Tag
MoVE proposes a Mixture-of-LoRA-Experts architecture that preserves laughter and crying in speech-to-speech translation, achieving 76% NV retention with only 30 minutes of curated data.