CVSS

CVSS is a massively multilingual-to-English speech-to-speech translation corpus, covering sentence-level parallel speech-to-speech translation pairs from 21 languages into English.

Papers using CVSS (11)

Scheduled Interleaved Speech-Text Training for Speech-to-Speech Translation with LLMs2025 · 1 citesBenchmarking Speech-to-Speech Translation Models2026Improving Direct Persian-English Speech-to-Speech Translation with Discrete Units and Synthetic Parallel Data2025SimulS2S-LLM: Unlocking Simultaneous Inference of Speech LLMs for Speech-to-Speech Translation2025Direct Text to Speech Translation System using Acoustic Units2023 · 10 citesSimple and Effective Unsupervised Speech Translation2022 · 2 citesImproving Cascaded Unsupervised Speech Translation with Denoising Back-translation2023 · 2 citesDiffNorm: Self-Supervised Normalization for Non-autoregressive Speech-to-speech Translation2024StreamSpeech: Simultaneous Speech-to-Speech Translation with Multi-task Learning2024Can We Achieve High-quality Direct Speech-to-Speech Translation without Parallel Speech Data?2024CVSS Corpus and Massively Multilingual Speech-to-Speech Translation2022