mrpro.utils.broadcasted_concatenate

mrpro.utils.broadcasted_concatenate(tensors: Sequence[Tensor], dim: int, reduce_views: bool = True) → Tensor[source]

Concatenate tensors while preserving broadcasting.

Parameters:

• tensors (Sequence[Tensor]) – The tensors to concatenate.

• dim (int) – The dimension to concatenate along.

• reduce_views (bool, default: True) – If True, reduce stride-0 dimensions to singleton dimensions after concatenating.

Returns:

The concatenated tensor.