---
license: mit
task_categories:
- image-to-text
language:
- en
tags:
- latex
- math
- formulas
- hard-2line
size_categories:
- 10K<n<100K
---

# LaTeX Hard 2-Line Dataset (Split)

Dataset ini adalah subset dari [OleehyO/latex-formulas](https://huggingface.co/datasets/OleehyO/latex-formulas) yang telah dikurasi untuk formula LaTeX "hard" dengan 2+ baris.

## Dataset Details

- **Source Dataset**: `OleehyO/latex-formulas` (config: `cleaned_formulas`)
- **Subset**: `hard_2line` - Formula dengan 2 atau lebih baris (mengandung `\\`)
- **Total Samples**: 99,305
- **Seed**: 42

## Splits

| Split | Samples | Percentage |
|-------|---------|------------|
| Train | 69,513 | 70% |
| Validation | 19,861 | 20% |
| Test | 9,931 | 10% |

## Usage

```python
from datasets import load_dataset

# Load all splits
dataset = load_dataset("wulanbhai/latex-hard-2line-split")

# Load specific split
train_data = load_dataset("wulanbhai/latex-hard-2line-split", split="train")
val_data = load_dataset("wulanbhai/latex-hard-2line-split", split="validation")
test_data = load_dataset("wulanbhai/latex-hard-2line-split", split="test")
```

## Curation Process

Dataset ini dikurasi menggunakan regex pattern untuk mendeteksi formula multi-baris:
- Pattern: `(?<!\\)\\\\(?![A-Za-z])` - break baris nyata (bukan escaped)
- Environment blocks: matrix, align, gather, dll.

## Generation Details

- **Generated**: 2025-10-28T04:42:49+00:00
- **Curation script**: [latex-hard-2line-matrix-indices](https://huggingface.co/datasets/wulanbhai/latex-hard-2line-matrix-indices)

## License

Indices only - no images or original LaTeX content redistributed. Refer to source dataset for licensing.