monocr 0.1.17

Optical Character Recognition for Mon text

Mon OCR

Optical Character Recognition for Mon (mnw) text.

Installation

pip install monocr | uv add monocr

Quick Start

Python Usage

from monocr import MonOCR

# Initialize
model = MonOCR()

# 1. Read an Image
text = model.read_text("image.png")
print(text)

# 2. Read with Confidence
result = model.predict_with_confidence("image.png")
print(f"Text: {result['text']}")
print(f"Confidence: {result['confidence']:.2%}")

Examples

See the examples/ folder to learn more.

• examples/run_ocr.py: A complete script that can process a folder of images or read a full PDF book.

CLI Usage

You can also use the command line interface:

# Process a single image
monocr read image.png

# Process a folder of images
monocr batch folder/path

# Manually download the model
monocr download

Related Tools

• mon_tokenizer
• hugging face mon_tokenizer model
• Mon corpus collection in unicode

License

MIT - do whatever you want with it.

Dev Setup

git clone [e-mail]:janakhpon/monocr.git
cd monocr
uv sync --dev

Update Model in Hugging Face

To update the model weights:

hf auth login
hf upload janakhpon/monocr path/to/model_dir --repo-type model

Release Workflow

uv version --bump patch
uv build
git add .
git commit -m "bump version"
git tag v0.1.17
git push origin main --tags

• monocr on pypi
• monocr on hugging face