先微调再结构化剪枝，让语音识别模型更小更快无损失
Fine-tune Before Structured Pruning: Towards Compact and Accurate Self-Supervised Models for Speaker Diarization
- 先微调后剪枝，保留模型精度的同时大幅压缩参数
- 在多个数据集上剪掉80%参数，推理速度提升2.6至4.0倍
- 适合部署在资源受限设备上的高精度语音分离系统
自监督学习模型如WavLM在说话人分离任务中表现优异，但模型庞大且推理慢，限制了其在资源受限场景的应用。现有压缩方法常以性能下降为代价。本文提出一种结合知识蒸馏的结构化剪枝方法，强调剪枝前对SSL模型进行微调的重要性。在远场单通道AMI、AISHELL-4和AliMeeting数据集上的实验表明，该方法可将WavLM Base+和WavLM Large模型冗余参数减少高达80%，且性能无下降。剪枝后，Base+和Large模型在单张GPU上的推理速度分别提升4.0倍和2.6倍。代码已公开。
原文摘要 · Abstract (English)
Self-supervised learning (SSL) models like WavLM can be effectively utilized when building speaker diarization systems but are often large and slow, limiting their use in resource constrained scenarios. Previous studies have explored compression techniques, but usually for the price of degraded performance at high pruning ratios. In this work, we propose to compress SSL models through structured pruning by introducing knowledge distillation. Different from the existing works, we emphasize the importance of fine-tuning SSL models before pruning. Experiments on far-field single-channel AMI, AISHELL-4, and AliMeeting datasets show that our method can remove redundant parameters of WavLM Base+ and WavLM Large by up to 80% without any performance degradation. After pruning, the inference speeds on a single GPU for the Base+ and Large models are 4.0 and 2.6 times faster, respectively. Our source code is publicly available.
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。