tabpfn-v2
TabPFN v2 — Tabular Classification & Regression
Built with PriorLabs-TabPFN · classify or predict on your table
Choose this to classify rows or predict a number per row from a small labelled table (up to 10,000 rows), without training. Inference-only.
Run it
$ pip install pfnstudio
$ pfnstudio run tabpfn-v2 --input <your file>
# ex: each task on the sample input pfnstudio ships
$ pfnstudio run tabpfn-v2 --task classify --input sample
$ pfnstudio run tabpfn-v2 --task regress --input sample
# download it without running · its tasks, inputs, license and digest
$ pfnstudio pull tabpfn-v2
$ pfnstudio show tabpfn-v2
# models loaded now · delete it from this machine
$ pfnstudio ps
$ pfnstudio rm tabpfn-v2
Python 3.10+. The first run downloads the model's adapter, asks before installing PyTorch and the model's own packages (-y installs without asking) and pulls the pretrained weights from Hugging Face. --input sample runs the sample input pfnstudio ships and prints its path: copy it, put your data in, and pass it as <your file>. Later runs reuse the cache.
Serves
| Task | Answers |
|---|---|
| classify | which class is this row? |
| regress | what number for this row? |
Provenance
| Publisher | Prior Labs ↗ |
| Weights | pretrained TabPFN v2 · classifier + regressor checkpoints from HuggingFace |
| Input | (N, F) table with a target column |
| License | Prior Labs License 1.1 (Apache-2.0 + attribution: Built with PriorLabs-TabPFN) |
| Adapter | basemodels/tabpfn-v2/adapters/ · tabpfn_v2 |
| Source | huggingface.co/Prior-Labs/TabPFN-v2-clf ↗ |
Readme
Predict a column of your table — churn, category, price, risk — from the rows where you already know it, without training a model. Built with PriorLabs-TabPFN.
TabPFN v2 (Prior Labs, Nature 2025) is a prior-fitted network for small tabular data: the labelled rows of your table are its examples and it predicts the target for the rows left empty, for classification and regression. This listing serves the v2 weights (HF Prior-Labs/TabPFN-v2-clf and -reg) through the upstream tabpfn package. Built with PriorLabs-TabPFN. The weights are under the Prior Labs License 1.1 (Apache-2.0 with an attribution requirement); newer TabPFN versions are under non-commercial licenses and are not offered here. Up to 10 classes, 500 features and 10,000 rows (1,000 on CPU).