Reproduce “Index-Translate: A Multilingual Translation Model Family -- Text, Speech, Controlled Dubbing, and Long-Document Translation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Tianjiao Li, Mengran Yu, Chenyu Shi, Lusheng Zhang, Qisi Chen, Yanshan Zhou, Ji Qi, Jingying Liu and 3 more
The authors present Index-Translate as a multilingual model family that shares a translation foundation and adapts it for speech, length-constrained output, and long documents. They train general, instruction, and cultural-translation specialists, with task-specific systems for dubbing and syllable control. On GuoFeng at 64K, Index-NativeLong-9B reaches document-SEGALE/COMET 0.7382, versus 0.4939 for Qwen3.8 Flash; this evaluated setting does not establish universal book-translation superiority.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.