Neural Networks with PyTorch (Python Tutorial)

Neural networks are often described as "inspired by the brain." But what does that actually mean, in code? In this video we build a working neural network from the ground up using PyTorch, Python's most popular deep learning library.

We begin with a quick refresher on tensors, then walk through the ideas that let a network learn: automatic differentiation with Autograd, the architecture of layers and neurons, activation functions like ReLU, loss functions and optimizers, and finally a complete training loop where you can watch the error shrink, epoch by epoch.

Support Socratica on Patreon:
https://www.patreon.com/socratica