NVIDIA Releases Kumo Tabular: Open Tabular Foundation Models That Predict New Rows in a Single Forward Pass
NVIDIA has released Kumo Tabular, a new family of tabular foundation models (TFMs) for classification and regression. If you have followed TabPFN or TabICL, the setup will look familiar. The model takes labeled rows as context and predicts new rows in one forward pass. There is no training, no hyperparameter tuning, and no feature engineering.
Kumo Tabular comes in Small, Medium, and Large versions, spanning roughly 28M to 215M parameters. It runs through NVIDIA's open-source structured-data-models (SDM) library.
Is it deployable? Yes. Weights ship under the OpenMDW-1.1 license, which permits commercial use. The SDM code is Apache-2.0, and it requires Python 3.11+ and PyTorch 2.7+, with examples targeting a CUDA GPU. As of 2026, this aligns with the industry's accelerating shift toward inference-based, foundation-model-driven tabular workflows that eliminate the traditional train-deploy-retrain pipeline.
What the SDM Library Adds
SDM is a GPU-native library for structured-data foundation models and preprocessing. Besides Kumo Tabular, it ships TabICLv2, Google's TabFM, and KumoRelational for multi-table data. All models share one in-context learning interface built on a TableTensor container. The library also handles preprocessing, ensembling, and many-class prediction.
How Kumo Tabular Works
Kumo Tabular extends the in-context learning paradigm to structured data. Instead of fitting parameters to a training set, the model conditions on a small set of labeled examples provided at inference time and directly outputs predictions for unseen rows. This zero-shot approach means a single forward pass replaces the entire classical ML pipelineβno gradient updates, no grid search, no feature scaling decisions. The SDM library wraps this via TableTensor, abstracting preprocessing, ensembling, and multi-class handling so developers can plug in tabular data and receive predictions with minimal code.
Key Takeaway
Kumo Tabular brings NVIDIA into the open tabular foundation model race alongside TabPFN and TabICL, offering commercial-friendly licensing and a unified, GPU-native ecosystem for structured data. For teams building predictive systems on tabular data in 2026, it represents a low-friction, training-free path from data to deployment.
via MarkTechPost