论文标题
快速神经架构搜索轻质密集的预测网络
Fast Neural Architecture Search for Lightweight Dense Prediction Networks
论文作者
论文摘要
我们提出了一份轻巧的密集预测神经体系结构搜索（NAS）框架。从预定义的通用主链开始，LDP应用了新颖的辅助禁忌搜索进行有效的体系结构探索。 LDP快速且适合各种密集估计问题，与以前的NAS方法相比，这些方法是计算要求或仅针对单个子任务而部署的。评估LPD的性能，对各种数据集的单眼深度估计，语义分割和图像超分辨率任务进行评估，包括NYU-DEPTH-V2，KITTI，CITTI，CITTISSCAPES，COCO-STUFF，DIV2K，SET5，SET5，SET14，SET14，BSD100，URBAN100，URBAN100，URBAN100。实验表明，所提出的框架在所有测试的密集预测任务上都能进行一致的改进，而在模型参数的数量方面，$ 5 \％-315 \％$比以前的艺术更紧凑。
We present LDP, a lightweight dense prediction neural architecture search (NAS) framework. Starting from a pre-defined generic backbone, LDP applies the novel Assisted Tabu Search for efficient architecture exploration. LDP is fast and suitable for various dense estimation problems, unlike previous NAS methods that are either computational demanding or deployed only for a single subtask. The performance of LPD is evaluated on monocular depth estimation, semantic segmentation, and image super-resolution tasks on diverse datasets, including NYU-Depth-v2, KITTI, Cityscapes, COCO-stuff, DIV2K, Set5, Set14, BSD100, Urban100. Experiments show that the proposed framework yields consistent improvements on all tested dense prediction tasks, while being $5\%-315\%$ more compact in terms of the number of model parameters than prior arts.
