<ul><li>Recent advancements in tabular classification leverage the in-context learning capability of Large Language Models (LLMs), with TabFlex being the latest model enhancing efficiency and scalability for larger datasets.</li><li>TabFlex incorporates linear attention mechanisms, enabling seamless scaling to handle tabular datasets with thousands of features and hundreds of classes, processing over a million samples in just 5 seconds.</li><li>Extensive evaluations show that TabFlex achieves over a 2x speedup compared to TabPFN and a 1.5x speedup over XGBoost, outperforming 25 tested baselines in terms of efficiency across diverse datasets.</li><li>TabFlex demonstrates strong performance on large-scale datasets with reduced computational costs when combined with data-efficient techniques such as dimensionality reduction and data sampling.</li></ul>

TabFlex: Scaling Tabular Learning to Millions with Linear Attention

Discover more