---
language:
- az
license: apache-2.0
pipeline_tag: image-to-text
tags:
- ocr
- vision-language
- markdown
- azerbaijani
library_name: transformers
---

# AZOCR

Azerbaijani OCR model based on GLM-OCR.

## Training

- Base model: GLM-OCR
- Language: Azerbaijani
- Dataset: 6699 page samples
- Method: LoRA
- GPUs: 2× NVIDIA Tesla T4
- Epochs: 2

## Notes

Fine-tuned for Azerbaijani document OCR.

## How to use?

https://github.com/orucexe/azerbaijani-ocr