Abstract | ||
---|---|---|
Structural re-parameterization has drawn increasing attention in various computer vision tasks. It aims at improving the performance of deep models without introducing any inference-time cost. Though efficient during inference, such models rely heavily on the complicated training-time blocks to achieve high accuracy, leading to large extra training cost. In this paper, we present online convolutional re-parameterization (OREPA), a two-stage pipeline, aiming to reduce the huge training overhead by squeezing the complex training-time block into a single convolution. To achieve this goal, we introduce a linear scaling layer for better optimizing the online blocks. Assisted with the reduced training cost, we also explore some more effective re-param components. Compared with the state-of-the-art re-param models, OREPA is able to save the training-time memory cost by about 70% and accelerate the training speed by around 2×. Meanwhile, equipped with OREPA, the models out-perform previous methods on ImageNet by up to +0.6%. We also conduct experiments on object detection and semantic segmentation and show consistent improvements on the downstream tasks. Codes are available at https://github.com/JUGGHM/OREPA_CVPR2022. |
Year | DOI | Venue |
---|---|---|
2022 | 10.1109/CVPR52688.2022.00065 | 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) |
Keywords | DocType | Volume |
Deep learning architectures and techniques, Efficient learning and inferences | Conference | 2022 |
Issue | ISSN | ISBN |
1 | 1063-6919 | 978-1-6654-6947-0 |
Citations | PageRank | References |
0 | 0.34 | 3 |
Authors | ||
7 |
Name | Order | Citations | PageRank |
---|---|---|---|
Mu Hu | 1 | 0 | 1.01 |
Junyi Feng | 2 | 0 | 0.34 |
Jiashen Hua | 3 | 0 | 0.34 |
Baisheng Lai | 4 | 9 | 2.54 |
Jianqiang Huang | 5 | 55 | 19.18 |
Xiaojin Gong | 6 | 195 | 17.08 |
Xian-Sheng Hua | 7 | 6566 | 328.17 |