 Corpus Name: NeuLab-TedTalks
     Package: NeuLab-TedTalks.en-vi in Moses format
     Website: http://opus.nlpl.eu/NeuLab-TedTalks-v1.php
     Release: v1
Release date: Wed Jan 18 12:27:33 EET 2023
   Copyright: Please respect the <a href="https://www.ted.com/about/our-organization/our-policies-terms">TED Talks Usage Policy</a>

This package is part of OPUS - the open collection of parallel corpora
OPUS Website: http://opus.nlpl.eu

Please <a href="http://opus.lingfil.uu.se/LREC2012.txt">cite the following article</a> if you use the OPUS packages and downloads in your own work:<br/> J. Tiedemann, 2012, <a href="http://www.lrec-conf.org/proceedings/lrec2012/pdf/463_Paper.pdf"><i>Parallel Data, Tools and Interfaces in OPUS.</i></a> In Proceedings of the 8th International Conference on Language Resources and Evaluation (LREC 2012)<br/>

This dataset is compiled from TED talk subtitles and distributed through phontron.com. The package here includes the training data only (development and test data are not included in this package). The transcripts have been translated by a global community of volunteers to more than 100 languages. The parallel corpus and the code fopr cerating it is available from https://www.ted.com/participate/translate
Note that this corpus is tokenized in its original form. For this package we applied automatic de-tokenization using the moses tools. This is not perfect and may miss some de-tokenization steps.
