通过压缩与蒸馏，让语音翻译模型变小50%仍保持高精度。
Efficient Speech Translation through Model Compression and Knowledge Distillation
- 用层重要性评估+迭代剪枝，减少模型参数。
- 剪枝后模型体积减半，翻译质量保留97%以上。
- 适合资源有限场景下的高效语音翻译部署。
大型音视频-语言模型在语音翻译中的高效部署仍面临巨大计算挑战。本文针对国际口语语言翻译会议（IWSLT 2025）的“模型压缩”赛道，提出结合迭代层剪枝、4比特量化低秩适配（QLoRA）及知识蒸馏的方案。实验以Qwen2-Audio-7B-Instruct模型为教师模型，进行德语和汉语的语音翻译。剪枝后的学生模型实现高达50%的参数量与存储空间缩减，同时在域内任务中保持97%-100%的翻译质量，显著提升部署效率。
原文摘要 · Abstract (English)
Efficient deployment of large audio-language models for speech translation remains challenging due to their significant computational requirements. In this paper, we address this challenge through our system submissions to the "Model Compression" track at the International Conference on Spoken Language Translation (IWSLT 2025). We experiment with a combination of approaches including iterative layer pruning based on layer importance evaluation, low-rank adaptation with 4-bit quantization (QLoRA), and knowledge distillation. In our experiments, we use Qwen2-Audio-7B-Instruct for speech translation into German and Chinese. Our pruned (student) models achieve up to a 50% reduction in both model parameters and storage footprint, while retaining 97-100% of the translation quality of the in-domain (teacher) models.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。