arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2511.08093eess.AScs.CLcs.SD
收藏

对Whisper-small的量化:设计选择如何影响语音识别性能

Quantizing Whisper-small: How design choices affect ASR performance

Copenhagen Business School(哥本哈根商学院);Danske Bank(丹麦银行);Jabra (GN Group)(Jabra(GN集团))

Arthur Söhler, Julian Irigoyen, Andreas Søeborg Kirkedal

AI总结:

本文研究了不同量化方案对Whisper-small模型性能的影响,发现动态int8量化在模型压缩和识别准确率之间取得了最佳平衡,同时展示了通过精心选择量化方法可以显著减少模型大小和推理成本,从而在受限硬件上实现高效部署。

摘要:

大型语音识别模型如Whisper-small虽然能实现高精度,但其高计算需求使其难以在边缘设备上部署。为此,我们提出了一种统一的跨库评估,评估了Whisper-small上的后训练量化(PTQ)方法,以分离量化方案、方法、粒度和位宽的影响。我们的研究基于四个库:PyTorch、Optimum-Quanto、HQQ和bitsandbytes。在LibriSpeech测试清洁和测试其他数据集上的实验表明,动态int8量化结合Quanto提供了最佳的权衡,将模型大小减少57%,同时在基线的词错误率上有所提升。静态量化表现较差,可能由于Whisper的Transformer架构,而更激进的格式(如nf4、int3)在嘈杂条件下以牺牲准确性为代价实现了高达71%的压缩。总体而言,我们的结果表明,精心选择的PTQ方法可以在不重新训练的情况下显著减少模型大小和推理成本,从而在受限硬件上实现Whisper-small的高效部署。

英文摘要:

Large speech recognition models like Whisper-small achieve high accuracy but are difficult to deploy on edge devices due to their high computational demand. To this end, we present a unified, cross-library evaluation of post-training quantization (PTQ) on Whisper-small that disentangles the impact of quantization scheme, method, granularity, and bit-width. Our study is based on four libraries: PyTorch, Optimum-Quanto, HQQ, and bitsandbytes. Experiments on LibriSpeech test-clean and test-other show that dynamic int8 quantization with Quanto offers the best trade-off, reducing model size by 57% while improving on the baseline's word error rate. Static quantization performed worse, likely due to Whisper's Transformer architecture, while more aggressive formats (e.g., nf4, int3) achieved up to 71% compression at the cost of accuracy in noisy conditions. Overall, our results demonstrate that carefully chosen PTQ methods can substantially reduce model size and inference cost without retraining, enabling efficient deployment of Whisper-small on constrained hardware.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