It was formed based on 1,200,000 court cases. The data has been preliminarily normalized and split into sentences. The dataset consists of 37 million sentences and approximately 500-600 million tokens.

Dataset Structure

Each row represents a single sentence extracted from a court case document.

Column Type Description

case_id… See the full description on the dataset page: https://huggingface.co/datasets/LocalDoc/court_cases_azerbaijani.