Crnn Base Fa V1

🚀 Hezar - Persian OCR CRNN Model

Hezar's CRNN model is designed for Persian OCR, offering a practical solution for extracting text from scanned Persian documents. It is based on a CNN + LSTM architecture inspired by this paper.

🚀 Quick Start

To get started with the model, follow these steps:

Installation

pip install hezar

Usage Examples

Basic Usage

from hezar.models import Model

crnn = Model.load("hezarai/crnn-base-fa-v1")
texts = crnn.predict(["sample_image.jpg"])
print(texts)

✨ Features

• Persian OCR: Specifically optimized for Persian scanned documents.
• Fine - Tunable: Can be fine - tuned on other domains such as license plates or handwritten texts.

📚 Documentation

Limitations

This model is tailored for the Persian alphabet and has limited ability to recognize numbers and digits. The development team plans to retrain the model to cover all scenarios in the near future.

Important Note

⚠️ Important Note

This model is only optimized for scanned documents and supports input characters of up to 32. For an end - to - end OCR pipeline, it is recommended to use a text detector model first to extract text boxes, preferably at the word level, and then use this model.

Usage Tip

💡 Usage Tip

Although the model is currently optimized for scanned documents, it can be fine - tuned on other domains like license plate or handwritten texts.

📦 Installation

pip install hezar

💻 Usage Examples

Basic Usage

from hezar.models import Model

crnn = Model.load("hezarai/crnn-base-fa-v1")
texts = crnn.predict(["sample_image.jpg"])
print(texts)

🔧 Technical Details

The model is based on a simple CNN + LSTM architecture inspired by this paper. More information about the data and training will be provided soon.