EBERT: Efficient BERT Inference with Dynamic Structured Pruning | |
Liu, Zejian1,2; Li, Fanrong1,2; Li, Gang1; Cheng, Jian1,2 | |
2021 | |
会议日期 | 2021 |
会议地点 | Online |
英文摘要 | Pruning has been demonstrated as an effective way of reducing computational complexity for deep networks, especially CNNs for computer vision tasks. In this paper, we investigate the opportunity to accelerate the inference of large-scale pre-trained language model via pruning. We propose EBERT, a dynamic structured pruning algorithm for efficient BERT inference. Unlike previous methods that randomly prune the model weights for static inference, EBERT dynamically determines and prunes the unimportant heads in multi-head self-attention layers and the unimportant structured computations in feed-forward network for each input sample at run-time. Experimental results show that our proposed EBERT outperforms other state-of-the-art methods on different tasks. |
语种 | 英语 |
URL标识 | 查看原文 |
内容类型 | 会议论文 |
源URL | [http://ir.ia.ac.cn/handle/173211/48622] |
专题 | 类脑芯片与系统研究 |
通讯作者 | Cheng, Jian |
作者单位 | 1.National Laboratory of Pattern Recognition, Institute of Automation, CAS 2.School of Future Technology, University of Chinese Academy of Sciences |
推荐引用方式 GB/T 7714 | Liu, Zejian,Li, Fanrong,Li, Gang,et al. EBERT: Efficient BERT Inference with Dynamic Structured Pruning[C]. 见:. Online. 2021. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论