Methods › General › Attention Modules › Multi-DConv-Head Attention

Multi-DConv-Head Attention

15 papers tagged archive 2025-07-28

archive 2025-07-28 Description, source and code snippet are the archive's method entry.

Multi-DConv-Head Attention, or MDHA, is a type of Multi-Head Attention that utilizes depthwise convolutions after the multi-head projections. It is used in the Primer Transformer architecture.

Specifically, 3x1 depthwise convolutions are added after each of the multi-head projections for query Q, key K and value V in self-attention. These depthwise convolutions are performed over the spatial dimension of each dense projection’s output. Interestingly, this ordering of pointwise followed by depthwise convolution is the reverse of typical separable convolution, which the authors find to be less effective. They also find that wider depthwise convolution and standard convolution not only do not improve performance, but in several cases hurt it.

MDHA is similar to Convolutional Attention, which uses separable convolution instead of depthwise convolution and does not apply convolution operations per attention head as in MDHA.

Source: Primer: Searching for Efficient Transformers for...

Papers archive 2025-07-28

15 shown of 15, newest first. Repository counts are the archive's code-links table. A Syntology line states what Syntology ran from that paper's harvested code; it is per sample and not a correctness claim.

Tasks archive 2025-07-28

18 tasks the archive attaches to papers tagged with this method, by distinct papers. A task without a page in the catalog is plain text.

TaskPapers
Language Modeling5
Language Modelling5
CPU1
Common Sense Reasoning1
Coreference Resolution1
Deep Learning1
Diversity1
Epidemiology1
GPU1
Natural Language Inference1
Protein Structure Prediction1
Question Answering1
Safety Alignment1
Sentiment Analysis1
Specificity1
TAR1
Text Classification1
Word Sense Disambiguation1

Usage over time archive 2025-07-28

Papers per year tagged with Multi-DConv-Head Attention: 2021 to 2025, peak 4 4 0 2021: 3 papers 2021 2022: 3 papers 2022 2023: 4 papers 2023 2024: 2 papers 2024 2025: 3 papers 2025
Papers per year the archive tags with this method, by the paper's archive date (15 dated). Bars are counts, not a trend claim.

Components: the archive holds no method-to-method composition, so PwC's Components table cannot be rebuilt; the Papers list carries no Results column for the same reason (the archive does not join its leaderboard rows to method tags).

Categories archive 2025-07-28

Attention Modules

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