types.core.ArrayLike

types.core.ArrayLike

Array-like type supporting multiple backends.

Can be NumPy array, PyTorch tensor, or JAX array. Use this for maximum flexibility across backends.

Shape conventions: - Scalars: () - Vectors: (n,) - Matrices: (m, n) - Batched: (batch, …)

Examples

>>> def process(data: ArrayLike) -> ArrayLike:
...     # Works with all backends
...     return data * 2

# types.core.ArrayLike { #cdesym.types.core.ArrayLike }

`types.core.ArrayLike`

Array-like type supporting multiple backends.

Can be NumPy array, PyTorch tensor, or JAX array.
Use this for maximum flexibility across backends.

Shape conventions:
- Scalars: ()
- Vectors: (n,)
- Matrices: (m, n)
- Batched: (batch, ...)

## Examples {.doc-section .doc-section-examples}

```python
>>> def process(data: ArrayLike) -> ArrayLike:
...     # Works with all backends
...     return data * 2
```