---
license: mit
language:
- en
tags:
- patents
- claims
- legal
- gpt-training
task_categories:
- text-generation
pretty_name: Granted EP Patent Independent Claims
size_categories:
- 100K<n<1M
---

# Granted EP Patent Independent Claims

One claim per line. Optional `<EOS>` and `<IDX>` placeholders.

## Splits
- Train: 90%
- Validation: 5%
- Test: 5%

## Usage
```python
from datasets import load_dataset
ds = load_dataset("mhurhangee/patent-ind-claim-en")
```