arXiv cs.CL· Tsz-To Wong, Shyang-En Weng·· 13 小时前AI 评分30
面向浏览器扩展的 CPU 端字幕翻译：Sotto-0.6B 与 Sotto-1.7B 的部署驱动训练与延迟设计
Context-Aware Subtitle Translation on CPU: Deployment-Driven Training, Evaluation, and Latency Design
AI 导读
基于 LMT-60 训练的 Sotto-1.7B 和已部署的 Sotto-0.6B 用于英译繁中（台湾）字幕，在 1,820 条留出字幕上对 Google Translate 胜率 0.576，并正面击败四个开源翻译模型，略低于一个 27B 模型。
正文 · 原文
Abstract:We study English-to-Traditional-Chinese (Taiwan) subtitle translation for a free browser extension, where the model must run on the viewer's CPU in about 500 ms per subtitle. We train two models built on the LMT-60 base models, Sotto-1.7B and the deployed Sotto-0.6B, with a task-specific vocabulary, continued pretraining, supervised fine-tuning, weight interpolation, logit distillation and on-policy preference optimization. Because reference translations of subtitles are unreliable, we evaluate by pairwise comparison against one fixed baseline, Google Translate, with each pair judged in both orders by a panel of two LLM judges. On 1,820 held-out subtitles, Sotto-1.7B has a win rate of 0.576 against the baseline and wins head-to-head against four open translation models, and is slightly below a 27B model. We split latency into prefill and decoding: idle-time prefill, which prefills the known part of the prompt while waiting for the next subtitle, cuts mean end-to-end latency by 31% on a CPU, and the task-specific vocabulary encodes the same Chinese text in 28% fewer tokens. The deployed model stays within 500 ms for every subtitle in our latency test. The code is available at this https URL .
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2607.09957 [cs.CL] |
| (or arXiv:2607.09957v3 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2607.09957 arXiv-issued DOI via DataCite |
Submission history
From: Tsz-To Wong [view email]
[v1]
Fri, 10 Jul 2026 20:22:13 UTC (12 KB)
[v2]
Thu, 16 Jul 2026 17:45:45 UTC (1,004 KB)
[v3]
Thu, 22 Oct 2026 13:39:31 UTC (1,184 KB)
来源：arXiv cs.CL · arxiv.org