caveman273/swe-11k

SWE-11K This dataset is generated from https://github.com/sdrobac/ijdar-2020. Quick Start from datasets import load_dataset dataset = load_dataset("parquet", data_files="train.parquet", split="train") Dataset Description SWE-11K is a Swedish OCR dataset containing 11,094 image-text pairs extracted from historic Swedish newspapers and journals. Suitable for training and evaluating OCR models on Swedish historical text. Supported Tasks… See the full description on the dataset page: https://huggingface.co/datasets/caveman273/swe-11k.