An Introduction to PyTorch & Autograd

Explore the fundamentals of PyTorch and its Autograd package in this 32-minute conference talk from EuroPython 2017. Dive into the world of GPU programming frameworks and discover how PyTorch's "define-by-run" approach enables dynamic computation graphs, offering new possibilities for neural network design. Learn about efficient tensor representations, hardware-agnostic operations, and the mechanics of Algorithmic Differentiation. Through simple examples, gain insights into key concepts such as tensors, variables, back propagation, and linear regression. Compare PyTorch with other frameworks and understand its unique features for deep learning and machine learning applications. Perfect for those interested in GPU processing, deep learning, and linear algebra.

Introduction
PyTorch
Tensors
Variables
Back Propagation
Back Propagation Example
Tangent Example
Linear Regression
Machine Learning
Training
Comparison