{"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/longt5-efficient-text-to-text-transformer-for","title":"LongT5: Efficient Text-To-Text Transformer for Long Sequences","arxiv_id":"2112.07916","date":"2021-12-15","proceeding":"Findings (NAACL) 2022 7","authors":["Mandy Guo","Joshua Ainslie","David Uthus","Santiago Ontanon","Jianmo Ni","Yun-Hsuan Sung","Yinfei Yang"],"abstract":"Recent work has shown that either (1) increasing the input length or (2) increasing model size can improve the performance of Transformer-based neural models. In this paper, we present a new model, called LongT5, with which we explore the effects of scaling both the input length and model size at the same time. Specifically, we integrated attention ideas from long-input transformers (ETC), and adopted pre-training strategies from summarization pre-training (PEGASUS) into the scalable T5 architecture. The result is a new attention mechanism we call {\\em Transient Global} (TGlobal), which mimics ETC's local/global attention mechanism, but without requiring additional side-inputs. We are able to achieve state-of-the-art results on several summarization tasks and outperform the original T5 models on question answering tasks.","url_abs":"https://arxiv.org/abs/2112.07916v2","url_pdf":"https://arxiv.org/pdf/2112.07916v2.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":"longt5-efficient-text-to-text-transformer-for","repo_url":"https://github.com/google-research/longt5","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null},{"paper_slug":"longt5-efficient-text-to-text-transformer-for","repo_url":"https://github.com/utsjiyaoli/qa-attack","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"longt5-efficient-text-to-text-transformer-for","repo_url":"https://github.com/MindSpore-scientific-2/code-14/tree/main/longt5","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"longt5-efficient-text-to-text-transformer-for","repo_url":"https://github.com/pwc-1/Paper-9/tree/main/longt5","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"abstractive-text-summarization","task_name":"Abstractive Text Summarization"},{"task_slug":"long-range-modeling","task_name":"Long-range modeling"},{"task_slug":"multi-document-summarization","task_name":"Multi-Document Summarization"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"text-summarization","task_name":"Text Summarization"}],"methods":[{"method_slug":"adafactor","method_name":"Adafactor"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"attention-dropout","method_name":"Attention Dropout"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"glu","method_name":"Gated Linear Unit"},{"method_slug":"inverse-square-root-schedule","method_name":"Inverse Square Root Schedule"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"sentencepiece","method_name":"SentencePiece"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"t5","method_name":"T5"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/abstractive-text-summarization-on-cnn-daily","task":"Abstractive Text Summarization","dataset":"CNN / Daily Mail","model":"LongT5","rank_in_archive_order":21,"of":53,"metrics":{"ROUGE-1":"43.94","ROUGE-2":"21.40 ","ROUGE-L":"41.28"},"uses_additional_data":false},{"leaderboard":"/sota/long-range-modeling-on-scrolls","task":"Long-range modeling","dataset":"SCROLLS","model":"LongT5 XL","rank_in_archive_order":2,"of":13,"metrics":{"Avg.":"42.53","CNLI":"88.2","GovRep":"61.1 / 32.3 / 33.7","Nrtv":"29.3","QALT EM-T/H":"46.0 / 42.1","QMSum":"34.9 / 11.8 / 23.5","Qspr":"53.1","SumScr":"35.8 / 9.6 / 21.1"},"uses_additional_data":false},{"leaderboard":"/sota/long-range-modeling-on-scrolls","task":"Long-range modeling","dataset":"SCROLLS","model":"LongT5 Large","rank_in_archive_order":3,"of":13,"metrics":{"Avg.":"41.03","CNLI":"87.3","GovRep":"61.3/32.2/33.8","Nrtv":"27.2","QALT EM-T/H":"40.6 / 38.6","QMSum":"35.1 / 12.0 / 23.3","Qspr":"52.3","SumScr":"60.3 / 31.1 / 32.8"},"uses_additional_data":false},{"leaderboard":"/sota/long-range-modeling-on-scrolls","task":"Long-range modeling","dataset":"SCROLLS","model":"LongT5 Base","rank_in_archive_order":5,"of":13,"metrics":{"Avg.":"38.6","CNLI":"85.6","GovRep":"57.7 / 30.0 / 31.4","Nrtv":"23.0","QALT EM-T/H":"37.9 / 36.6","QMSum":"33.9 / 11.0 / 22.8","Qspr":"46.6","SumScr":"34.8 / 9.6 / 21.1"},"uses_additional_data":false},{"leaderboard":"/sota/multi-document-summarization-on-multi-news","task":"Multi-Document Summarization","dataset":"Multi-News","model":"LongT5","rank_in_archive_order":2,"of":6,"metrics":{"ROUGE-1":"48.17","ROUGE-2":"19.43","ROUGE-SU4":"24.94"},"uses_additional_data":false},{"leaderboard":"/sota/text-summarization-on-arxiv","task":"Text Summarization","dataset":"Arxiv HEP-TH citation graph","model":"LongT5","rank_in_archive_order":9,"of":28,"metrics":{"ROUGE-1":"48.35","ROUGE-2":"21.92","ROUGE-L":" 44.27"},"uses_additional_data":false},{"leaderboard":"/sota/text-summarization-on-bigpatent","task":"Text Summarization","dataset":"BigPatent","model":"LongT5","rank_in_archive_order":1,"of":2,"metrics":{"ROUGE-1":"76.87","ROUGE-2":"66.06","ROUGE-L":"70.76"},"uses_additional_data":false},{"leaderboard":"/sota/text-summarization-on-pubmed-1","task":"Text Summarization","dataset":"Pubmed","model":"LongT5","rank_in_archive_order":4,"of":29,"metrics":{"ROUGE-1":"50.23","ROUGE-2":"24.76","ROUGE-L":"46.67"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2112.07916","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2112.07916"}},"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/MindSpore-scientific-2/code-14/tree/main/longt5","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/google-research/longt5","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/utsjiyaoli/qa-attack","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/pwc-1/Paper-9/tree/main/longt5","reach":null}],"summary":{"ran_draft_wrong":1},"by_repo_kind":{},"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":1,"samples":[{"code_sha256_prefix":"3df820a3fa7f5f80","entry":"parse_json_line","repo":null,"repo_kind":null,"path":null,"file_url":null,"link_basis":"identical_code_first_harvested_elsewhere","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":null,"inline_ok":false,"mcp_get_code":{"code_sha256":"3df820a3fa7f5f80"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}