Deep Dive into the Google Inception Network Architecture!
In recent years, mainly due to advances in deep learning, more concretely convolutional networks, the quality of image recognition and object detection has improved at a spectacular rate. Most of this progress is not only the result of more powerful hardware, larger data sets, and larger models, but primarily the result of new ideas, algorithms, and improved network architectures. In this article, we will focus on one such innovative idea — Inception networks.
we previously discussed other popular ideas in this domain such as LeNet, VGG, ResNet and AlexNet. Let’s suppose now, you want to build a more complex deep neural network like these ones, what are the challenges you face when adding complexity?
When building a deep neural network, we are faced two main challenges while adding a new layer:
- What should be the filter size should we choose — 3x3, 5x5 or 1x3 or something else?
- should we choose a convolutional layer or a pooling layer?
The question is how I should proceed? Do I need to build multiple networks testing every time with a different layer or maybe a different kernel size? Is there a way to do them all and let the network decide what is…
