Efficient Deformable ConvNets: Rethinking Dynamic and Sparse Operator for Vision Applications
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911756049186816 |
|---|---|
| author | Xiong, Yuwen Li, Zhiqi Chen, Yuntao Wang, Feng Zhu, Xizhou Luo, Jiapeng Wang, Wenhai Lu, Tong Li, Hongsheng Qiao, Yu Lu, Lewei Zhou, Jie Dai, Jifeng |
| author_facet | Xiong, Yuwen Li, Zhiqi Chen, Yuntao Wang, Feng Zhu, Xizhou Luo, Jiapeng Wang, Wenhai Lu, Tong Li, Hongsheng Qiao, Yu Lu, Lewei Zhou, Jie Dai, Jifeng |
| contents | We introduce Deformable Convolution v4 (DCNv4), a highly efficient and effective operator designed for a broad spectrum of vision applications. DCNv4 addresses the limitations of its predecessor, DCNv3, with two key enhancements: 1. removing softmax normalization in spatial aggregation to enhance its dynamic property and expressive power and 2. optimizing memory access to minimize redundant operations for speedup. These improvements result in a significantly faster convergence compared to DCNv3 and a substantial increase in processing speed, with DCNv4 achieving more than three times the forward speed. DCNv4 demonstrates exceptional performance across various tasks, including image classification, instance and semantic segmentation, and notably, image generation. When integrated into generative models like U-Net in the latent diffusion model, DCNv4 outperforms its baseline, underscoring its possibility to enhance generative models. In practical applications, replacing DCNv3 with DCNv4 in the InternImage model to create FlashInternImage results in up to 80% speed increase and further performance improvement without further modifications. The advancements in speed and efficiency of DCNv4, combined with its robust performance across diverse vision tasks, show its potential as a foundational building block for future vision models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_06197 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Efficient Deformable ConvNets: Rethinking Dynamic and Sparse Operator for Vision Applications Xiong, Yuwen Li, Zhiqi Chen, Yuntao Wang, Feng Zhu, Xizhou Luo, Jiapeng Wang, Wenhai Lu, Tong Li, Hongsheng Qiao, Yu Lu, Lewei Zhou, Jie Dai, Jifeng Computer Vision and Pattern Recognition We introduce Deformable Convolution v4 (DCNv4), a highly efficient and effective operator designed for a broad spectrum of vision applications. DCNv4 addresses the limitations of its predecessor, DCNv3, with two key enhancements: 1. removing softmax normalization in spatial aggregation to enhance its dynamic property and expressive power and 2. optimizing memory access to minimize redundant operations for speedup. These improvements result in a significantly faster convergence compared to DCNv3 and a substantial increase in processing speed, with DCNv4 achieving more than three times the forward speed. DCNv4 demonstrates exceptional performance across various tasks, including image classification, instance and semantic segmentation, and notably, image generation. When integrated into generative models like U-Net in the latent diffusion model, DCNv4 outperforms its baseline, underscoring its possibility to enhance generative models. In practical applications, replacing DCNv3 with DCNv4 in the InternImage model to create FlashInternImage results in up to 80% speed increase and further performance improvement without further modifications. The advancements in speed and efficiency of DCNv4, combined with its robust performance across diverse vision tasks, show its potential as a foundational building block for future vision models. |
| title | Efficient Deformable ConvNets: Rethinking Dynamic and Sparse Operator for Vision Applications |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2401.06197 |