allenai/olmOCR-2-7B-1025

Note

This is a remotely-sourced model from the olmOCR-2 plugin, maintained by the community. It is not part of FiftyOne core and may have special installation requirements. Please review the plugin documentation and license before use.

olmOCR-2 is an advanced OCR model from AllenAI that uses Qwen2.5-VL architecture for document text extraction. Returns markdown output with YAML front matter containing document metadata (language, rotation, tables, diagrams). Converts equations to LaTeX and tables to HTML..

Details

• Model name: allenai/olmOCR-2-7B-1025

• Model source: https://huggingface.co/allenai/olmOCR-2-7B-1025

• Model author: AllenAI

• Model license: Apache 2.0

• Exposes embeddings? no

• Tags: ocr, document-understanding, VLM, vision-language

Requirements

• Packages: huggingface-hub, transformers, torch, Pillow, numpy

• CPU support

  • yes

• GPU support

  • yes

Example usage

import fiftyone as fo
import fiftyone.zoo as foz

foz.register_zoo_model_source("https://github.com/harpreetsahota204/olmOCR-2")

dataset = foz.load_zoo_dataset(
    "coco-2017",
    split="validation",
    dataset_name=fo.get_default_dataset_name(),
    max_samples=50,
    shuffle=True,
)

model = foz.load_zoo_model("allenai/olmOCR-2-7B-1025")

dataset.apply_model(model, label_field="predictions")

session = fo.launch_app(dataset)