--- dataset_info: features: - name: incorrect dtype: string - name: correct dtype: string splits: - name: train num_bytes: 193960468 num_examples: 540000 - name: test num_bytes: 21546211 num_examples: 60000 download_size: 129134550 dataset_size: 215506679 configs: - config_name: default data_files: - split: train path: data/train-* - split: test path: data/test-* language: - ur --- This dataset contains paired Urdu text for error correction tasks. ## Data Format Each dialogue entry in the dataset is structured as follows: - `input_text`: The text containing simulated errors. - `correct_text`: The corresponding correct text.