Arrays vs tensors

Arrays

Use arrays for:

• general-purpose lists,

• mapping/reducing,

• heterogeneous data (pairs, objects, strings, etc.).

Tensors

Use tensors for:

• numeric computation where dimensionality matters,

• optimization/AD-heavy code,

• neural components and linear algebra.

Key gotcha: “flat storage”

When you build a tensor from a list, it is often flat + dims. So for a 2x2 tensor you typically pass 4 values, not nested lists.

Example: Tensor([2, 2], [1, 2, 3, 4]).