Papers › An Empirical Model of Large-Batch Training

An Empirical Model of Large-Batch Training

14 Dec 2018arXiv:1812.06162archive 2025-07-28

Sam McCandlish, Jared Kaplan, Dario Amodei, OpenAI Dota Team

In an increasing number of domains it has been demonstrated that deep learning models can be trained using relatively large batch sizes without sacrificing data efficiency. However the limits of this massive data parallelism seem to differ from domain to domain, ranging from batches of tens of thousands in ImageNet to batches of millions in RL agents that play the game Dota 2. To our knowledge there is limited conceptual understanding of why these limits to batch size differ or how we might choose the correct batch size in a new domain. In this paper, we demonstrate that a simple and easy-to-measure statistic called the gradient noise scale predicts the largest useful batch size across many domains and applications, including a number of supervised learning datasets (MNIST, SVHN, CIFAR-10, ImageNet, Billion Word), reinforcement learning domains (Atari and Dota), and even generative model training (autoencoders on SVHN). We find that the noise scale increases as the loss decreases over a training run and depends on the model size primarily through improved model performance. Our empirically-motivated theory also describes the tradeoff between compute-efficiency and time-efficiency, and provides a rough model of the benefits of adaptive batch-size training.

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

Syntology 6 code samples harvested from 1 repository linked to this paper: 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 3 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): 6 samples from 1 repository, 3 ran (3 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 3 of its repositories that Syntology harvested for other papers:

Levels written between 2026-09-25 and 2026-09-26 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

RexGLiu/rlpyt_crbp mentioned on GitHubpytorchMITSyntology: not harvested report
akterskii/rlpyt mentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to rlpyt: A Research Code Base for Deep Reinforcement Learning in PyTorch (arXiv:1909.01500; that paper's own run record: 8 ran (of which 0 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method. report
alexsax/testing_rlpyt mentioned on GitHubpytorchnot reachable when probed 2026-09-17 — repositories for recent papers often appear after camera-readySyntology: not harvested report
astooke/rlpyt mentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Decoupling Representation Learning from Reinforcement Learning (arXiv:2009.08319; that paper's own run record: 2 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report
davidandym/task-conflict-in-text-to-text-learners mentioned on GitHubpytorchMITSyntology: 6 samples, 3 ran (0 instrument-failed) · 3 unverified report
eac-replication/eac-replication mentioned on GitHubpytorchMITSyntology: not harvested report
hal-314/fastai-batch-size-finder mentioned on GitHubApache-2.0Syntology: not harvested report
petuum/adaptdl mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to AdaScale SGD: A User-Friendly Algorithm for Distributed Training (arXiv:2007.05105; that paper's own run record: 0 ran · 2 unverified); nothing checks that they implement this paper's method. report
sandeeprockstar/Pose_Estimation mentioned on GitHubpytorchSyntology: not harvested report
sarahisyoung/rlpyt mentioned on GitHubpytorchMITSyntology: not harvested report
DanyWind/fastai_bs_finder pytorchSyntology: not harvested 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

6 samples harvested: 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 3 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.

3ran
3unverified

Licence: 0 of the 6 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 davidandym/task-conflict-in-text-to-text-learners. “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. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

davidandym/task-conflict-in-text-to-text-learners community (archive-listed) 6 samples, 3 ran (0 instrument-failed) · 3 unverified
get_conflict_metric_keys src/measure_conflict.py ran MIT (permissive) · 357c3ce4115b3f1b · report fixture: signature train_tasks:list/1/str, args:Args/0/? · returned list · level recorded 2026-09-26
masked_softmax in 51 papers
census: 10 implementations from 10 papers, 10 distinct outputs across 6 buckets, one shared input per bucket
src/models/canonical_heads.py ran MIT (permissive) · 854be3983683cc44 · report fixture: signature logits:Tensor/2/float, mask:Tensor/2/float, dim:int/0/int, log_softmax:bool/0/bool · returned Tensor · level recorded 2026-09-26
score in 54 papers src/deca_metrics.py ran MIT (permissive) · e81a8138a9405edf · report fixture: signature answer:str/0/str, gold:list/1/str · returned ndarray · level recorded 2026-09-25
compute_big_batch_gradient src/measure_conflict.py unverified MIT (permissive) · a862dac46152a1e9 · report
measure_conflict src/measure_conflict.py unverified MIT (permissive) · ae6e307cdcec33d5 · report
set_seed in 49 papers src/utils.py unverified MIT (permissive) · 7e5f9c2aee507dd4 · report

Tasks

Dota 2Reinforcement Learningmodel

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

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