# SD-Conv: Towards the Parameter-Efficiency of Dynamic Convolution

Source: [SD-Conv: Towards the Parameter-Efficiency of Dynamic Convolution](<https://www.sota2.com/research/paper/68fa6f7e1e209131cea45f5b>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-26T01:21:27.030Z

Authors: Shwai He, Chenbo Jiang, Daize Dong, Liang Ding

Published: 2022-04-05T14:03:54.000Z

## Abstract

Dynamic convolution achieves better performance for efficient CNNs at the cost of negligible FLOPs increase. However, the performance increase can not match the significantly expanded number of parameters, which is the main bottleneck in real-world applications. Contrastively, mask-based unstructured pruning obtains a lightweight network by removing redundancy in the heavy network. In this paper, we propose a new framework, \\textbf{Sparse Dynamic Convolution} (\\textsc{SD-Conv}), to naturally integrate these two paths such that it can inherit the advantage of dynamic mechanism and sparsity. We first design a binary mask derived from a learnable threshold to prune static kernels, significantly reducing the parameters and computational cost but achieving higher performance in Imagenet-1K. We further transfer pretrained models into a variety of downstream tasks, showing consistently better results than baselines. We hope our SD-Conv could be an efficient alternative to conventional dynamic convolutions.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2204.02227>)

[PDF](<https://arxiv.org/pdf/2204.02227.pdf>)