06·Python, NumPy and the first tensor — a bridge to PyTorch

Introduction

All the math from the previous five lessons (tensors, operations, derivatives, gradients, gradient descent) becomes useful only in code. That code is Python, and the most common library for numerical array computing is NumPy. NumPy taught the world to think of data as n-dimensional arrays — and all modern AI libraries (PyTorch, TensorFlow, JAX) hook into its API pattern. In this lesson we do a quick Python refresher, then move on to NumPy: how to create arrays, check their shape and dtype, index them, use broadcasting, and perform operations. At the end we show the "bridge" — how a NumPy array turns into a PyTorch tensor with one line, torch.from_numpy(). After this lesson you are ready for R13 (PyTorch) where you will start building your first models.