lightonai/LightOnOCR-2-1B

Note

This is a remotely-sourced model from the LightOnOCR-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.

LightOnOCR is a 2.1B-parameter vision-language model optimized for optical character recognition. It uses a chat-based interface to extract text from images with high accuracy across various document types, handwritten text, and scene text..

Details

• Model name: lightonai/LightOnOCR-2-1B

• Model source: https://huggingface.co/lightonai/LightOnOCR-2-1B

• Model author: LightOn AI

• Model license: Apache-2.0

• Exposes embeddings? no

• Tags: ocr, text-extraction, VLM, document-understanding

Requirements

• Packages: huggingface-hub, transformers, torch, torchvision, pillow

• 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/LightOnOCR-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("lightonai/LightOnOCR-2-1B")

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

session = fo.launch_app(dataset)