:

Interactive Colab Tutorial

🌐 TabPFN Ecosystem

Choose the right TabPFN implementation for your needs:

TabPFN Client: Easy-to-use API client for cloud-based inference
TabPFN Extensions: Community extensions and integrations
TabPFN (this repo): Core implementation for local deployment and research

Try our Interactive Colab Tutorial to get started quickly.

🏁 Quick Start

Installation

# Simple installation
pip install tabpfn

# Local development installation
git clone https://github.com/PriorLabs/TabPFN.git
pip install -e "tabpfn[dev]"

Basic Usage

from sklearn.datasets import load_breast_cancer
from sklearn.metrics import accuracy_score, roc_auc_score
from sklearn.model_selection import train_test_split

from tabpfn import TabPFNClassifier

# Load data
X, y = load_breast_cancer(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)

# Initialize a classifier
clf = TabPFNClassifier()
clf.fit(X_train, y_train)

# Predict probabilities
prediction_probabilities = clf.predict_proba(X_test)
print("ROC AUC:", roc_auc_score(y_test, prediction_probabilities[:, 1]))

# Predict labels
predictions = clf.predict(X_test)
print("Accuracy", accuracy_score(y_test, predictions))

💡 Usage Tips

TabPFN is designed to work out-of-the-box with minimal preprocessing:

No preprocessing needed: TabPFN handles normalization internally
Categorical variables: Use numerical encodings (floats for ordered, OrdinalEncoder for unordered)
Automatic ensembling: Controls with n_estimators
Independent predictions: Test samples can be predicted individually or in batch
Differentiable: Core model is differentiable (except preprocessing)
GPU Support: Use device='cuda' for GPU acceleration

📜 License

Prior Labs License (Apache 2.0 with additional attribution requirement)

📚 Citation

@article{hollmann2024tabpfn,
 title={Accurate predictions on small data with a tabular foundation model},
 author={Hollmann, Noah and M{\"u}ller, Samuel and Purucker, Lennart and
         Krishnakumar, Arjun and K{\"o}rfer, Max and Hoo, Shi Bin and
         Schirrmeister, Robin Tibor and Hutter, Frank},
 journal={Nature},
 year={2024},
 month={01},
 day={09},
 doi={10.1038/s41586-024-08328-6},
 publisher={Springer Nature},
 url={https://www.nature.com/articles/s41586-024-08328-6},
}

🤝 Join Our Community

We're building the future of tabular machine learning and would love your involvement:

Connect & Learn:
- Join our Discord Community
- Read our Documentation
- Check out GitHub Issues
Contribute:
- Report bugs or request features
- Submit pull requests
- Share your research and use cases
Stay Updated: Star the repo and join Discord for the latest updates

🛠️ Development

Setup environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
git clone https://github.com/PriorLabs/TabPFN.git
cd tabpfn
pip install -e ".[dev]"
pre-commit install

Before committing:

pre-commit run --all-files

Run tests:

pytest tests/