Receptive Fields: Why 3x3 conv layer is the best?

Описание к видео Receptive Fields: Why 3x3 conv layer is the best?

In this video, we understand the concept of receptive fields in deep learning and the reason why using 3x3 conv layer is more effective than 5x5 or 7x7 layers.

Table of content:
00:00 Intro
00:18 Receptive Fields
03:55 Max-pooling effect
04:24 Why 3x3 conv layers!?

Icon made by Freepik from flaticon.com

Комментарии

Информация по комментариям в разработке