README.md

---
license: apache-2.0
base_model: meta-llama/Llama-3.2-11B-Vision-Instruct
tags:
- thai
- handwriting-recognition
- vision-language
- fine-tuned
- vision
datasets:
- iapp/thai_handwriting_dataset
language:
- th
pipeline_tag: image-to-text
---

# Thai Handwriting Recognition Vision-Language Model

A LoRA-adapted vision-language model based on Llama-3.2-11B-Vision-Instruct that transcribes Thai handwritten text from images.

## Model Description
- Base Model: Llama-3.2-11B-Vision-Instruct
- Training Technique: LoRA adaptation
- Quantization: Supports 4-bit inference
- Dataset: iapp/thai_handwriting_dataset

## Demo

Try the model via our web interface:
🔗 [Thai-HandWriting-to-Text](https://huggingface.co/spaces/Aekanun/Thai-HandWriting-to-Text)

### Example Output

![Medical Prescription Recognition](https://video.aekanun.com/Zq4F7xd8)

The model can accurately transcribe complex medical prescriptions, including:
- Medication names and dosages
- Treatment instructions
- Clinical notes

### Features
- Supports both general handwriting and medical prescriptions
- Simple drag-and-drop interface
- Real-time text recognition
- No setup required

### Example Use Cases
1. Medical prescription digitization
2. Clinical document processing
3. General Thai handwriting transcription

## Limitations
- Designed specifically for Thai handwriting
- Performance may vary with image quality
- Requires clear handwriting for best results

## License
This model is released under the Apache 2.0 license.