This self-paced tutorial aims to provide a comprehensive, one-stop resource for understanding the fundamentals of neural networks without the need for GPUs or internet access. This tutorial series is split into five parts: (1) an overview of neural networks (2) the math behind forward and back propagation (3) MNIST classification with a MLP in NumPy (4) MNIST classification with a MLP in PyTorch (5) MNIST classification with a CNN in PyTorch Along the way, I provide offline PDF resources with more detailed and/or supplemental information.
After completing the following steps, you will be able to get working with no need for anymore internet access or GPUs.