{"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/deebert-dynamic-early-exiting-for","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","arxiv_id":"2004.12993","date":"2020-04-27","proceeding":"ACL 2020 6","authors":["Ji Xin","Raphael Tang","Jaejun Lee","Yao-Liang Yu","Jimmy Lin"],"abstract":"Large-scale pre-trained language models such as BERT have brought significant improvements to NLP applications. However, they are also notorious for being slow in inference, which makes them difficult to deploy in real-time applications. We propose a simple but effective method, DeeBERT, to accelerate BERT inference. Our approach allows samples to exit earlier without passing through the entire model. Experiments show that DeeBERT is able to save up to ~40% inference time with minimal degradation in model quality. Further analyses show different behaviors in the BERT transformer layers and also reveal their redundancy. Our work provides new ideas to efficiently apply deep transformer-based models to downstream tasks. Code is available at https://github.com/castorini/DeeBERT.","url_abs":"https://arxiv.org/abs/2004.12993v1","url_pdf":"https://arxiv.org/pdf/2004.12993v1.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":"deebert-dynamic-early-exiting-for","repo_url":"https://github.com/castorini/deebert","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null},{"paper_slug":"deebert-dynamic-early-exiting-for","repo_url":"https://github.com/JulesBelveze/bert-squeeze","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null},{"paper_slug":"deebert-dynamic-early-exiting-for","repo_url":"https://github.com/huggingface/transformers","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"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":"bpe","method_name":"BPE"},{"method_slug":"deebert","method_name":"DeeBERT"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"early-exiting","method_name":"Early exiting"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"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":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"position-wise-feed-forward-layer","method_name":"Position-Wise Feed-Forward Layer"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"},{"method_slug":"weight-decay","method_name":"Weight Decay"},{"method_slug":"wordpiece","method_name":"WordPiece"}],"datasets_introduced":[],"methods_introduced":[{"slug":"deebert","name":"DeeBERT","full_name":"DeeBERT"}],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2004.12993","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2004.12993"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/huggingface/transformers","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/castorini/deebert","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/JulesBelveze/bert-squeeze","reach":null}],"summary":{"ran":4,"ran_honours":1,"unverified":1},"by_repo_kind":{"official":{"samples":1,"ran":1,"repositories":1},"listed":{"samples":5,"ran":4,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":5,"samples":[{"code_sha256_prefix":"3c68e7d3c7a22462","entry":"BertHighway","repo":"castorini/deebert","repo_kind":"official","path":"transformers/modeling_highway_bert.py","file_url":"https://github.com/castorini/deebert/blob/HEAD/transformers/modeling_highway_bert.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"3c68e7d3c7a22462"}},{"code_sha256_prefix":"acec206d3de79819","entry":"DeeBertEncoderOutput","repo":"JulesBelveze/bert-squeeze","repo_kind":"listed","path":"bert_squeeze/models/custom_transformers/deebert.py","file_url":"https://github.com/JulesBelveze/bert-squeeze/blob/HEAD/bert_squeeze/models/custom_transformers/deebert.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"acec206d3de79819"}},{"code_sha256_prefix":"a05459465715ee48","entry":"OffRamp","repo":"JulesBelveze/bert-squeeze","repo_kind":"listed","path":"bert_squeeze/models/custom_transformers/deebert.py","file_url":"https://github.com/JulesBelveze/bert-squeeze/blob/HEAD/bert_squeeze/models/custom_transformers/deebert.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"a05459465715ee48"}},{"code_sha256_prefix":"2e045a653a51bbdc","entry":"RampOutput","repo":"JulesBelveze/bert-squeeze","repo_kind":"listed","path":"bert_squeeze/models/custom_transformers/deebert.py","file_url":"https://github.com/JulesBelveze/bert-squeeze/blob/HEAD/bert_squeeze/models/custom_transformers/deebert.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"2e045a653a51bbdc"}},{"code_sha256_prefix":"76ecddd69557ac5b","entry":"entropy","repo":"JulesBelveze/bert-squeeze","repo_kind":"listed","path":"bert_squeeze/models/custom_transformers/deebert.py","file_url":"https://github.com/JulesBelveze/bert-squeeze/blob/HEAD/bert_squeeze/models/custom_transformers/deebert.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"76ecddd69557ac5b"}},{"code_sha256_prefix":"dc1656c6e19154ad","entry":"DeeBertEncoder","repo":"JulesBelveze/bert-squeeze","repo_kind":"listed","path":"bert_squeeze/models/custom_transformers/deebert.py","file_url":"https://github.com/JulesBelveze/bert-squeeze/blob/HEAD/bert_squeeze/models/custom_transformers/deebert.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"dc1656c6e19154ad"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}