#4141st of 126 in ML frameworks and libraries
tabpfn
Pretrained transformer for tabular classification and regression, no per-dataset training
- Stars
- 8.3k
- License
- Apache-2.0
- Last commit
- Oct 2026
- Last release
- Oct 2026
- Language
- Python
5,928 on 4 Apr 20268,339 on 31 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
TabPFN is a Python library with scikit-style TabPFNClassifier and TabPFNRegressor estimators built on a transformer pretrained on synthetic datasets. Calling fit stores the training data and predict runs a single forward pass, so no per-dataset gradient training is needed. The default TabPFN-3.5 checkpoint accepts up to 1,000,000 rows and 20,000 features. Model weights carry non-commercial licenses, separate from the Apache-2.0 code.
Who it is for: Data scientists applying ML to tabular datasets in Python
Strengths
- Drop-in fit/predict estimators for classification and regression, installed with pip
- Data stays local; works offline once weights are downloaded
- Several checkpoints selectable, including smaller TabPFN-3.5-Fast and older V2, V3
- Runs on Nvidia, AMD ROCm and Apple Silicon GPUs, with CPU fallback
Weaknesses
- TabPFN-2.5 to 3.5 weights are non-commercial; commercial use needs a license
- First use requires browser login and license acceptance, or a TABPFN_TOKEN
- CPU use capped at 5000 samples for recent models unless overridden
- Fast distillation inference mode is Enterprise-only
What it needs
- GPU optional
- Needs PyTorch, PriorLabs account token for weight download
- Models: TabPFN-3.5, TabPFN-3.5-Fast, TabPFN-3, TabPFN-2
Quick answers
Is tabpfn open source?
Yes. Its code is on GitHub under the Apache-2.0 license.
Does tabpfn 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 |