Machine Translation (MT) Evaluation Scripts
-
Updated
May 19, 2024 - Python
#
mt-evaluation
Here are 4 public repositories matching this topic...
The Suboptimal Quality of WMT Test Sets and Their Impact on HumanParity
-
Updated
May 13, 2024 - Python
TamilLingBench: English→Tamil machine translation benchmark for obligatory verb agreement (rationality, gender, number). Tests whether MT evaluation metrics such as COMET, xCOMET, MetricX and CometKiwi catch wrong Tamil agreement. Code, data and model outputs for the WMT 2026 paper "Obligatory Slots".
nlp benchmark translation morphology machine-translation comet computational-linguistics tamil quality-estimation finite-state-transducer evaluation-metrics interpretability wmt mt-evaluation dravidian-languages llm-evaluation activation-patching metricx challenge-set english-tamil
-
Updated
Sep 12, 2026 - Python
EN-UK parallel corpus pipeline built from Baldur's Gate 3 localization files: 186k aligned pairs, 7 text types, Gemini API translation and MT evaluation (BLEU, METEOR, TER, ChrF++, BERTScore). BA thesis project.
nlp machine-translation ukrainian corpus-linguistics parallel-corpus gemini-api mt-evaluation baldurs-gate-3
-
Updated
Sep 22, 2026 - Python
Add this topic to your repo
To associate your repository with the mt-evaluation topic, visit your repo's landing page and select "manage topics."