CNN Architectures in PyTorch

Welcome to the CNNArchitectures repository! In this project, I have implemented four classic Convolutional Neural Network (CNN) architectures using PyTorch: AlexNet, VGG, GoogLeNet and ResNet. Each architecture has its own dedicated directory containing the implementation code and a explanation of the corresponding paper.

Directory Structure

1. AlexNet Implementation (alexnet-implementation):

   • Code: alexnet-implementation/
   • Paper: ImageNet Classification with Deep Convolutional Neural Networks
   • About: AlexNet, introduced in 2012, played a pivotal role in popularizing deep learning for image classification. It won the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) in 2012.

2. VGG Implementation (vggnet-implementation):

   • Code: vggnet-implementation/
   • Paper: Very Deep Convolutional Networks for Large-Scale Image Recognition
   • About: VGGNet, proposed in 2014, is known for its simplicity and effectiveness. It achieved state-of-the-art results in the ILSVRC 2014 competition. The network architecture is characterized by its repeated use of small convolutional filters.

3. GoogLeNet Implementation (googlenet-implementation):

   • Code: googlenet-implementation/
   • Paper: Going Deeper with Convolutions
   • About: GoogLeNet, also known as Inception v1, was proposed in 2014. It is characterized by the use of inception modules, which consist of multiple convolutions with different filter sizes. This design allows the network to capture features at different scales efficiently.

4. ResNet Implementation (resnet-implementation):

   • Code: resnet-implementation/
   • Paper: Deep Residual Learning for Image Recognition
   • About: ResNet, introduced in 2015, addresses the deep networks deterioration problem by introducing skip connections. The architecture enables the training of very deep networks and has been widely adopted in various computer vision tasks.

Getting Started

To explore and run the implementations, navigate to the respective directories and follow the instructions provided in each README file.

Feel free to experiment, modify, and build upon these implementations for your own projects. If you have any questions or suggestions, please don't hesitate to reach out :D