Pooling池化是卷积神经网络中的重要组成部分, 根据任务需求, 模型架构和计算资源所使用的池化也有所不同,其中最常见的是最大池化(Max Pooling)和平均池化(Average Pooling), 而空间金字塔池化统一输入图像的尺度,全局池化是为了减小参数

Layer Normalization

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression

ADAM: A METHOD FOR STOCHASTIC OPTIMIZATION

HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis