Papers › Accelerating Neural Transformer via an Average Attention Network

Accelerating Neural Transformer via an Average Attention Network

2 May 2018ACL 2018 7arXiv:1805.00631archive 2025-07-28

Biao Zhang, Deyi Xiong, Jinsong Su

With parallelizable attention networks, the neural Transformer is very fast to train. However, due to the auto-regressive architecture and self-attention in the decoder, the decoding procedure becomes slow. To alleviate this issue, we propose an average attention network as an alternative to the self-attention network in the decoder of the neural Transformer. The average attention network consists of two layers, with an average layer that models dependencies on previous positions and a gating layer that is stacked over the average layer to enhance the expressiveness of the proposed attention network. We apply this network on the decoder part of the neural Transformer to replace the original target-side self-attention model. With masking tricks and dynamic programming, our model enables the neural Transformer to decode sentences over four times faster than its original version with almost no loss in training time and translation performance. We conduct a series of experiments on WMT17 translation tasks, where on 6 different language pairs, we obtain robust and consistent speed-ups in decoding.

PaperPDFConference PDFCodeCode 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 19 code samples harvested from 1 repository linked to this paper: 7 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 0 where Syntology's instrument failed) · 12 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: official repository: 19 samples from 1 repository, 7 ran (7 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.

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

bzhangXMU/transformer-aan officialmentioned in papertfBSD-3-ClauseSyntology: 19 samples, 7 ran (0 instrument-failed) · 12 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.

Code Syntology ran Syntology

19 samples harvested: 7 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 0 where Syntology's instrument failed) · 12 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.

7ran
12unverified

Licence: 0 of the 19 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 bzhangXMU/transformer-aan. “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.

bzhangXMU/transformer-aan official repository 19 samples, 7 ran (0 instrument-failed) · 12 unverified
closest_length code/thumt/utils/bleu.py ran BSD-3-Clause (permissive) · dd985e59ae3a38a4 · report fixture: signature candidate:str/0/str, references:list/1/str · returned int · level recorded 2026-09-25
compute_batch_indices code/thumt/utils/search_aan.py ran BSD-3-Clause (permissive) · 6cc541726e9562c3 · report fixture: signature batch_size:int/0/int, beam_size:int/0/int · returned EagerTensor · level recorded 2026-09-25
compute_topk_scores_and_seq code/thumt/utils/search_aan.py ran BSD-3-Clause (permissive) · 90f2d69b85bf60d5 · report fixture: signature sequences:EagerTensor/3/int, scores:EagerTensor/2/float, scores_to_gather:EagerTensor/2/float, flags:EagerTensor/2/bool, beam_size:int/0/int, batch_size:int/0/int, decoded_states:list/1/EagerTensor · returned tuple · level recorded 2026-09-25
modified_precision in 9 papers code/thumt/utils/bleu.py ran BSD-3-Clause (permissive) · 9c5e87513bca1bae · report fixture: signature candidate:list/1/int, references:list/1/list, n:int/0/int · returned tuple · level recorded 2026-09-25
parse_args in 355 papers code/trainer.py ran BSD-3-Clause (permissive) · 69dd5f64c584f7eb · report fixture: signature args:list/1/str · returned Namespace · level recorded 2026-09-25
residual_fn code/thumt/models/transformer.py ran fingerprinted BSD-3-Clause (permissive) · f3ce51f764fb2afc · report fixture: signature x:EagerTensor/2/float, y:EagerTensor/2/float, keep_prob:float/0/float · returned EagerTensor · level recorded 2026-09-25
shortest_length code/thumt/utils/bleu.py ran BSD-3-Clause (permissive) · 531aab5936f2e596 · report fixture: signature references:list/1/str · returned int · level recorded 2026-09-25
add_timing_signal in 5 papers code/thumt/layers/attention.py unverified BSD-3-Clause (permissive) · d159075513daf35b · report
collect_params in 15 papers code/trainer.py unverified BSD-3-Clause (permissive) · c55c6aba7b709f18 · report
combine_heads in 6 papers code/thumt/layers/attention.py unverified BSD-3-Clause (permissive) · 273d5e370a58f518 · report
get_weights in 34 papers code/thumt/models/transformer.py unverified BSD-3-Clause (permissive) · 1b441b0d053c905a · report
import_params in 9 papers code/translator.py unverified BSD-3-Clause (permissive) · 149a90364d6911d9 · report
import_params in 9 papers code/trainer.py unverified BSD-3-Clause (permissive) · ba3612d14359b1ea · report
layer_norm in 32 papers code/thumt/layers/nn.py unverified BSD-3-Clause (permissive) · 788092fa1ba69066 · report
linear in 88 papers code/thumt/layers/nn.py unverified BSD-3-Clause (permissive) · 9676af2224c507fe · report
log_prob_from_logits code/thumt/utils/search_aan.py unverified BSD-3-Clause (permissive) · b4a5e930784a8355 · report
maxout code/thumt/layers/nn.py unverified BSD-3-Clause (permissive) · 52b62c3783732f3f · report
merge_parameters in 5 papers code/translator.py unverified BSD-3-Clause (permissive) · 8be2667aba5e919e · report
split_heads in 10 papers code/thumt/layers/attention.py unverified BSD-3-Clause (permissive) · 98b6d2c44a099591 · report

Tasks

DecoderMachine TranslationTranslation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Machine Translation WMT2014 English-German Average Attention Network BLEU score 26.31 #60 of 91 Archive leaderboard report
Machine Translation WMT2014 English-German Average Attention Network (w/o FFN) BLEU score 26.05 #63 of 91 Archive leaderboard report
Machine Translation WMT2014 English-German Average Attention Network (w/o gate) BLEU score 25.91 #65 of 91 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

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerReLUResidual ConnectionSoftmaxTransformer

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