Companion repo for A 2×2 Controlled Ablation of Data Quality and Capacity in Transformer MT. Contains the QE-filtered fine-tuning pipeline (Section 7), CometKiwi-22 scoring/filtering scripts, and the full paper LaTeX source. Pretraining lives in Machine_translation
nlp machine-translation pytorch transformer neural-machine-translation mt arxiv reproducibility quality-estimation data-quality sentencepiece ablation-study model-scaling sacrebleu wmt14 cometkiwi
-
Updated
May 4, 2026 - TeX