Tensors

Tensors are the fundamental data structure: multi-dimensional arrays that generalize scalars, vectors, and matrices for all modern machine learning computation.

Tensors are the core data unit in AI and deep learning, essentially multi-dimensional arrays (or 'n-dimensional arrays'). They generalize lower-rank structures: a scalar is rank-0, a vector is rank-1, and a matrix is rank-2. In practice, a 3-rank tensor might represent a color image (height, width, RGB channels), while a 4-rank tensor handles a batch of images (batch size, height, width, channels). Frameworks like TensorFlow and PyTorch use these structures to store and manipulate all data, weights, and gradients, making them indispensable for training neural networks and optimizing performance on specialized hardware (GPUs, Google’s TPUs).

https://www.tensorflow.org/