This repository offers a collection of predefined deep learning models and dataset fetchers for PyTorch beginners. It includes popular architectures like ResNet, VGG, Inception, and SqueezeNet along with standard benchmarks (MNIST, CIFAR-10/100, SVHN, STL-10). The project also provides training scripts and a quantization module supporting linear, minmax, and non-linear methods to compress models to specified bit-widths.