Blogs

Recently, it has been shown that depth-wise convolutions are very effective in designing efficient networks, such as MobileNet and ShuffleNet. However, such efficient networks take longer to train, usually 300 – 400 epochs, to achieve state-of-the-art accuracy on the ImageNet dataset. In this article, we describe an effective learning rate scheduler introduced in the ESPNetv2 paper (CVPR’19) that allows training efficient networks in about 100 epochs without having any significant impact on the accuracy.
Click here for more details.

Nowadays, a number of real-world applications, such as autonomous vehicles, involves visual scene understanding. Semantic segmentation is one of the main tasks that opens the way for visual scene understanding. However, it is one of the most computationally expensive tasks in computer vision. This article provides an overview of an efficient semantic segmentation network that is used in the ESPNetv2 paper.
Click here for more details.