{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/mobilebert-a-compact-task-agnostic-bert-for","title":"MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices","arxiv_id":"2004.02984","date":"2020-04-06","proceeding":"ACL 2020 6","authors":["Zhiqing Sun","Hongkun Yu","Xiaodan Song","Renjie Liu","Yiming Yang","Denny Zhou"],"abstract":"Natural Language Processing (NLP) has recently achieved great success by using huge pre-trained models with hundreds of millions of parameters. However, these models suffer from heavy model sizes and high latency such that they cannot be deployed to resource-limited mobile devices. In this paper, we propose MobileBERT for compressing and accelerating the popular BERT model. Like the original BERT, MobileBERT is task-agnostic, that is, it can be generically applied to various downstream NLP tasks via simple fine-tuning. Basically, MobileBERT is a thin version of BERT_LARGE, while equipped with bottleneck structures and a carefully designed balance between self-attentions and feed-forward networks. To train MobileBERT, we first train a specially designed teacher model, an inverted-bottleneck incorporated BERT_LARGE model. Then, we conduct knowledge transfer from this teacher to MobileBERT. Empirical studies show that MobileBERT is 4.3x smaller and 5.5x faster than BERT_BASE while achieving competitive results on well-known benchmarks. On the natural language inference tasks of GLUE, MobileBERT achieves a GLUEscore o 77.7 (0.6 lower than BERT_BASE), and 62 ms latency on a Pixel 4 phone. On the SQuAD v1.1/v2.0 question answering task, MobileBERT achieves a dev F1 score of 90.0/79.2 (1.5/2.1 higher than BERT_BASE).","url_abs":"https://arxiv.org/abs/2004.02984v2","url_pdf":"https://arxiv.org/pdf/2004.02984v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"mobilebert-a-compact-task-agnostic-bert-for","repo_url":"https://github.com/tensorflow/models/tree/master/official/nlp/projects/mobilebert","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null},{"paper_slug":"mobilebert-a-compact-task-agnostic-bert-for","repo_url":"https://github.com/Milan-Chicago/GLG-Automated-Meta-data-Tagging","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"mobilebert-a-compact-task-agnostic-bert-for","repo_url":"https://github.com/tchebonenko/Automated-Topic_Modeling-and-NER","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"mobilebert-a-compact-task-agnostic-bert-for","repo_url":"https://github.com/2023-MindSpore-1/ms-code-159","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":{"status":"ok"}},{"paper_slug":"mobilebert-a-compact-task-agnostic-bert-for","repo_url":"https://github.com/MS-P3/code5/tree/main/mobilebert","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"mobilebert-a-compact-task-agnostic-bert-for","repo_url":"https://github.com/nosaydomore/MobileBert_paddle","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"paddle","reach":{"status":"ok"}},{"paper_slug":"mobilebert-a-compact-task-agnostic-bert-for","repo_url":"https://github.com/pwc-1/Paper-5/tree/main/mobilebert","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"natural-language-inference","task_name":"Natural Language Inference"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"semantic-textual-similarity","task_name":"Semantic Textual Similarity"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"}],"methods":[{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"attention-dropout","method_name":"Attention Dropout"},{"method_slug":"bert","method_name":"BERT"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"ib-bert","method_name":"IB-BERT"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"linear-warmup-with-linear-decay","method_name":"Linear Warmup With Linear Decay"},{"method_slug":"mobilebert","method_name":"MobileBERT"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"weight-decay","method_name":"Weight Decay"},{"method_slug":"wordpiece","method_name":"WordPiece"}],"datasets_introduced":[],"methods_introduced":[{"slug":"mobilebert","name":"MobileBERT","full_name":"MobileBERT"}],"results":[{"leaderboard":"/sota/semantic-textual-similarity-on-mrpc","task":"Semantic Textual Similarity","dataset":"MRPC","model":"MobileBERT","rank_in_archive_order":20,"of":45,"metrics":{"Accuracy":"88.8%"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2004.02984","atlas_url":"https://app.syntology.ai/?focus=2004.02984","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}