README.md
| 1 | --- |
| 2 | pipeline_tag: tabular-regression |
| 3 | library_name: tabpfn |
| 4 | license: other |
| 5 | license_name: priorlabs-1-1 |
| 6 | license_link: https://github.com/PriorLabs/TabPFN/blob/49394b053a6759cfe68e90c21a2d51c31b396768/LICENSE |
| 7 | --- |
| 8 | |
| 9 | # TabPFN v2: A Tabular Foundation Model |
| 10 | |
| 11 | TabPFN is a transformer-based foundation model for tabular data that leverages prior-data based learning to achieve strong performance on small tabular regression tasks without requiring task-specific training. |
| 12 | |
| 13 | ## Installation |
| 14 | ```bash |
| 15 | pip install tabpfn |
| 16 | ``` |
| 17 | |
| 18 | ## Model Details |
| 19 | - **Developed by:** Prior Labs |
| 20 | - **Model type:** Transformer-based foundation model for tabular data |
| 21 | - **License:** [Prior Labs License (Apache 2.0 with additional attribution requirement)](https://priorlabs.ai/tabpfn-license/) |
| 22 | - **Paper:** Published in Nature (January 2025) |
| 23 | - **Repository:** [GitHub - priorlabs/tabpfn](https://github.com/priorlabs/tabpfn) |
| 24 | |
| 25 | ### 📚 Citation |
| 26 | |
| 27 | ```bibtex |
| 28 | @article{hollmann2025tabpfn, |
| 29 | title={Accurate predictions on small data with a tabular foundation model}, |
| 30 | author={Hollmann, Noah and M{\"u}ller, Samuel and Purucker, Lennart and |
| 31 | Krishnakumar, Arjun and K{\"o}rfer, Max and Hoo, Shi Bin and |
| 32 | Schirrmeister, Robin Tibor and Hutter, Frank}, |
| 33 | journal={Nature}, |
| 34 | year={2025}, |
| 35 | month={01}, |
| 36 | day={09}, |
| 37 | doi={10.1038/s41586-024-08328-6}, |
| 38 | publisher={Springer Nature}, |
| 39 | url={https://www.nature.com/articles/s41586-024-08328-6}, |
| 40 | } |
| 41 | ``` |
| 42 | |
| 43 | ## Quick Start |
| 44 | |
| 45 | 📚 For detailed usage examples and best practices, check out: |
| 46 | - [Interactive Colab Tutorial](https://tinyurl.com/tabpfn-colab-api) |
| 47 | |
| 48 | |
| 49 | ## Technical Requirements |
| 50 | - Python ≥ 3.9 |
| 51 | - PyTorch ≥ 2.1 |
| 52 | - scikit-learn ≥ 1.0 |
| 53 | - Hardware: 16GB+ RAM, CPU (GPU optional) |
| 54 | |
| 55 | ## Limitations |
| 56 | - Not designed for very large datasets |
| 57 | - Not suitable for non-tabular data formats |
| 58 | |
| 59 | ## Resources |
| 60 | - **Documentation:** https://priorlabs.ai/docs |
| 61 | - **Source:** https://github.com/priorlabs/tabpfn |
| 62 | - **Paper:** https://www.nature.com/articles/s41586-024-08328-6 |
| 63 | |
| 64 | ### Team |
| 65 | - Noah Hollmann |
| 66 | - Samuel Müller |
| 67 | - Lennart Purucker |
| 68 | - Arjun Krishnakumar |
| 69 | - Max Körfer |
| 70 | - Shi Bin Hoo |
| 71 | - Robin Tibor Schirrmeister |
| 72 | - Frank Hutter |
| 73 | - Eddie Bergman |
| 74 | - Léo Grinsztajn |