Datasets:
Add Turkish Text Normalization (TN/ITN) dataset: 17k rule-based pairs + reproducible pipeline, validation, dataset card
5c0e24f verified | { | |
| "total": 17109, | |
| "train": 15398, | |
| "test": 1711, | |
| "by_category": { | |
| "cardinal": 4259, | |
| "currency": 3051, | |
| "date": 3082, | |
| "decimal": 3497, | |
| "ordinal": 1683, | |
| "percentage": 436, | |
| "time": 1101 | |
| }, | |
| "seed": 42 | |
| } |