Practical PyTorch — tensors, autograd, training loops, and best practices.
PyTorch tensors — creation, manipulation, and GPU acceleration.