Sakhanda Wire
NVDA $154.00 -1.37% MSFT $506.28 -2.26% GOOGL $280.87 -1.00% META $620.34 -0.97% AMZN $256.88 -0.34%
← Back to the news

Nums AI releases Causilo tabular foundation model

Nums AI has released Causilo, a pretrained foundation model for tabular classification and regression. It provides a scikit-learn interface, accepts NumPy arrays and pandas DataFrames, and can handle categorical features and missing values. The model supports classification with up to 10 classes. Its regression interface returns mean predictions by default, while version 1.0.1 adds median and quantile outputs based on 999 native quantiles.

On the official TabArena evaluation, covering 51 datasets and 816 Full splits, Causilo achieved the highest Elo among individual models overall, as well as in classification and regression. Its overall Elo was 1792.9, compared with 1764.4 for Google Research’s TabFM. A maintainer reproduced the result; when system entries were included, Causilo ranked third among 88 entries. The leading models’ confidence intervals overlap, indicating a narrow margin.

Causilo uses in-context learning: fitting stores training rows as context without changing pretrained weights. Nums AI says it was trained only on synthetic data. The code is available under Apache-2.0, while commercial, production and hosted API use require a separate license. Research and evaluation deployment is supported on CUDA or CPU.

Written by Sakhanda Wire AI from reporting by MarkTechPost. Facts, quotes and full context are in the original article below.

Originally published by MarkTechPost on

Read the original on MarkTechPost ↗

The reporting and the image belong to MarkTechPost and are credited and linked above; the text on this page is Sakhanda Wire's own summary.

← Back to the news

More stories

All the latest news