Papers › Micro-Batch Training with Batch-Channel Normalization and Weight Standardization
Micro-Batch Training with Batch-Channel Normalization and Weight Standardization
Siyuan Qiao, Huiyu Wang, Chenxi Liu, Wei Shen, Alan Yuille
Batch Normalization (BN) has become an out-of-box technique to improve deep network training. However, its effectiveness is limited for micro-batch training, i.e., each GPU typically has only 1-2 images for training, which is inevitable for many computer vision tasks, e.g., object detection and semantic segmentation, constrained by memory consumption. To address this issue, we propose Weight Standardization (WS) and Batch-Channel Normalization (BCN) to bring two success factors of BN into micro-batch training: 1) the smoothing effects on the loss landscape and 2) the ability to avoid harmful elimination singularities along the training trajectory. WS standardizes the weights in convolutional layers to smooth the loss landscape by reducing the Lipschitz constants of the loss and the gradients; BCN combines batch and channel normalizations and leverages estimated statistics of the activations in convolutional layers to keep networks away from elimination singularities. We validate WS and BCN on comprehensive computer vision tasks, including image classification, object detection, instance segmentation, video recognition and semantic segmentation. All experimental results consistently show that WS and BCN improve micro-batch training significantly. Moreover, using WS and BCN with micro-batch training is even able to match or outperform the performances of BN with large-batch training.
In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.
Code
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
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
1 archive task tag without a task page not shown.
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Instance Segmentation | COCO minival | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | AP50 | 61.07 | #79 of 93 | Archive leaderboard | report |
| Instance Segmentation | COCO minival | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | AP75 | 40.82 | #79 of 93 | Archive leaderboard | report |
| Instance Segmentation | COCO minival | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | APL | 56.08 | #79 of 93 | Archive leaderboard | report |
| Instance Segmentation | COCO minival | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | APM | 41.73 | #79 of 93 | Archive leaderboard | report |
| Instance Segmentation | COCO minival | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | APS | 18.32 | #79 of 93 | Archive leaderboard | report |
| Instance Segmentation | COCO minival | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | mask AP | 38.34 | #79 of 93 | Archive leaderboard | report |
| Object Detection | COCO minival | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | AP50 | 64.15 | #144 of 220 | Archive leaderboard | report |
| Object Detection | COCO minival | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | AP75 | 47.11 | #144 of 220 | Archive leaderboard | report |
| Object Detection | COCO minival | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | APL | 56.39 | #144 of 220 | Archive leaderboard | report |
| Object Detection | COCO minival | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | APM | 47.19 | #144 of 220 | Archive leaderboard | report |
| Object Detection | COCO minival | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | APS | 25.49 | #144 of 220 | Archive leaderboard | report |
| Object Detection | COCO minival | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | box AP | 43.12 | #144 of 220 | 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
Introduced by this paper: Weight Standardization
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