Datasets:

bilirkesi
/

osmanlica-bench-v1

name string	license string	description string	size string	split dict	metrics dict	sources list
Osmanlica-Bench-v1	mit	Benchmark dataset for Ottoman Turkish transliteration	6500 samples	{ "train": 5000, "validation": 500, "test": 1000 }	{ "CER": "6.46%", "WER": "20.69%", "BLEU": "77.18" }	[ "Servet-i Funun (1896-1901)", "Ruznamçe Registers (17th century)", "TBMM Proceedings (1920s)", "OTC Corpus samples" ]

Osmanlica-Bench-v1

A benchmark dataset for Ottoman Turkish transliteration systems.

from datasets import load_dataset

dataset = load_dataset("bilirkesi/osmanlica-bench-v1")
print(dataset["test"][0])

pip install datasets