Inception Net [V1] Deep Neural Network - Explained with Pytorch

GoogLeNet or Inception V1 is an interesting 22-layer deep neural network that made heavy use of 1x1 convolution in order to construct a competitive deep network in 2014.

In this tutorial, we’ll take a look at it’s usage in Pytorch with CIFAR-10, a step-by-step explanation of the paper “Going Deeper with Convolutions” as well as a walkthrough of the Pytorch implementation.

#Tabl** of Content**
- Introduction: 0:00
- GoogLeNet with Pytorch on CIFAR-10: 0:22
- Background: 3:57
- Architecture Overview: 4:45
- Inception Module: 7:04
- Architecture Details: 9:22
- GoogLeNet Training: 11:03
- GoogLeNet Result: 11:23
- GoogLeNet Pytorch Overview: 12:01
- GoogLeNet Pytorch - Inception: 13:55
- GoogLeNet Pytorch - BasicConv2D: 17:06
- GoogLeNet Pytorch - InceptionAux: 17:44
- GooGleNet Pytorch - GoogLeNet: 18:46
- Conclusion: 22:36

*Important Links:*
📌 Paper: https://arxiv.org/pdf/1409.4842
📌 Github: https://github.com/yacineMahdid/deep-...
📌 Original Code: https://github.com/pytorch/vision/blo...

*Abstract for Going Deeper with Convolutions:*
"We propose a deep convolutional neural network architecture codenamed "Inception", which was responsible for setting the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC 2014).

The main hallmark of this architecture is the improved utilization of the computing resources inside the network. This was achieved by a carefully crafted design that allows for increasing the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC 2014 is called GoogLeNet, a 22-layer deep network, the quality of which is assessed in the context of classification and detection. “

#machinelearning #deeplearning

----
Join the Discord for general discussion:   / discord  

----
Follow Me Online Here:

Twitter:   / yacineaxya  
LinkedIn:   / yacinemahdid  
___

Have a great week! 👋