razisayyed/

expense-extraction-dataset

A multilingual intent-classification dataset containing Arabic and English student questions and requests.

The dataset is divided into three splits:

Train: approximately 80% Validation: approximately 10% Test: approximately 10%

The splits were generated deterministically using a random seed of 42.