File size: 5,143 Bytes

---
license: apache-2.0
library_name: onnx
pipeline_tag: image-to-text
tags:
- ocr
- paddleocr
- onnx
- computer-vision
- text-detection
- text-recognition
- multilingual
datasets:
- community-datasets
language:
- en
- zh
- multilingual
widget:
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/tiger.jpg
  example_title: Sample OCR
---

# PaddleOCR ONNX Models

🔥 **ONNX format models converted from PaddleOCR for easy deployment and testing**

## 📖 Model Description

This repository contains ONNX format models converted from [PaddleOCR](https://github.com/PaddlePaddle/PaddleOCR), a practical ultra-lightweight OCR system. These models are optimized for production deployment and cross-platform compatibility.

## 📦 Model Files

| File Name | Size | Description |
|-----------|------|-------------|
| `PP-OCRv5_server_det_infer.onnx` | 84MB | Text detection model - locates text regions in images |
| `PP-OCRv5_server_rec_infer.onnx` | 81MB | Text recognition model - recognizes text content |
| `UVDoc_infer.onnx` | 30MB | Document rectification model - corrects document perspective |
| `PP-LCNet_x1_0_doc_ori_infer.onnx` | 6.5MB | Document orientation detection |
| `PP-LCNet_x1_0_textline_ori_infer.onnx` | 6.5MB | Text line orientation detection |
| `PP-OCRv5_server_rec_infer.yml` | 145KB | Recognition model configuration file |

**Total Size:** ~208MB

## 🚀 Quick Start

### Installation

```bash
pip install huggingface_hub onnxruntime
```

### Download Models

```python
from huggingface_hub import hf_hub_download
import os

def download_paddleocr_models():
    """Download all PaddleOCR ONNX models"""
    model_files = [
        "PP-OCRv5_server_det_infer.onnx",
        "PP-OCRv5_server_rec_infer.onnx", 
        "UVDoc_infer.onnx",
        "PP-LCNet_x1_0_doc_ori_infer.onnx",
        "PP-LCNet_x1_0_textline_ori_infer.onnx",
        "PP-OCRv5_server_rec_infer.yml"
    ]
    
    cache_dir = "models"
    os.makedirs(cache_dir, exist_ok=True)
    
    for file in model_files:
        print(f"Downloading {file}...")
        hf_hub_download(
            repo_id="marsena/paddleocr-test",
            filename=file,
            local_dir=cache_dir
        )
    print("All models downloaded!")

# Download models
download_paddleocr_models()
```

### Basic Usage

```python
import onnxruntime as ort
import numpy as np
from PIL import Image

# Load detection model
det_session = ort.InferenceSession("models/PP-OCRv5_server_det_infer.onnx")

# Load recognition model  
rec_session = ort.InferenceSession("models/PP-OCRv5_server_rec_infer.onnx")

# Your OCR pipeline implementation here...
```

## 🏷️ Model Tags

- **Framework:** ONNX
- **Task:** Computer Vision, OCR
- **Language:** Multi-language support
- **Domain:** Text Detection, Text Recognition

## 🔧 Technical Details

### Conversion Process

These models were converted from PaddlePaddle format to ONNX format for broader compatibility:

1. **Source:** Original PaddleOCR models from PaddlePaddle Hub
2. **Conversion:** PaddlePaddle → ONNX format
3. **Optimization:** Model optimization for inference speed
4. **Validation:** Output consistency verification

### System Requirements

- **Runtime:** ONNX Runtime
- **Python:** 3.7+
- **Memory:** Minimum 2GB RAM recommended
- **Platform:** Cross-platform (Windows, Linux, macOS)

## 📄 License

This project follows the **Apache 2.0 License**, consistent with the original PaddleOCR project.

### Original PaddleOCR License

```
Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
```

## 🙏 Acknowledgments

- **Original Project:** [PaddleOCR](https://github.com/PaddlePaddle/PaddleOCR) by PaddlePaddle Team
- **Framework:** [PaddlePaddle](https://github.com/PaddlePaddle/Paddle)
- **Conversion Tools:** ONNX ecosystem

## 📚 Citation

If you use these models in your research, please cite the original PaddleOCR paper:

```bibtex
@misc{paddleocr2020,
    title={PaddleOCR: Awesome multilingual OCR toolkits},
    author={PaddlePaddle Authors},
    year={2020},
    howpublished={\url{https://github.com/PaddlePaddle/PaddleOCR}}
}
```

## ❓ Issues & Support

For issues related to:
- **Model conversion:** Create an issue in this repository
- **Original PaddleOCR:** Visit [PaddleOCR Issues](https://github.com/PaddlePaddle/PaddleOCR/issues)
- **ONNX Runtime:** Visit [ONNX Runtime Issues](https://github.com/microsoft/onnxruntime/issues)

---

**Note:** This is a community contribution for easier deployment of PaddleOCR models. For production use, please ensure compliance with your specific requirements and test thoroughly.