What Is Tensor

A tensor is a data container used in frameworks such as PyTorch. You can think of it as a generalized form of numbers, vectors, and matrices:

• A single number is a 0D tensor.
• A list of numbers is a 1D tensor.
• A table of numbers is a 2D tensor.
• Higher-dimensional arrays are also tensors.

Why Tensors Matter

Tensors are the standard way to represent model inputs, outputs, parameters, and intermediate computation results in deep learning.

Where You See Them

• Images stored as width, height, and channel values
• Batches of training data
• Robot observations and actions
• Neural network weights and activations

A Practical Mental Model

If you already understand arrays in NumPy, a tensor is very similar, but with better support for GPU computation, automatic differentiation, and deep learning pipelines.