Dataset Description

TabFQuAD (Table French Question Answering Dataset) is designed to evaluate TableQA models in realistic industry settings. Using a vision language model (GPT4V), we create additional queries to augment the existing human-annotated ones.

Data Curation

To ensure homogeneity across our benchmarked datasets, we subsampled the original test set to 280 pairs, leaving the rest for training and renaming the different columns.

Load the dataset

from datasets import load_dataset
ds = load_dataset("vidore/tabfquad_test_subsampled", split="test")

Dataset Structure

Here is an example of a dataset instance structure:

features:
  - name: query
    dtype: string
  - name: image_filename
    dtype: string
  - name: generated_by
    dtype: string
  - name: GPT4 caption
    dtype: string
  - name: image
    dtype: image
  - name: source
    dtype: string