Implementation of ResNet using PyTorch

Todos

• Dataloaders for TinyImageNet, Dogs and Cifar10
• Train with imported ResNet from torchvision
• Implement residual block
• Resnet34
• Hyper-Parameters via CLI
• Save logs and checkpoints with hyperparameters in name
• Evaluate on Tiny Imagenet
• Image preprocessing
• Build Resnet50 (resnet34 with bottlenecks)
• Evaluate again o Tiny Imagenet
• Try Dogs and Cifar10 (follow paper for cifar10)
• Implement ResNet50 with pre-activation
• Implement ResNext29