Receipt-KIE-200 is a compact, high-quality dataset designed for Key Information Extraction (KIE) from receipt and invoice images. It focuses on structured extraction of textual and semantic information using OCR + reasoning-based formatting in Markdown. This dataset is suitable for training and evaluating multimodal models, especially Vision-Language Models (VLMs) that perform document understanding and structured text generation.… See the full description on the dataset page:
https://huggingface.co/datasets/prithivMLmods/Receipt-KIE-200.