1. Convolution Neural Networks are a core class of Neural Networks that are often operated on visual imagery. CNN architectures are commonly used for Image Classification, implement ResNet18 from scratch. Make your submission using the following steps:
- Fork and clone this Github repository.
- Create a new branch and set its name to your “Name”.
- Create a folder named "Name" (or Handle) in which you have all the contents of whatever submission you wanna make.
- Merge your branch with the master branch.
- Pull origin & add files, commit and push to fork.
- Create pull request on main repo.
Brownie Points for implementing any ResNet (takes block and layers)
2. Hint: You have to implement the torch.nn.module class for the models.
3. Tip: Look at the resources attached, and try exploring more articles/papers to develop a deeper understanding.
- Recording for the Session
- Colab Notebook:
- An Overview of ResNet and its variants
- An implementation of ResNet
- Difference between AlexNet, VGGNet, ResNet, and Inception
- The Original paper on ResNet
Caution: Do not copy pre-existing implementations blindly.