metythorn/ocr-stn-cnn-transformer-base

Khmer OCR CNN + Transformer

This repository contains a ResNet + Transformer decoder checkpoint for Khmer OCR, I don’t have a public paper for this model — everything comes from thousands of experiments across different model architectures and datasets.

Installation

pip install mer

Usage

from mer import Mer

model = Mer(markdown=True, device='cuda')
result = model.predict("sample_image.png")
print("Predicted text:", result)