Papers › mT5: A massively multilingual pre-trained text-to-text transformer

mT5: A massively multilingual pre-trained text-to-text transformer

22 Oct 2020NAACL 2021 4arXiv:2010.11934archive 2025-07-28

Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel

The recent "Text-to-Text Transfer Transformer" (T5) leveraged a unified text-to-text format and scale to attain state-of-the-art results on a wide variety of English-language NLP tasks. In this paper, we introduce mT5, a multilingual variant of T5 that was pre-trained on a new Common Crawl-based dataset covering 101 languages. We detail the design and modified training of mT5 and demonstrate its state-of-the-art performance on many multilingual benchmarks. We also describe a simple technique to prevent "accidental translation" in the zero-shot setting, where a generative model chooses to (partially) translate its prediction into the wrong language. All of the code and model checkpoints used in this work are publicly available.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology View this paper on Syntology: its repositories, every harvested function with whether it ran, its licence and the call to fetch it.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2010.11934")

Code

Syntology Ran 5 of 13 code samples harvested from 3 repositories linked to this paper; 8 have no recorded run. Of those that ran: 5 ran with no contract checked.

By repository: official repository: 5 samples from 1 repository, 0 ran; community (archive-listed): 8 samples from 2 repositories, 5 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

google-research/multilingual-t5 officialmentioned on GitHubtfApache-2.0 report
KoshiroSato/Flask_NLP_App mentioned on GitHubMIT report
MorenoLaQuatra/bart-it mentioned on GitHubpytorchMIT report
google-research/byt5 mentioned on GitHubtfApache-2.0 report
huggingface/transformers mentioned on GitHubpytorch report
manshri/tesum mentioned on GitHubpytorch report
pwc-1/Paper-5 mindspore report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

13 samples harvested; 5 ran; 0 honoured the contract we drafted; 8 have no recorded run. Read from Syntology's graph 2026-09-25; that is when this build read the record, not when the samples ran.

5ran
8unverified

Licence: 0 of the 13 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 3 repositories linked to this paper, official or community; each sample names its own and says which. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: 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. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

normalize_mlqa google-research/multilingual-t5/multilingual_t5/evaluation/metrics.py official repository unverified Apache-2.0 (permissive) · 4ad4262f32bec052 · report
span_f1 google-research/multilingual-t5/multilingual_t5/evaluation/metrics.py official repository unverified Apache-2.0 (permissive) · 3a70b94fccf1c165 · report
wikiann google-research/multilingual-t5/multilingual_t5/preprocessors.py official repository unverified Apache-2.0 (permissive) · 6a5b181d28b784b6 · report
xnli_map_hypothesis_premise google-research/multilingual-t5/multilingual_t5/preprocessors.py official repository unverified Apache-2.0 (permissive) · f04d1c1706891df6 · report
xquad google-research/multilingual-t5/multilingual_t5/preprocessors.py official repository unverified Apache-2.0 (permissive) · 64c4f17ea92b71d7 · report
bert_predict KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/bert_model.py community (archive-listed) ran MIT (permissive) · ab4db4a59f94ef40 · report
mt5_post_processing KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/mt5_model.py community (archive-listed) ran MIT (permissive) · 0d9ef0d0dfe26859 · report
predict_with_post_processing KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/bert_model.py community (archive-listed) ran MIT (permissive) · 0d4436476da0cc30 · report
preprocess_logits_for_metrics MorenoLaQuatra/bart-it/summarization/finetune_summarization.py community (archive-listed) ran MIT (permissive) · 2641f706223d1e5c · report
read_txt KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/utils.py community (archive-listed) ran fingerprinted MIT (permissive) · e445d0574fe31243 · report
app_logger KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/logger.py community (archive-listed) unverified MIT (permissive) · d10b1ac136fbfec1 · report
model_io_logger KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/logger.py community (archive-listed) unverified MIT (permissive) · baf8e460eed595db · report
read_yaml KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/utils.py community (archive-listed) unverified MIT (permissive) · c9737e28fdfdcfe9 · report

Tasks

Common Sense ReasoningNatural Language InferenceQuestion AnsweringReading ComprehensionTranslation

Datasets

Introduced by this paper, per the archive.

mC4

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Common Sense Reasoning PARus MT5 Large Accuracy 0.504 #14 of 22 Archive leaderboard report
Common Sense Reasoning RWSD MT5 Large Accuracy 0.669 #13 of 22 Archive leaderboard report
Common Sense Reasoning RuCoS MT5 Large Average F1 0.57 #10 of 22 Archive leaderboard report
Common Sense Reasoning RuCoS MT5 Large EM 0.562 #10 of 22 Archive leaderboard report
Natural Language Inference LiDiRus MT5 Large MCC 0.061 #16 of 22 Archive leaderboard report
Natural Language Inference RCB MT5 Large Accuracy 0.454 #11 of 22 Archive leaderboard report
Natural Language Inference RCB MT5 Large Average F1 0.366 #11 of 22 Archive leaderboard report
Natural Language Inference TERRa MT5 Large Accuracy 0.561 #16 of 22 Archive leaderboard report
Question Answering DaNetQA MT5 Large Accuracy 0.657 #10 of 22 Archive leaderboard report
Reading Comprehension MuSeRC MT5 Large Average F1 0.844 #2 of 22 Archive leaderboard report
Reading Comprehension MuSeRC MT5 Large EM 0.543 #2 of 22 Archive leaderboard report
Zero-Shot Cross-Lingual Transfer XTREME mT5 Avg 40.9 #23 of 25 Archive leaderboard report
Zero-Shot Cross-Lingual Transfer XTREME mT5 Question Answering 73.6 #23 of 25 Archive leaderboard report
Zero-Shot Cross-Lingual Transfer XTREME mT5 Sentence Retrieval NA #23 of 25 Archive leaderboard report
Zero-Shot Cross-Lingual Transfer XTREME mT5 Sentence-pair Classification 89.8 #23 of 25 Archive leaderboard report
Zero-Shot Cross-Lingual Transfer XTREME mT5 Structured Prediction NA #23 of 25 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

AdafactorAttentionAttention DropoutBPEDense ConnectionsDropoutGated Linear UnitInverse Square Root ScheduleLayer NormalizationLinear LayerMulti-Head AttentionResidual ConnectionSentencePieceSoftmaxT5mT5

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections