Khmer Trocr Ocr V1.0

🚀 Khmer TrOCR OCR 📝🇰🇭

This is a fine - tuned model based on [microsoft/trocr - base - stage1](https://huggingface.co/microsoft/trocr - base - stage1), designed for Khmer name recognition. It utilizes synthetic image - text pairs of Khmer personal names to achieve accurate identification of Khmer names.

🚀 Quick Start

Install required packages

pip install transformers torch pillow

Python Inference Example

import torch
from PIL import Image
from transformers import TrOCRProcessor, VisionEncoderDecoderModel

# Load model and processor
model = VisionEncoderDecoderModel.from_pretrained("your_username/khmer-trocr-ocr")
processor = TrOCRProcessor.from_pretrained("your_username/khmer-trocr-ocr")

# Load and process image
image = Image.open("khmer_name_images/khmer_name_00001.png").convert("RGB")
pixel_values = processor(images=image, return_tensors="pt").pixel_values

# Move to GPU if available
device = "cuda" if torch.cuda.is_available() else "cpu"
model = model.to(device)
pixel_values = pixel_values.to(device)

# Generate prediction
generated_ids = model.generate(pixel_values)
predicted_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]

print("🔤 Predicted:", predicted_text)

✨ Features

• Khmer Name Recognition: Specifically designed for recognizing Khmer names.
• Based on TrOCR: Utilizes the pre - trained microsoft/trocr - base - stage1 model for fine - tuning.
• Synthetic Data Training: Trained on a custom - generated synthetic dataset of Khmer names.

📦 Installation

To use this model, you need to install the necessary packages. You can install them using the following command:

pip install transformers torch pillow

💻 Usage Examples

Basic Usage

import torch
from PIL import Image
from transformers import TrOCRProcessor, VisionEncoderDecoderModel

# Load model and processor
model = VisionEncoderDecoderModel.from_pretrained("your_username/khmer-trocr-ocr")
processor = TrOCRProcessor.from_pretrained("your_username/khmer-trocr-ocr")

# Load and process image
image = Image.open("khmer_name_images/khmer_name_00001.png").convert("RGB")
pixel_values = processor(images=image, return_tensors="pt").pixel_values

# Move to GPU if available
device = "cuda" if torch.cuda.is_available() else "cpu"
model = model.to(device)
pixel_values = pixel_values.to(device)

# Generate prediction
generated_ids = model.generate(pixel_values)
predicted_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]

print("🔤 Predicted:", predicted_text)

📚 Documentation

📌 Model Details

🧠 Training

The model was fine - tuned on a synthetic dataset of rendered Khmer names using a Khmer Unicode font (KhmerOS_muol.ttf). Each image is paired with a corresponding text label for supervised training.

• Input: RGB image (512x64) of a Khmer name
• Output: Unicode Khmer text
• Dataset: Custom - generated dataset of Khmer names (10,000+ samples)
• Preprocessing: Images were rendered from text using PIL and paired with their ground - truth labels

📊 Evaluation

Coming soon — evaluation on a labeled test set with CER/WER metrics.

✅ Applications

• Khmer ID OCR

📄 License

The model is released under the Apache - 2.0 license, which allows free use for both research and commercial applications.

🤝 Acknowledgements

• [Microsoft TrOCR](https://huggingface.co/microsoft/trocr - base - stage1)
• Hugging Face Transformers
• Khmer Unicode fonts project