Xiaomi today released Xiaomi-TabLDM, a general-purpose foundation model for tabular data.
According to Xiaomi, Xiaomi-TabLDM uses a single pre-trained model and a unified default configuration, so it can be applied directly to different tabular datasets. It handles classification and regression predictions without per-task retraining, parameter tuning, or post-hoc ensembling.
Xiaomi says the model is designed to lower the barrier to entry for tabular data modeling, providing general-purpose capabilities for structured data analysis and prediction across different scenarios.