106 datasets found

The COMET dataset contains human-annotated scores for machine translation candidates.

The WMT2020 Metrics Shared Task dataset contains human-annotated scores for machine translation candidates.

RoBLEURT is a robustly optimizing the training of BLEURT, a trainable metric model for evaluating the semantic consistency between machine translation candidates and golden...

The Umsuka English-isiZulu Parallel Corpus provides a novel, high-quality parallel dataset for machine translation, containing English sentences sampled from both News Crawl...

The dataset used in the paper is the WAT2015 translation task from Japanese (ja) to/from English (en) and Chinese (zh).

Two types of datasets: traditional bilingual and domain adaptation datasets.

The dataset used in the competition for Sentence-Level Direct Assessment shared task is composed of data extracted from Wikipedia for six language pairs, consisting of...

The dataset used for training and testing the machine translation systems.

The dataset used for training and testing the Naive Bayes classifier for quality estimation of English-Hindi outputs.

Newstest2012 and Newstest2013 are used for testing the proposed approach.

The dataset used in this paper is the WMT2021 shared task on machine translation using terminologies, which consists of 4.53M sentence pairs.

The dataset used in this paper is the WMT2014 German-English translation task, which consists of 4.51M parallel sentence pairs.

The Shifts Machine Translation dataset consists of pairs of source and target sentences in English and Russian.

The ParCor dataset is a parallel corpus of annotated pronouns.

The WIT3 parallel corpus is a large-scale corpus of transcribed and translated talks.

The WMT 2016 task on cross-lingual pronoun prediction is a classification task in which participants are asked to provide predictions on what pronoun class label should replace a...

The dataset used for building the NMT model, which is a German-to-English parallel corpus.

The dataset used in the paper is the WMT dataset, which contains machine translation data for various language pairs.

The dataset used in the paper for validation studies of automatic metrics in natural language generation evaluation

A neural machine translation toolkit that uses maximum likelihood as the training criterion.