README.md

---
license: apache-2.0
tags:
- ocr
- rapidocr
- computer-vision
- onnx
---

# RapidOCR Models

This repository contains RapidOCR ONNX model files for optical character recognition.

## Models Included

- `arabic_PP-OCRv3_rec_infer.onnx`
- `ch_PP-OCRv4_det_infer.onnx`
- `ch_PP-OCRv4_rec_infer.onnx`
- `ch_ppocr_mobile_v2.0_cls_infer.onnx`

## Usage

Download these models and use them with RapidOCR:

```python
from rapidocr_onnxruntime import RapidOCR

# Initialize engine with custom model paths
engine = RapidOCR(
    det_model_path='path/to/ch_PP-OCRv4_det_infer.onnx',
    rec_model_path='path/to/ch_PP-OCRv4_rec_infer.onnx',
    cls_model_path='path/to/ch_ppocr_mobile_v2.0_cls_infer.onnx',
)

# Use the engine
result, elapse = engine('path/to/image.jpg')
```

## License

These models are provided under Apache 2.0 license. Please check the original PaddleOCR repository for specific model licenses.