多语言语音识别系统在11种语言上达到9.83%错误率，排名全球第三。
Transsion Multilingual Speech Recognition System for MLC-SLM 2025 Challenge
- 用冻结的Whisper大模型提取声学特征，搭配可训练适配器对齐声纹与文本。
- 集成冻结Qwen大模型+可训练LoRA，实现上下文感知的精准解码。
- 适合多语言语音识别、端到端模型优化等场景研究者参考。
本文介绍由传音语音团队为MLC-SLM 2025挑战赛第1赛道设计的新型多语言自动语音识别（ASR）系统架构与性能表现。该系统包含三个核心组件：1）基于冻结的Whisper-large-v3语音编码器，利用大规模预训练确保鲁棒的声学特征提取；2）采用线性-激活-线性变换机制的可训练适配模块，有效对齐语音与文本表示；3）集成冻结的Qwen2.5-7B-Instruct大语言模型（LLM）并结合可训练的LoRA进行优化的上下文语言解码。通过系统性融合预训练模型与任务特定微调，该系统在评估集11种语言上实现了9.83%的词错误率（WER）和字符错误率（CER），在全球参赛者中位列第三。
原文摘要 · Abstract (English)
This paper presents the architecture and performance of a novel Multilingual Automatic Speech Recognition (ASR) system developed by the Transsion Speech Team for Track 1 of the MLC-SLM 2025 Challenge. The proposed system comprises three key components: 1) a frozen Whisper-large-v3 based speech encoder, leveraging large-scale pretraining to ensure robust acoustic feature extraction; 2) a trainable adaptor module using Linear-ReLU-Linear transformation mechanisms to effectively align speech and text representations; and 3) a frozen Qwen2.5-7B-Instruct large language model (LLM) integrated with trainable LoRA for optimized contextual linguistic decoding. By systematically combining pretrained models with task specific fine-tuning, the system achieved a word/character error rate (WER/CER) of 9.83% across 11 languages in the evaluation set and ranked third place among global participants.
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。