Papers › Scale Efficiently: Insights from Pre-training and Fine-tuning Transformers

Scale Efficiently: Insights from Pre-training and Fine-tuning Transformers

22 Sep 2021arXiv:2109.10686archive 2025-07-28

Yi Tay, Mostafa Dehghani, Jinfeng Rao, William Fedus, Samira Abnar, Hyung Won Chung, Sharan Narang, Dani Yogatama, Ashish Vaswani, Donald Metzler

There remain many open questions pertaining to the scaling behaviour of Transformer architectures. These scaling decisions and findings can be critical, as training runs often come with an associated computational cost which have both financial and/or environmental impact. The goal of this paper is to present scaling insights from pretraining and finetuning Transformers. While Kaplan et al. presents a comprehensive study of the scaling behaviour of Transformer language models, the scope is only on the upstream (pretraining) loss. Therefore, it is still unclear if these set of findings transfer to downstream task within the context of the pretrain-finetune paradigm. The key findings of this paper are as follows: (1) we show that aside from only the model size, model shape matters for downstream fine-tuning, (2) scaling protocols operate differently at different compute regions, (3) widely adopted T5-base and T5-large sizes are Pareto-inefficient. To this end, we present improved scaling protocols whereby our redesigned models achieve similar downstream fine-tuning quality while having 50% fewer parameters and training 40% faster compared to the widely adopted T5-base model. We publicly release over 100 pretrained checkpoints of different T5 configurations to facilitate future research and analysis.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

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

For agents, on Syntology's MCP service (how to connect):

Code

No sample from the archive's official repository (google-research/google-research) is linked to this paper; the 12 samples below come from 2 community repositories. Syntology harvested code from google-research/google-research for 24 other papers.

google-research/google-research: Syntology's graph links these samples to Efficient Neural Architecture Search via Parameter Sharing (arXiv:1802.03268; that paper's own run record: 18 ran (of which 10 constructed an object rather than computing a result; 12 with no instrument failure: 1 honoured, 0 violated, 11 with no contract checked; 6 where Syntology's instrument failed) · 20 unverified), Milking CowMask for Semi-Supervised Image Classification (arXiv:2003.12022; that paper's own run record: 1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified; the one sample that ran constructed an object rather than computing a result), Neural Additive Models: Interpretable Machine Learning with Neural Nets (arXiv:2004.13912; that paper's own run record: 26 ran (of which 14 constructed an object rather than computing a result; 26 with no instrument failure: 5 honoured, 0 violated, 21 with no contract checked; 0 where Syntology's instrument failed) · 7 unverified) and 21 more (every one in this page's JSON twin); nothing checks that they implement this paper's method.

Syntology 12 code samples harvested from 2 repositories linked to this paper: 12 ran (of which 0 constructed an object rather than computing a result; 12 with no instrument failure: 4 honoured, 0 violated, 8 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: community (archive-listed): 12 samples from 2 repositories, 12 ran (12 with no instrument failure; 0 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

No sample is linked to this paper from 1 of its repositories that Syntology harvested for other papers:

Levels written between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

google-research/google-research officialjaxSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Efficient Neural Architecture Search via Parameter Sharing (arXiv:1802.03268; that paper's own run record: 18 ran (of which 10 constructed an object rather than computing a result; 12 with no instrument failure: 1 honoured, 0 violated, 11 with no contract checked; 6 where Syntology's instrument failed) · 20 unverified), Milking CowMask for Semi-Supervised Image Classification (arXiv:2003.12022; that paper's own run record: 1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified; the one sample that ran constructed an object rather than computing a result), Neural Additive Models: Interpretable Machine Learning with Neural Nets (arXiv:2004.13912; that paper's own run record: 26 ran (of which 14 constructed an object rather than computing a result; 26 with no instrument failure: 5 honoured, 0 violated, 21 with no contract checked; 0 where Syntology's instrument failed) · 7 unverified) and 21 more (every one in this page's JSON twin); nothing checks that they implement this paper's method. report
aaronmueller/emergent-syntax mentioned on GitHubMITSyntology: 7 samples, 7 ran (0 instrument-failed) · 0 unverified report
gsarti/it5 mentioned on GitHubpytorchApache-2.0Syntology: 5 samples, 5 ran (0 instrument-failed) · 0 unverified 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.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

