olmOCR-mix-0225 是一个大规模、高质量的 PDF 文档数据集，专为训练和优化光学字符识别 (OCR) 模型而设计。该数据集由 Allen Institute for AI 于 2025 年发布，相关论文成果为「olmOCR: Unlocking Trillions of Tokens in PDFs with Vision Language Models」。
# olmOCR-mix-0225 olmOCR-mix-0225 is a dataset of ~250,000 PDF pages which have been OCRed into plain-text in a natural reading order using gpt-4o-2024-08-06 and a special prompting strategy that pre
This is just 10,000 PDFs randomly sampled from https://huggingface.co/datasets/allenai/olmOCR-mix-0225 that we use internally at AI2 for speed benchmarking and also quantization calibration.
