--- base_model: - Qwen/Qwen2.5-VL-7B-Instruct datasets: - allenai/olmOCR-mix-0225 language: - en library_name: transformers license: apache-2.0 new_version: allenai/olmOCR-7B-0825-FP8 pipeline_tag: image-to-text tags: - ocr - qwen --- olmOCR Logo # olmOCR-7B-0725-FP8 This is a quantized (FP8) version of [olmOCR-7B-0725](https://huggingface.co/allenai/olmOCR-7B-0725), using llmcompressor. This model is part of the olmOCR family of powerful OCR systems for converting digitized print documents, like PDFs, into clean, naturally ordered plain text. The latest advancements in this family are presented in the paper [olmOCR 2: Unit Test Rewards for Document OCR](https://huggingface.co/papers/2510.19817). This olmOCR model is fine-tuned from Qwen2.5-VL-7B-Instruct using the [olmOCR-mix-0225](https://huggingface.co/datasets/allenai/olmOCR-mix-0225) dataset. Quick links: - 📃 [Paper](https://huggingface.co/papers/2510.19817) - 🤗 [Dataset](https://huggingface.co/datasets/allenai/olmOCR-mix-0225) - 🛠️ [Code](https://github.com/allenai/olmocr) - 🎮 [Demo](https://olmocr.allenai.org/) The best way to use this model is via the [olmOCR toolkit](https://github.com/allenai/olmocr). The toolkit comes with an efficient inference setup via sglang that can handle millions of documents at scale. ## Usage This model expects as input a single document image, rendered such that the longest dimension is 1288 pixels. The prompt must then contain the additional metadata from the document, and the easiest way to generate this is to use the methods provided by the [olmOCR toolkit](https://github.com/allenai/olmocr). ## Sample Usage For quick testing, try the [web demo](https://olmocr.allen.ai/). To run locally, a GPU is required. Here's an example of converting a single PDF using the `olmocr` pipeline, as provided in the GitHub repository: ```bash # Download a sample PDF curl -o olmocr-sample.pdf https://olmocr.allenai.org/papers/olmocr_3pg_sample.pdf # Convert it to markdown python -m olmocr.pipeline ./localworkspace --markdown --pdfs olmocr-sample.pdf ``` Results will be stored as markdown files inside of `./localworkspace/markdown/`. ## License and use olmOCR is licensed under the Apache 2.0 license. olmOCR is intended for research and educational use. For more information, please see our [Responsible Use Guidelines](https://allenai.org/responsible-use).