mrdbourke/pytorch-deep-learning

Covers fundamental tensor operations, neural network architectures, computer vision, custom dataset loading, and modular code organization through hands-on notebooks and exercises. The curriculum progresses from PyTorch basics through classification and transfer learning, then advances to practical applications including experiment tracking, research paper replication, and model deployment. PyTorch 2.0 compatible materials are supplemented with video lectures, slides, and a reference cheatsheet.

17,355 stars.