This is the public, feature-limited version of the conference webpage. After Registration and login please visit the full version.

DynaBERT: Dynamic BERT with Adaptive Width and Depth

Lu Hou, Zhiqi Huang, Lifeng Shang, Xin Jiang, Xiao Chen, Qun Liu

Spotlight presentation: Orals & Spotlights Track 03: Language/Audio Applications
on 2020-12-07T20:00:00-08:00 - 2020-12-07T20:10:00-08:00
Poster Session 1 (more posters)
on 2020-12-07T21:00:00-08:00 - 2020-12-07T23:00:00-08:00
Abstract: The pre-trained language models like BERT, though powerful in many natural language processing tasks, are both computation and memory expensive. To alleviate this problem, one approach is to compress them for specific tasks before deployment. However, recent works on BERT compression usually compress the large BERT model to a fixed smaller size, and can not fully satisfy the requirements of different edge devices with various hardware performances. In this paper, we propose a novel dynamic BERT model (abbreviated as DynaBERT), which can flexibly adjust the size and latency by selecting adaptive width and depth. The training process of DynaBERT includes first training a width-adaptive BERT and then allowing both adaptive width and depth, by distilling knowledge from the full-sized model to small sub-networks. Network rewiring is also used to keep the more important attention heads and neurons shared by more sub-networks. Comprehensive experiments under various efficiency constraints demonstrate that our proposed dynamic BERT (or RoBERTa) at its largest size has comparable performance as BERT-base (or RoBERTa-base), while at smaller widths and depths consistently outperforms existing BERT compression methods. Code is available at https://github.com/huawei-noah/Pretrained-Language-Model/tree/master/DynaBERT.

Preview Video and Chat

To see video, interact with the author and ask questions please use registration and login.