# Kashmiri OCR Dataset

## Dataset Description
This dataset contains 5,000 samples of Kashmiri text images for OCR training.

## Dataset Structure
- **Total samples**: 5,000
- **Image dimensions**: 1024x128
- **DPI**: 300
- **Format**: PNG images with UTF-8 text annotations

## Usage
```python
from datasets import load_from_disk
dataset = load_from_disk('kashmiri_hf_dataset')
```

## Generation Details
- Source: CSV file with 46,542 Kashmiri sentences
- Augmentation: Enabled
- Generated on: 2025-07-30 15:02:40