Self-supervised speech representation learning has shown promising results in various speech processing tasks. However, the pre-trained models, e.g., HuBERT, are storage-intensive Transformers, limiting their scope of applications under low-resource settings. To this end, we propose LightHuBERT, a once-for-all Transformer compression framework, to find the desired architectures automatically by pruning structured parameters. More precisely, we create a Transformer-based supernet that is nested with thousands of weight-sharing subnets and design a two-stage distillation strategy to leverage the contextualized latent representations from HuBERT. Experiments on automatic speech recognition (ASR) and the SUPERB benchmark show the proposed LightHuBERT enables over $10^9$ architectures concerning the embedding dimension, attention dimension, head number, feed-forward network ratio, and network depth. LightHuBERT outperforms the original HuBERT on ASR and five SUPERB tasks with the HuBERT size, achieves comparable performance to the teacher model in most tasks with a reduction of 29% parameters, and obtains a $3.5\times$ compression ratio in three SUPERB tasks, e.g., automatic speaker verification, keyword spotting, and intent classification, with a slight accuracy loss. The code and pre-trained models are available at https://github.com/mechanicalsea/lighthubert.

我们提出了 once-for-all Transformer 压缩框架 LightHuBERT，通过剪枝结构化参数自动找到所需的结构，成功地将嵌套数千个共享权重子网的基于Transformer的超网进行了设计，并使用两阶段蒸馏策略从HuBERT利用上下文化的潜在表征。在自动语音识别（ASR）和 SUPERB 基准测试上实验表明，LightHuBERT 实现了超过 $10^9$ 种结构，深度, 宽度, 注意力维度，前馈网络比例和网络深度, 在ASR 和五个 SUPERB 任务上表现优于原始的 HuBERT，与该教师模型在大多数任务中表现相当，在三个 SUPERB 任务中具有 $3.5	imes$ 的压缩比，即自动说话人验证、关键词检测和意图分类，在稍微减少准确率的情况下，可以实现 29％ 的参数减少，并提供代码和预训练模型。

LightHuBERT: 一种轻量级可配置的语音表征学习方法，采用一次性隐藏单元BERT