ak.to_backend#

Defined in awkward.operations.ak_to_backend on line 16.

ak.to_backend(array, backend, *, highlevel=True, behavior=None, attrs=None)#

Returns an array on a different backend (kernel set).

Any components that are already in the desired backend are viewed, rather than copied, so this operation can be an inexpensive way to ensure that an array is ready for a particular library.

To use "cuda", the cupy package must be installed, either with:

pip install cupy

or:

conda install -c conda-forge cupy

To use "jax", the jax package must be installed, either with:

pip install jax

or:

conda install -c conda-forge jax

See ak.kernels.

Parameters:

  • array – Array-like data (anything ak.to_layout recognizes).

  • backend ("cpu", "cuda", "jax", or "typetracer") – If "cpu", the array structure is recursively copied (if need be) to main memory for use with the default Numpy backend; if "cuda", the structure is copied to the GPU(s) for use with CuPy. If "jax", the structure is copied to the CPU for use with JAX.

  • highlevel (bool) – If True, return an ak.Array; otherwise, return a low-level ak.contents.Content subclass.

  • behavior (None or dict) – Custom ak.behavior for the output array, if high-level.

  • attrs (None or dict) – Custom attributes for the output array, if high-level.

Returns:

An array with the same data as the input, moved to the requested backend (kernel set).