make_contiguous¶

These operations are equivalent of numpy.ascontiguousarray() in numpy and torch.Tensor.contiguous() in torch.

metatensor.make_contiguous(tensor: TensorMap) → TensorMap[source]¶

Returns a new TensorMap where all values and gradient values arrays are made to be contiguous.

metatensor.make_contiguous_block(block: TensorBlock) → TensorBlock[source]¶

Returns a new TensorBlock where the values and gradient (if present) arrays are made to be contiguous.