Datasets:
language:
- roh
license: cc-by-4.0
task_categories:
- translation
configs:
- config_name: monolingual
data_files: monolingual/*.jsonl.gz
- config_name: poly_aligned
data_files: polylingual/Aligned/*.jsonl.gz
- config_name: poly_non_aligned
data_files: polylingual/Non_Aligned/*.jsonl.gz
- config_name: synthetic
data_files: Synthetic/**/*.jsonl.gz
This dataset consist of three differnt parts. Monolingual Romansh Data, Polylingual data or more precisely translated data from Romansh into either German, French, Italian or English and Sythetic Data.
The Polylingual data consists of aligned and non aligned data. The synthetic data was created by interweaving the translational data and prefacing it with the sentence " This is a text translated from SOURCE LANGUAGE to Rumantsch Grischun".
The data has a metadata "idiom" if the if specific idiom was provided otherwise it is implicitly assumed that the data is in Romansh Grischun.
The data constists of:
- Law Texts and announcements from the Muncipalities of: Sagogn (Sursilvan), Lantsch (Surmiran), Zernez (Vallader), Ilanz (Sursilvan)
- Law Texts from the Canton of Grisons in Rumantsch Grischun (https://www.gr-lex.gr.ch/app/rm/systematic/texts_of_law)
- The Bilingual Corpus on GitHub (https://github.com/ZurichNLP/RumantschCorpora/tree/master)
- Online Dictionaries from the Lia Rumantscha in Surmiran, Sutsilvan and Sursilvan (www.pledarigrond.ch)
- Romansh Websites on Wikipedia
Below I give a token count using alehc/swissai-tokenizer. Mixed languages like 'de/roh' refer to the fact that either the transational text was not aligned or that it is synthetic data using the two languages. Note that the synthetic data token count is inflated the above-mentioned prefix.
Note that all data has been preprocessed using the pipeline in https://github.com/swiss-ai/Swiss-AI-Romansh-Scripts.
Please feel free to contact me if you have any comments regarding the data ([email protected]).