README.md

---
pretty_name: ocr_data
task_categories:
- image-to-text
language:
- ar
tags:
- ocr
- arabic
- synthetic
- webdataset
---

# ocr_data

Synthetic Arabic document images with layout annotations, for OCR training.

## Layout

WebDataset `.tar` shards. Files sharing a basename are one sample, so the image
becomes the `png`/`jpg` column and the annotation the `json` column.

```
data/shard_001.tar ... data/shard_103.tar                 originals (PNG + JSON)
data_aug/shard_001_aug1.tar ... data_aug/shard_103_aug3.tar   augmented variants (JPEG + JSON)
```

Each shard holds up to 9990 samples (~1 GB). `data/` and `data_aug/` are separate
so you can train on clean originals alone.

## Loading

```python
from datasets import load_dataset

# one shard
ds = load_dataset("webdataset",
                  data_files="hf://datasets/OCR-Data/ocr_data/data/shard_001.tar",
                  split="train", streaming=True)

# a range of shards
ds = load_dataset("webdataset",
                  data_files="hf://datasets/OCR-Data/ocr_data/data/shard_{001..010}.tar",
                  split="train", streaming=True)

# everything, originals + augmented
ds = load_dataset("webdataset", data_files={"train": [
        "hf://datasets/OCR-Data/ocr_data/data/*.tar",
        "hf://datasets/OCR-Data/ocr_data/data_aug/*.tar"]},
      split="train", streaming=True)
```

`meta.augmentation` in each annotation is `null` for originals and
`{"name": ..., "params": {...}}` for augmented variants.