TableSuite-1K is a source-grounded benchmark for tabular prediction, cell
grounding, and table question answering over 1,000 OpenML-referenced datasets.
This repository stores compact task specifications, not copied source tables,
rendered answers, model outputs, or checkpoints.
Configuration
Records
Unit
datasets
1,000
source, schema, target, split, and license metadata
table_prediction_tasks
998
prediction-eligible… See the full description on the dataset page:
https://huggingface.co/datasets/Lester1996/TableSuite-1K.
