摘要： 一种简单的解释是用来降维。 For example, an image of 200*200 with 50 features on convolution with 20 filters of 1*1 would result in size of 200*200*20. 但是，1*1卷积核的作用 阅读全文