Models & Labs
NVIDIA Kumo Tabular beats GBDT on benchmarks
Hugging Face Bloghigh confidence
Why it matters
- →Replaces manual feature engineering and hyperparameter tuning with zero-shot inference.
- →Achieves state-of-the-art accuracy on tabular benchmarks while being faster than existing foundation models.
- →Trained entirely on synthetic data, offering a scalable path for enterprise structured data tasks.
NVIDIA has released Kumo Tabular, an open-source foundation model for tabular data available on Hugging Face and GitHub. The model, trained entirely on synthetic data generated by structural causal models, performs zero-shot classification and regression in a single forward pass without requiring feature engineering or hyperparameter tuning. It ranks first on the TabArena, BeyondArena, TALENT, and ScoringBench benchmarks, outperforming traditional gradient-boosted trees and other foundation models in both accuracy and inference speed. The release includes model weights for three sizes (28M to 215M parameters) under the OpenMDW-1.1 license.
Read originalMore from Hugging Face Blog
More in Models & Labs
© TechCrunch AIModels & Labsproductivity
OpenAI launches collaborative office suite features
OpenAI is directly challenging Microsoft’s dominance by embedding a full office suite into ChatGPT. The new 'Space' feature acts as a shared workspace where users and AI agents collaborate in real-time, while 'Pages' serves as an agent-native word processor. Collaborative slides allow teams to co-edit presentations generated from conversation. This marks a strategic pivot from pure chatbot utility to comprehensive workplace infrastructure, forcing incumbents to defend their core business models against an AI-native competitor.
TechCrunch AI
© Wes RothInvestment · $42 billion net loss
