论文标题
EAUTODET：有效的体系结构搜索对象检测
EAutoDet: Efficient Architecture Search for Object Detection
论文作者
论文摘要
由于大型数据集和复杂的网络模块，用于检测的训练CNN正在耗时，因此很难直接在检测数据集中搜索体系结构，这通常需要巨大的搜索成本（通常是数十个甚至数百个GPU日）。相比之下，本文引入了一个名为Eautodet的高效框架，该框架可以在1.4 GPU天内发现实用的骨干和FPN架构以进行对象检测。具体而言，我们为骨干和FPN模块构建了一个超网，并采用了可区分的方法。为了减少GPU内存需求和计算成本，我们通过在一个边缘上共享候选操作的权重并将其整合到一个卷积中来提出一种内核重复使用技术。搜索通道编号还引入了动态通道改进策略。广泛的实验显示了我们方法的显着疗效和效率。特别是，发现的体系结构超过了最新的对象检测NAS方法，并在可可Test-DEV集合设置的41.3 fps实现了40.1 MAP，并具有41.3 fps的49.2 MAP。我们还将发现的体系结构转移到旋转检测任务中，该任务达到了77.05地图$ _ {\ text {50}} $，dota-v1.0在带有211万参数的测试设置上。
Training CNN for detection is time-consuming due to the large dataset and complex network modules, making it hard to search architectures on detection datasets directly, which usually requires vast search costs (usually tens and even hundreds of GPU-days). In contrast, this paper introduces an efficient framework, named EAutoDet, that can discover practical backbone and FPN architectures for object detection in 1.4 GPU-days. Specifically, we construct a supernet for both backbone and FPN modules and adopt the differentiable method. To reduce the GPU memory requirement and computational cost, we propose a kernel reusing technique by sharing the weights of candidate operations on one edge and consolidating them into one convolution. A dynamic channel refinement strategy is also introduced to search channel numbers. Extensive experiments show significant efficacy and efficiency of our method. In particular, the discovered architectures surpass state-of-the-art object detection NAS methods and achieve 40.1 mAP with 120 FPS and 49.2 mAP with 41.3 FPS on COCO test-dev set. We also transfer the discovered architectures to rotation detection task, which achieve 77.05 mAP$_{\text{50}}$ on DOTA-v1.0 test set with 21.1M parameters.
