Datasets for Finnish OCR tasks. The aida-xxx datasets are generated from the AIDA_ocr_training_data in parquet format.