AI
TabFM: A Zero-Shot Foundation Model for Tabular Data
Publication: Hugging Face Daily Papers
Author: Not available in collected metadata
Source published: 28/09/2026, 20:00:00 UTC
Briefing created: 30/09/2026, 04:33:52 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 30/09/2026, 04:18:24 UTC.
Briefings
TabFM is a 400M-parameter foundation model that predicts tabular data via in-context learning, achieving zero-shot performance on 51 benchmark datasets. It outperforms tuned AutoML pipelines and default foundation models on classification and regression tasks. The model was trained on synthetic tables derived from structural causal models, enabling generalizable representations for real-world applications.
Key points
- TabFM is a 400M-parameter foundation model that predicts tabular data using in-context learning, achieving zero-shot performance on 51 benchmark datasets.
- Zero-shot TabFM ranks first among default tabular foundation models and outperforms tuned AutoML pipelines on classification and regression tasks.
Context and limitations
Results are limited to 51 benchmark datasets in TabArena, covering 38 classification and 13 regression tasks. Performance on real-world tasks remains unverified as the model was trained on synthetic data.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 30/09/2026, 04:33:52 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeiebpgo6sfuevbwjpat353ujmqmvc5ullr4tmvee6xrpfnikh65fkq