Papers › On the Ideal Number of Groups for Isometric Gradient Propagation
On the Ideal Number of Groups for Isometric Gradient Propagation
Bum Jun Kim, Hyeyeon Choi, Hyeonah Jang, Sang Woo Kim
Recently, various normalization layers have been proposed to stabilize the training of deep neural networks. Among them, group normalization is a generalization of layer normalization and instance normalization by allowing a degree of freedom in the number of groups it uses. However, to determine the optimal number of groups, trial-and-error-based hyperparameter tuning is required, and such experiments are time-consuming. In this study, we discuss a reasonable method for setting the number of groups. First, we find that the number of groups influences the gradient behavior of the group normalization layer. Based on this observation, we derive the ideal number of groups, which calibrates the gradient scale to facilitate gradient descent optimization. Our proposed number of groups is theoretically grounded, architecture-aware, and can provide a proper value in a layer-wise manner for all layers. The proposed method exhibited improved performance over existing methods in numerous neural network architectures, tasks, and datasets.
Code
No code repository is listed for this paper in the archive or in Syntology's graph.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Fine-Grained Image Classification | Caltech-101 | ResNet-101 (ideal number of groups) | Top-1 Error Rate | 22.247% | #15 of 18 | Archive leaderboard | report |
| Fine-Grained Image Classification | Oxford-IIIT Pets | ResNet-101 (ideal number of groups) | Accuracy | 77.076 | #18 of 19 | Archive leaderboard | report |
| Image Classification | MNIST | MLP (ideal number of groups) | Percentage error | 1.67 | #53 of 81 | Archive leaderboard | report |
| Object Detection | COCO 2017 | Faster R-CNN (ideal number of groups) | AP | 40.7 | #7 of 24 | Archive leaderboard | report |
| Object Detection | COCO 2017 | Faster R-CNN (ideal number of groups) | AP50 | 61.2 | #7 of 24 | Archive leaderboard | report |
| Object Detection | COCO 2017 | Faster R-CNN (ideal number of groups) | AP75 | 44.6 | #7 of 24 | Archive leaderboard | report |
| Panoptic Segmentation | COCO panoptic | PFPN (ideal number of groups) | PQ | 42.147 | #2 of 2 | Archive leaderboard | report |
| Panoptic Segmentation | COCO panoptic | PFPN (ideal number of groups) | PQst | 30.572 | #2 of 2 | Archive leaderboard | report |
| Panoptic Segmentation | COCO panoptic | PFPN (ideal number of groups) | PQth | 49.816 | #2 of 2 | 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
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