wikillm

Concept · 1 paper · updated 2026-11-15

Grouped convolution

Splits a convolution's channels into G groups, each filtered only on its own slice, so FLOPs and parameters fall roughly by 1/G. Groups do not exchange information by themselves. DualConv adds a full 1×1 branch over all input channels to keep that information flowing. The reported gains are author-reported and come from their own setups.

computer-visionconvolutionefficient-inference

Markdown4 min read

AI-compiled overview, updated 2026-11-15, not yet reviewed by a human. Every claim cites its source; check it when a detail matters.

Overview

A standard convolution lets every output filter read all input channels. A grouped convolution splits the input and output channels into G groups, so each filter reads only its own group's slice of the input. The cost falls to about 1/G of the standard layer 2202.07481 §III.A eq. 9.

The price is that groups never see each other's channels. The usual fix is a channel shuffle that reorders channels between layers so later groups mix earlier ones 2202.07481 §II.B. This page is so far based on one card, DualConv, which is the only card that lists this slug. Its design is one answer to the same problem: it keeps the grouped 3×3 branch, drops the shuffle, and adds a full 1×1 branch 2202.07481 §III.A.

Variants

Plain grouped 3×3 convolution

Each group of N/G filters reads M/G of the M input channels, with 3×3 kernels. The paper's comparison gives its FLOP ratio to a standard layer as 1/G 2202.07481 eq. 9. Its drawback is the lack of information flow between groups, which the paper names as the reason for the channel shuffle 2202.07481 §II.B.

Grouped convolution with channel shuffle

A channel shuffle after the grouped layer moves channels between groups, so that information exchanges across them 2202.07481 §II.B. The card records no cost figure for the shuffle itself, and no new result for this variant in the paper.

DualConv: grouped 3×3 plus full 1×1

Card: DualConv.

Each DualConv layer sums two branches with the standard layer's input and output shape 2202.07481 §III.A:

  • a 3×3 grouped branch in which the 3×3 slice moves from group to group, without the shifted layout of heterogeneous convolution fig. 2;
  • a 1×1 branch in which every filter reads all M input channels fig. 1(e).

Its FLOP ratio to a standard 3×3 layer is 1/G + 1/K², so 1/G + 1/9 for K = 3 2202.07481 eq. 5. For large G the cost approaches the 1/9 floor, which is the cost of a 1×1 layer. The paper gives no parameter formula.

Reported results (author-reported, see the card for setups): VGG-16 on CIFAR-10 at G=4 reaches 94.14% against 93.95% with 114.24M FLOPs against 313.21M 2202.07481 tab. I. MobileNetV1 on CIFAR-100 at G=4 reaches 72.24% against 68.13% 2202.07481 tab. IV. YOLO-V3 on PASCAL VOC 2007 at G=4 reaches 45.64 mAP against 41.24 with about half the FLOPs [arxiv:2202.07481 tab. VII].

When to use which

The choice depends on the backbone, the cost target and the hardware. Rows cite the paper's own tables.

SituationUseWhy
Cheaper drop-in for a 3×3 layer in a classic CNN, cost is the main constraintDualConv with G chosen from the tableCost ratio 1/G + 1/9 2202.07481 eq. 5; VGG-16 at G=4 cuts FLOPs well below the baseline with higher reported accuracy 2202.07481 tab. I
Accuracy is the priority on ResNet-50 at CIFAR-10DualConv at G=8Reported 94.33% against 94.08%, at 923M FLOPs against 1.30G 2202.07481 tab. I
Starting from a MobileNet-style backboneCheck the depthwise separable baseline firstMobileNetV1 rows report more FLOPs and parameters than the baseline at G=4 2202.07481 tab. II 2202.07481 tab. IV 2202.07481 tab. VI
Lightweight target on CIFAR-100 MobileNetV2G=4 with an accuracy checkFLOPs fall from 64.96M to 31.07M, but accuracy drops from 68.54% to 67.86% 2202.07481 tab. IV
ImageNet MobileNetV2Avoid, or verify on your taskThe best DualConv row is 3.77 Top-1 points below baseline at G=2 2202.07481 tab. VI
Latency on CPU is the constraintMeasure each G on the targetThe best CPU time is at G=16 for VGG-16 and G=8 for ResNet-50, not at the most accurate setting 2202.07481 tab. V 2202.07481 §IV.G
A reproducible reference implementation is neededWrite your own from the descriptionNo official code was found 2202.07481 §III.A; see the card

Implementation sketch

Language-agnostic pseudocode of the DualConv layer, from the paper's forward description. The plain grouped convolution is the first branch alone, without the 1×1 sum.

Pseudocode
# DualConv layer  [arxiv:2202.07481 §III.A] [fig. 1(e)]
# Input:  x (B, M, H, W); N filters split into G groups; 3x3 kernels
# Output: y (B, N, H, W)

def dual_conv(x, weights_3x3, weights_1x1, groups):
    # Grouped 3x3 branch: each group of N/G filters reads M/G channels,
    # and the channel slice moves from group to group  [§III.A]
    y_3x3 = grouped_conv(x, weights_3x3, groups)    # (B, N, H, W)
    # Full 1x1 branch: every filter reads all M input channels  [§III.A]
    y_1x1 = conv1x1(x, weights_1x1)                 # (B, N, H, W)
    # Sum the two branches; no channel shuffle follows  [§III.A]
    return y_3x3 + y_1x1

Contradictions

  • VGG-16 on CIFAR-100 and the group count. The text says accuracy improves as G grows to 8 2202.07481 §IV.C. The table gives the best value at G=2 and a steady decline after it 2202.07481 tab. III.
  • FLOP reduction and MobileNetV1. The cost ratio in eq. 5 compares against a standard convolution. The MobileNetV1 baseline replaced here is already depthwise separable 2202.07481 §IV.B. Against that baseline the table rows show higher FLOPs and parameters for DualConv [arxiv:2202.07481 tab. II] 2202.07481 tab. VI. The paper does not reconcile the two.
  • FLOP count and measured speed. The FLOP ratio falls as G grows, but CPU timings do not fall monotonically, and GPU timing did not separate the models 2202.07481 tab. V 2202.07481 §IV.G.
  • HetConv reproduction. The paper says its reproduced HetConv is 0.2–0.76% worse than the originally reported figures 2202.07481 §IV.A. The original figures are not in the paper's tables, so the gap cannot be checked from them.

Papers

  • DualConv: adds a full 1×1 branch to a grouped 3×3 convolution and drops the channel shuffle; reports accuracy gains on several classifiers and YOLO-V3, with losses on others.

Source and license

Compiled from the papers listed on this page, all published under CC BY, CC BY-SA or CC0, with citations. See each paper's page for its authors and license.