This assumes prior experience with PyTorch or related frameworks. The content is presented in a dense manner so that it can serve as a quick reference. Readers are encouraged to run the code themselves to get a better understanding.

AI was heavily involved in some sections. However, AI still doesn't have a strong understanding of niche PyTorch behaviors, so I have had to make many corrections. I welcome any feedback about inaccuracies or suggestions for general improvements.

Tensor Creation


Tensor Fundamentals


Einsum / Einops


Model Creation


Training and Evaluation


Data Handling


Model Management


Advanced Modeling


Memory and Performance