12 samples harvested: 12 ran (of which 0 constructed an object rather than computing a result; 12 with no instrument failure: 4 honoured, 0 violated, 8 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

4ran · honoured contract
8ran

Licence: 1 of the 12 samples is pointer only, meaning Syntology does not serve that copy's text. For that sample, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 2 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “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. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (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 are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. 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 six 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; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. 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.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity.

aaronmueller/emergent-syntax community (archive-listed) 7 samples, 7 ran (0 instrument-failed) · 0 unverified
compute_input_and_target_lengths in 5 papers run_t5_mlm_flax.py ran · honoured contract fingerprinted MIT (permissive) · eb29f6682a2426ad · report fixture: signature inputs_length:int/0/int, noise_density:float/0/float, mean_noise_span_length:float/0/float · returned tuple · level recorded 2026-09-22
evaluate_predictions in 9 papers models/pred_eval.py ran MIT (permissive) · 0c5a6c34f4b57d8b · report fixture: signature pred_filename:str/0/str, gold_filename:str/0/str, passiv:bool/0/bool · returned tuple · level recorded 2026-09-24
exact_match in 23 papers models/metrics.py ran · honoured contract MIT (permissive) · 66552297bbb6cd38 · report fixture: signature pred_sentence:str/0/str, gold_sentence:str/0/str, src_sentence:str/0/str · returned int · level recorded 2026-09-22
first_word models/metrics.py ran · honoured contract MIT (permissive) · 30008b8c34f29539 · report fixture: signature pred_sentence:str/0/str, gold_sentence:str/0/str, src_sentence:str/0/str · returned int · level recorded 2026-09-22
generate_batch_splits run_t5_mlm_flax.py ran MIT (permissive) here; the stored copy is pointer only · 26eb4e2856b3d7a1 · report fixture: signature samples_idx:ndarray/1/int, batch_size:int/0/int, drop_last:bool/0/bool · returned ndarray · level recorded 2026-09-24
get_dirs in 5 papers analysis/plot_scaling.py ran MIT (permissive) · 219961813fa6d977 · report fixture: signature prefixes:list/1/str, suffixes:list/1/str · returned list · level recorded 2026-09-24
second_word models/metrics.py ran · honoured contract MIT (permissive) · bd0d4ea62311e45d · report fixture: signature pred_sentence:str/0/str, gold_sentence:str/0/str, src_sentence:str/0/str · returned int · level recorded 2026-09-22
gsarti/it5 community (archive-listed) 5 samples, 5 ran (0 instrument-failed) · 0 unverified
get_bsz finetuning/vars.py ran Apache-2.0 (permissive) · a747070e3ebf347f · report fixture: signature model:str/0/str · returned int · level recorded 2026-09-24
get_bsz inference/vars.py ran Apache-2.0 (permissive) · d79ddc3312ab8f78 · report fixture: signature model:str/0/str · returned int · level recorded 2026-09-24
get_lr in 91 papers finetuning/vars.py ran Apache-2.0 (permissive) · c4f59346ce8a05d9 · report fixture: signature model:str/0/str · returned float · level recorded 2026-09-24
get_params in 53 papers inference/vars.py ran Apache-2.0 (permissive) · f74861ceb9e003dd · report fixture: signature model:str/0/str, config:str/0/str · returned dict · level recorded 2026-09-24
get_time in 20 papers finetuning/vars.py ran Apache-2.0 (permissive) · 07d60309bc634d37 · report fixture: signature model:str/0/str · returned int · level recorded 2026-09-24

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Absolute Position EncodingsAdafactorAdamAttentionAttention DropoutBPEDense ConnectionsDropoutGated Linear UnitInverse Square Root ScheduleLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSentencePieceSoftmaxT5Transformer

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