Tools
PriorLabs/TabPFN: Nature-Published Foundation Model for Tabular Data Trending at +218 Stars/Day
TabPFN is a transformer-based foundation model for tabular data classification and regression, published in Nature (2025) and developed by Prior Labs. Version 7.1.1 (April 9) supports datasets up to 50K rows and 2K features with minimal preprocessing — no scaling or one-hot encoding needed. Includes TabPFN-2.6, unsupervised extensions (outlier detection, embeddings, data generation), and SHAP-based interpretability. At 6.7K stars with 218/day, it challenges traditional ML pipelines for small-to-medium tabular workloads.
Source
↳ Follow the thread