Inception v2 and v3

Inception v2 and v3 were released together and propose several improvements over the original Inception block (Rethinking the Inception Architecture for Computer Visionhttps://arxiv.org/abs/1512.00567). The first is the factorization of the 5×5 convolution in two stacked 3×3 convolutions. We discussed the advantages of this in the Introduction to Visual Geometry Group section.

We can see the new Inception block in the following diagram:

Inception block A, inspired by https://arxiv.org/abs/1512.00567

The next improvement is the factorization of an n×n convolution in two stacked asymmetrical 1×n and n×1 convolutions. For ...

Get Advanced Deep Learning with Python now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.